Palgrave Studies in

Financial Services
Technology

Insurance
Transformed
Technological Disruption

Michael Naylor
Palgrave Studies in Financial Services Technology

Series Editor
Bernardo Nicoletti
Rome
Italy
The Palgrave Studies in Financial Services Technology series features
original research from leading and emerging scholars on contemporary
issues and developments in financial services technology. Falling into
4 broad categories: channels, payments, credit, and governance; topics
covered include payments, mobile payments, trading and foreign trans-
actions, big data, risk, compliance, and business intelligence to support
consumer and commercial financial services. Covering all topics within
the life cycle of financial services, from channels to risk management,
from security to advanced applications, from information systems to
automation, the series also covers the full range of sectors: retail bank-
ing, private banking, corporate banking, custody and brokerage, whole-
sale banking, and insurance companies. Titles within the series will be
of value to both academics and those working in the management of
financial services.

More information about this series at

http://www.springer.com/gp/series/14627
 Michael Naylor

Insurance
Transformed
Technological Disruption
Michael Naylor
School of Economics and Finance
Massey University
Palmerston North, New Zealand

Palgrave Studies in Financial Services Technology

ISBN 978-3-319-63834-8 ISBN 978-3-319-63835-5  (eBook)
DOI 10.1007/978-3-319-63835-5

Library of Congress Control Number: 2017948716

© The Editor(s) (if applicable) and The Author(s) 2017

This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by
similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein
or for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.

Printed on acid-free paper

This Palgrave Macmillan imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
 Contents

1 Exponential Change 1

2 Key Technological Disruptors 15

3 Types of Insurance 41

4 The Impact of Disruptive Technology 47

5 What Insurance Companies Need to Do 93

6 Impact by Type of Insurance 175

7 The Dynamics of Decline 209

8 The Response of Incumbents 221

9 Regulatory and Legal Issues 263

10 Consequences for Insurance Workers 281

v
vi    
Contents

11 Impacted Occupations 297

12 Conclusion 321

Useful Sources 325

Index 335
 List of Figures

Fig. 1.1 Linear vs. Exponential change 5

Fig. 1.2 What we expect based on where we are at the moment vs.
what the future is actually like 6
Fig. 1.3 Hype vs. Product maturity. © Gartner, used with permission 9
Fig. 1.4 Gartner Hype Cycle phases. © Gartner, used with permission 10
Fig. 1.5 Exponential change and the Hype-cycle 11
Fig. 2.1 A dynamic node network 27
Fig. 2.2 Neutral Network 28
Fig. 4.1 Fixed vs. marginal cost diagram 49
Fig. 4.2 Changes in the spread of insurance rates 84
Fig. 5.1 Possible insurance ecosystems 160
Fig. 8.1 Market Entry Dynamics 230
Fig. 8.2 Stages of Digital Disruption 235
Fig. 8.3 The Digital matrix. © Boston Consulting Group,
used with permission 238
Fig. 8.4 The Action Space for Digital Transformation.
© Boston Consulting Group, used with permission 240

vii
 1
Exponential Change

Introduction
Insurance has traditionally been a very conservative industry, and this
includes conservatism in the way it has used information technology.
This conservatism applies to both companies’ interactions with clients
and to interactions within the company. This conservatism is about to
change as the insurance industry is currently on the crest of a combina-
tion of technological advances in technology which, when combined,
will utterly disrupt the current insurance industry and its interactions
with clients, and disrupt up to 80% of current insurance job activities.
Coping with this challenge will be the key insurance management issue
for the next few decades, and success at adapting to technological dis-
ruption will be the defining characteristic of industry survivors.
Some of the elements of the disruptive technologies have been
around for a while; others are new, whilst some are not yet useable. Each
of these technologies has individually, as yet, not had much impact on
insurance, and this has lolled the insurance industry into complacency.
This complacency is starting to crack with the insurance consultants
from 2015 waking up to the possibilities, so that by 2017 discussion

© The Author(s) 2017 1

M. Naylor, Insurance Transformed, Palgrave Studies in Financial Services Technology,
DOI 10.1007/978-3-319-63835-5_1
2    
M. Naylor

of technological disruption was the fashion. No commentators, how-

ever, have realized that it is the combination of these technologies which
is the disruptor, not the impact of each individually. No commentator
has realized that when added together, the small impacts of each tech-
nology will combine to create utter disruption in the insurance indus-
try, so that insurance will emerge as an industry transformed.
While these technologies will disrupt all industrial and service sec-
tors, PWC (2015) and the World Economic Forum (2015) argue that
in the next decade insurance will be the most disrupted of all the major
service sectors. This disruption is not a once-off; it will be continuous
and geometrically increasing challenge, that is, the rate of change will
increase, so disruption will increasingly increase. Insurance compa-
nies who do not respond dynamically will not survive as the ability to
respond innovatively to disruption will become as vital in insurance as
it has in mobile phones. The key management skill in the future will
be the ability to fundamentally transform companies so that they are
able to cope with continuous change. Most current insurers will struggle
with this, resulting in a sector ripe for disruptive external entrants.
Disruption is also occurring because customers’ experiences with
technological leaders in other sectors are transforming their expectations
of insurance customer service. Consumers do not now judge their cus-
tomer experience of an insurance company against that of other insur-
ance companies, but instead judge it against their superior experiences
with digital leaders such as Amazon. Yet insurers still tend to benchmark
themselves against other insurers instead of benchmarking against pos-
sible external disruptors. Who would have guessed that the PC manu-
facturers would create a product which would destroy incumbents in
diverse industries such as telephones, diaries, maps, and photography?
Insurers need to understand the scale of the transformation needed or
risk being left behind.
This meta-study outlines the key disruptors, outlines how these will
impact on the insurance sector, including advisers, and what insurers
will have to do to survive. It is vital to understand that these changes
will not occur as a stand-alone Uber moment, but will consist of a series
of increasingly high waves of disruption. There is no end point - the
 1  Exponential Change    
3

future is one of increasingly fast change. Insurers need to reimagine

their product and their industry.
The end result and the timing are uncertain, so the research can only
provide informed speculation. While society as a whole will undoubtedly
be better off because of the coming changes, some insurance companies,
and their workers, will not be. The question is: ‘what is the likely impact
of the looming technological disruption on the insurance industry, on
insurance advisers, and on jobs?’

The Third Wave

Historians generally recognize two waves1 of technological change in
the modern era: (i) the late eighteenth-century steam/mechanical indus-
trial revolution, and (ii) the late nineteenth-century electric power/
motor car revolution. Both of these transformed society and led to far
higher average income and better lives for society. Both were, however,
quite disruptive to existing social structures and employment. The first
industrial revolution, for example, involved a change from a society
which was 90% agricultural and rural, to one which was 90% industrial
and urban. A lot of the benefits, which we now enjoy and see as essen-
tial, were not appreciated by common people at the time, with major
protests against both the water wheel and the motorcar. However, in
every case, society as a whole gained. The current IT revolution, which
started about 1970, should be recognized as constituting a third wave,
one which will transform our society as thoroughly as the two earlier
industrial revolutions did.
It is useful to understanding to note that each of the first two waves
took 80–100 years or more to work its way through its impact on soci-
ety, with most of the social impacts occurring only towards the end of
the wave. This is because technological revolutions involve a combina-
tion of inter-linked technologies, each embodying a small change which
builds on the others, ensuring that the depth and spread of change

1Some Historians include a third or fourth wave, but these do not change my point.
4    
M. Naylor

speeds up only over time. The initial stages of the first industrial rev-
olution, for example, had little impact outside the cotton industry for
about 50 years, and then it transformed society. The social impact of the
IT revolution as it nears its 50-year mark, while it has changed soci-
ety in a number of ways, should thus be seen as only at an initial stage,
with the biggest social changes due to occur over the next 50 years. The
scope and scale of those substantial changes are only now starting to
become evident. Many of these future changes will be unexpected and
unforeseeable.

Exponential Change
Exponential change is change which follows an ever-rising upward
curve, so that changes occur at an ever-increasing rate. For example, it
took 38 years for radio to reach 50 million people, 20 years for phone,
13 years for the television, 3.6 years for Facebook, 88 days for Google
Plus, and 35 days for Angry Birds. Most people find this type of change
hard to deal with, as even though we think we are now used to change,
humans typically have an inherent tendency to be unable to visualize
exponential change. For example, the Internet was originally invented in
the mid-1970s, but it only become useful outside a narrow niche after
the i­nvention of a graphical browser by Netscape in 1994. At that stage,
nearly all forecasters failed to see its potential. Cliff Stoll, a respected net-
work expert, argued in a 1995 Newsweek article2 that ‘online shopping
and entertainment were an unrealistic fantasy’ and ‘the truth is that no
online database will replace your newspaper’ and then he poured scorn on
predictions that we would soon be downloading books via the Internet.
Commercial use was seen as impossible. Thus, looking back, while the
Internet revolution which has occurred now seems obvious, in 1995 it was
nearly unimaginable. Also, nobody predicted in 1980 that in 30 years’ time
each of us would be carrying the power of a 1980s mainframe computer

2Stoll, C (1995) ‘Why the Web won’t be Nirvana,’ Newsweek, Feb 27 (original title ‘The

Internet? Bah!’).
 1  Exponential Change    
5

Technological Change Exponential

Progress

change:
how tech
People under changes
expect

Linear
Expectations:
People over what people
expect expect

Years

Fig. 1.1  Linear vs. Exponential change. Source Author

in our pockets. We currently find it nearly impossible to forecast the

potential of a mobile device with 1000x today’s capability.
The reason for this inability to predict the future is because we are
hard-wired to expect change to be linear. This means that when they
forecast a transformative event, people will assume that future changes
continue at about the same rate as past changes, with an equal rate of
change and equal social impact over each time period. However, tech-
nological change is predominately nonlinear, normally exponential.
This means that initial predictions of transformative change tend to
be wrong. An inherent issue with complex exponential change is that
80% of the social change tends to happen in the last quarter of the
period. This is because for most technologies to be useful, a number of
structural transformations have to occur first, and these have to work
together to be transformative. Most technologies also require changes
in businesses/social structures before they can have a major impact. As
most social change does not occur until a number of technologies inter-
act, and this takes a while, social changes tend to accumulate during the
last stage of the technological revolution period. These assumptions are
contrasted in Fig. 1.1.
6    
M. Naylor

Current position What we are

Progress

Progress
in time about to face

Time Time

Fig. 1.2  What we expect based on where we are at the moment vs. what the
future is actually like. Source Author

The result of this faulty expectation is that people tend to be under-

impressed by the rate of change during the first 2/3rds of a disruptive
revolution, just before they are shocked by the rate of change during
the last 1/3rd,3 as shown in Fig. 1.2. This has occurred with the impact
of IT, which so far in occupational terms, typists apart, has not had as
much disruptive impact as people predicted in the 1990s. Even experts
get this wrong. In the mid-1980s, AT&T hired McKinsey to forecast
mobile phone adoption by the year 2000. They predicted 900,000 sub-
scribers. The actual number was 109M, off by 120 times!
It is important to remember‚ however‚ that the underlying technolog-
ical change has been unprecedented. Brynjolfsson and McAffe (2011)
state that the effectiveness of computers increased 43m-fold between
1988 and 2003, and Nordhaus (2007) notes that during the 1980s and
1990s the computing costs fell by an average of 64% per year. The para-
dox is that up until about 2006, the impact of IT on overall rates of
productivity was minor. Even since then, change has occurred at a rate
we can cope with. Over the next 50 years, however, this will reverse,
with society currently very ill-prepared for the drastic changes which
can be expected. The rate of change will particularly accelerate over the
last 3rd of the IT revolution, creating society-wide disruption.
Society is often initially skeptical about transformative technolo-
gies and tends to focus on the transformative technology’s potential for
destruction, rather than creation; in Plato’s Phaedrus‚ Socrates worried

3For a commentary on rate of change of technology see: http://waitbutwhy.com/2015/01/artifi-

cial-intelligence-revolution-1.html.
 1  Exponential Change    
7

that the recent invention of writing would have a deleterious effect on

youth, as it would harm their memories; when books started to roll off
Gutenberg’s press, many argued that the books would be confusing and
harmful by overwhelming young people with excess information; when
Marconi invented the radio, many worried that children’s minds would
be damaged by dangerous ideas and family life would be ruined; many
experts warned against the cancer risk from mobile phones. Possibly,
some early homo sapiens warned of the dangers of fire?
Yet once transformations have taken place, society tends to regard
them as obvious. For example, experts struggled for decades to achieve
even basic language recognition. Many experts regarded IBM’s Jeopardy
goal as impossible. Yet people now regard the ability of Siri to talk and
respond naturally as a trivial achievement. Larry Tesler quipped that
‘Intelligence is whatever machines haven’t done yet.’

The Ah-Ha Moment

Within every technological revolution, there comes moments, ‘ah-ha’
events, when humanity realizes with the shock the potential and scale of
the looming revolution.
One of the core elements of the current revolution is the ability of
software to handle complex tasks. One of the key ‘ah-ha’ moments
came at the 2007 3rd Darpa4 autonomous vehicle challenge. When
the challenge started in 2004, creating software able to handle driving
was widely regarded as far too complex to be automated; driving was a
‘complex task,’ which is defined as a situation involving ill-defined rules,
a fluid, dynamic, environment, and the need for instantaneous adjust-
ment based on real-time feedback.
Up to that point, analysts had assumed that software could only be
used where there were well-defined and repetitive tasks amenable to
yes/no rules, within a restricted environment. Sure enough, vehicles in
the 2004 Darpa challenge failed to handle open desert, mostly crash-
ing shortly after starting. Two of the top experts in the field, Levy and

4Defence Advanced Research Projects Agency.

8    
M. Naylor

Murnane, in their 2004 book, explained why handling ‘complex tasks’

like driving was just about impossible for a computer and would take
many decades to solve. They stated that the idea of software being able
to handle the complexity of ensuring a computer-driven car made a safe
turn across a flow of irregular oncoming traffic was just laughable, as
detecting and predicting the movement of other vehicles was extremely
complex, and was too unpredictable to be reduced to set rules.
The software community was thus stunned when in the Darpa 2007
challenge, autonomous vehicles successfully handled a complex urban
mock-up, happily obeying road rules, and safely interacting with other
vehicles. The speed of improvement in the ability of software to handle
complex feedback was stunning and unprecedented. You can now watch
Google cars do even more complex tasks on YouTube. Autonomous cars
have now driven themselves across the USA (99% of the time).
The idea that software could handle complex tasks changed everything -
programmers worldwide have since then been busy re-examining complex
areas previously thought too hard. By 2016, the sense of wonder at auton-
omous driving is starting to fade and we now find it hard to understand
why we were stunned that cars could drive themselves. Our expectations
of what software can handle have been transformed.
This ah-ha moment had huge implications for the disruptive impact
of technology, as programmers eagerly attacked many tasks previously
regarded as impossible. For example, speech recognition was for decades
seen as impossible, yet 5-year olds now expect Siri to understand them.
These complex tasks often involve a type of programming called ‘artifi-
cial intelligence,’ which involves creating a generalized program which is
‘teachable’ by importing large data sets and running these through the
program so it weights responses based on past events. In 2011, these
AI techniques were used to create software which could win at the TV
game show ‘Jeopardy,’ which was previously regarded as impossible for
software because of its subtle and complex word-play questions, and
then, even more impressively, in 2016, to beat the reigning world cham-
pion at the game of Go, also previously regarded as impossible. As most
insurance tasks involve easier algorithms than these tasks, there are few
areas of insurance inherently immune to computerization. Compared to
playing Go‚ most underwriting tasks are child’s play.
 1  Exponential Change    
9

There is no question that the insurance industry is far behind tech-

nologically. Millennials find that frustrating. We live in a world where
if you want something you go on your phone and get it instantly. The
insurance industry currently just is not like that.

Hype and the Technological Innovation Cycle

Gartner Corporation5 argues that the inability of humans to cope with
exponential events leads to new technologies going through a ‘hype-
cycle,’ whereby expectations about technologies tend to be initially
over-hyped, and then because they cannot deliver fast, there is disillu-
sionment, followed later by actual solid products. Gartner Hype Cycles
contrast product maturity versus hype/expectations and provide a
graphic representation of the maturity and adoption of technologies and
applications. The Hype Cycle is shown in Fig. 1.3. Each Hype Cycle
drills down into the five key phases of a technology’s life cycle:

Gartner’s Five-Step Hype Cycle

Visibility/ Hype/ Expectations

Peak of Inflated
Expectations
Plateau of
Productivity

Slope of
Enlightenment

Trough of
Disillusionment
Technology @ Gartner.com
Trigger

Product Maturity

Fig. 1.3  Hype vs. Product maturity. © Gartner, used with permission

5Gartner.com; http://www.gartner.com/technology/research/methodologies/.
10    
M. Naylor

Technology Trigger: A potential technology breakthrough kicks things off.

Early proof-of-concept stories and media interest trigger significant public-
ity. Often, no usable products exist and commercial viability is unproven.
Peak of Inflated Expectations: Expectations reach a fever pitch. Early
publicity produces a number of success stories - the scores of failures are
ignored.
Trough of Disillusionment: Interest wanes as experiments and beta
products fail to match expectations. Producers of the technology shake
out or fail. Investments continue only if the surviving providers improve
their products to the satisfaction of early adopters.
Slope of Enlightenment: Product users start to understand how to use
the technology productively. More instances of how the technology can
benefit the enterprise start to crystallize and become more widely under-
stood. Second-generation products with far fewer faults appear from
technology providers. More enterprises fund pilots; conservative compa-
nies remain cautious and start to fall behind (Fig. 1.4).
Plateau of Productivity: Polished third-generation products arrive.
Mainstream adoption starts to take off. Criteria for assessing provider
viability are more clearly defined. The technology’s broad market appli-
cability and relevance are clearly paying off.

First-generation
products, high price, Negative press
customization needed begins
Hype

Consolidation
and failures High growth phase starts,
approx 30% of target audience
has adopted or is adopting
At the Peak Second round technology
of venture capital Third-generation
funding products, work out
No working Technology accepted
of-the-box as trival/normal
products

Mass media
hype begins Less than 5% Plateau
On the Rise adoption
Sliding into
the Trough
Start-up companies,
venture capital Post-
Climbing Plateau
Second-generation the Slope
Laboratory
prototypes products, some services

R&D
@Gartner.com
Technology Trigger
Maturity

Fig. 1.4  Gartner Hype Cycle phases. © Gartner, used with permission

 1  Exponential Change    
11

$FWXDO

3URJUHVV7HFK,QQRYDWLRQF\FOH+\SH
FKDQJH

([SHFWHG
&KDQJH

+\SH

7LPH
%HWD +\SH 'LVDS 6KRFN ,QWHJ

Fig. 1.5  Exponential change and the Hype-cycle. Source Author

Maturity: Mature products drop off the chart, as do failed products.

The hype is over and the products are seen as ‘normal.’
When the faulty reaction to exponential change shown in Fig. 1.1
is combined with the Gartner Hype Cycle this creates, a technological
innovation cycle. This is shown in Fig. 1.5.
Beta—New technologies are normally only understood by and avail-
able to a small number of high-tech users. The technology is normally
immature, in beta format, and both users and producers are unsure of
how the technology can be best used. Products are difficult to use. The
only users are those prepared to spend time and effort working with
imperfect technology, either because they are fascinated by it or because
they like the prestige of having tech which no one else does. The tech-
nology tends to undergo substantial change during this stage as it tends
to just be a vague collection of exciting ideas. Many of the initial firms
do not survive.
Hype—Technologies which survive the beta stage get hyped up
initially in the technology media, and then in the general media.
Excitement spreads as new possibilities are discussed. These tend to
assume that the technology will have a major social impact in the near
future. The difficulties and imperfections of the technology are ignored.
As stock valuations soar, heavy investment takes place. This is the best
time for market leaders to issue a stock market IPO.
Disappointment—Generally, the new technology takes a while
to spread, and little actual impact on social structures or life style is
12    
M. Naylor

observed. Expectations of change exceed actual progress, so impa-

tience with the slow pace of impact grows. This is despite the technol-
ogy maturing from a vague collection of ideas into useful products. The
media move onto the ‘next big thing.’ Non-using companies reassure
themselves that there is little threat. Producers are thinned out to those
with useable products or deep pockets.
Shock—Products become useful, easy-to-use, and increasingly widely
available. As actual progress catches up with expectations, the technol-
ogy spreads and market share soars. Consumers discuss usefulness and
whether or not to buy the product rather than fantasizing about possibil-
ities. When progress soars past existing incumbent technology, producers
of now-non-viable existing products find themselves scrambling to keep
their products viable and rush to produce some version of the superior
technology. A substantial proportion of old-product firms goes bankrupt
or are taken over. Profits of new-tech firms soar, though cash flow can be
under pressure due to rapid expansion.
Integration—The product becomes the new ‘normal,’ and consumer
focus turns to the comparative merits of features. If the market is com-
petitive, profits margins drop as firm struggle to provide the most fea-
tures at the lowest cost. As technology is standardized, low-cost producers
expand market share. Old-tech producers become a fading memory.
For incumbents, the lesson is that the beta versions of disruptive tech-
nologies will initially seem to be a minor threat and then will seem to
‘come-from-nowhere’ to take over a market. In a world of exponential
change, where everything is changing, the biggest risk is standing still.
It is useful to distinguish between normal ‘incremental technologies,’
those which lead to little societal change, and ‘disruptive technologies,’
those which change the way society or business runs. Nearly all the
technological changes which insurance companies have encountered so
far have been incremental and could be dealt with by iterative change
to existing business systems. What the insurance sector is facing now
is a collection of technologies, which when combined, creates a perfect
storm of technological disruption. This cannot be handled by incremen-
tal adjustments, but needs a transformative change.
 1  Exponential Change    
13

References
Brynjolfsson, E. R., Hitt, L., & Kim, H. (2011). Strength in numbers: How
does data-driven decision-making affect firm performance? Social Science
Review Network, April.
Economist (2014). The Third Great Wave‚ special report‚ October 4th.
Levy, F., & Murnane, R. J. (2004). The new division of labor: How computers
are creating the new job market. USA: Princeton University Press.
Nordhaus, W. D. (2007). Two centuries of productivity growth in computing.
The Journal of Economic History, 67(1), 128.
PWC. (2015). Insurance 2020 & Beyond: Necessity is the mother of reinvention.
World Economic Forum. (2015). The future of financial services: How dis-
ruptive innovations are reshaping the way financial services are structured,
provisioned and consumed.
 2
Key Technological Disruptors

The coming disruptive revolution will involve not just AI programming,

but also a range of disruption drivers of the IT revolution, which are
evolving at about the same time. The drivers are:

Internet of Things
One of the key elements of a connected world is sensor chips, called
‘telematics’, which give constant real-time feedback to data centers on
their current state. An unheralded breakthrough for telematics came in
July 2015, when it was announced that scientists had discovered how
to power these devices by Wi-fi.1 They can also be powered by radio
waves. This is revolutionary as it means that no internal power will be
needed, no need for wires or changing batteries, thus enabling the inser-
tion of these devices into everything. In 2015, five quintillion (10 to
the power of 18) chips were added to things which were not computers.

1For a discussion, see Talla et al. (2015). In general, these devices need no internal power to trans-
mit as they use the energy provided by radio and other waves used to connect with them.

© The Author(s) 2017 15

M. Naylor, Insurance Transformed, Palgrave Studies in Financial Services Technology,
DOI 10.1007/978-3-319-63835-5_2
16    
M. Naylor

The recent drastic reduction in the price of these sensor chips means
that they are now being embedded in everyday items like cars, fridges,
groceries, street lights, and drains.
Morgan-Stanley/BCG (2013) forecast that wearable telematics will
rise from 6M in 2013 to 248M in 2017. McKinsey (2015) estimates
that the number of self-supporting interconnected devices, in general,
will grow from 5M now to 50B by 2025, which is probably an under-
estimate as they assumed that these devices would need a power source.
The total number of networked items is far higher than that once you
include all the unintelligent items which will be tagged so that telemat-
ics can notice them and thus control them. General estimates are that
there were 10 million sensors of all types connected to the Internet in
2007, and in excess of 3.5 billion by 2013. The total number of tele-
matic devices, including non-networked and non-self-supporting, could
exceed 10 trillion by 2025 and is expected to surpass 100 trillion by
2030. These sensors will provide a flow of real-time data whose size
exceeds most analysts’ imaginations. A new Internet protocol has just
been agreed to enable these huge numbers of telematics to be individu-
ally labeled. Generation 6 Internet is being built to deal with the mind-
boggling data flows.
The widespread use of telematics has huge implications. The obvi-
ous implication is examples like: if every grocery item is tagged/linked,
your fridge can record what is used and help the cupboard create a
shopping list; your car can tell your insurer’s computer how you drive;
street lights can tell the data center when they fail, or turn on only
when your car tells them you are approaching; swarms of nanobots can
examine pipes; your doctor can embed a small chip which can analyze
your blood, warn you of dangerous trends, make changes to and adjust
your automatic injector, alert emergency services if you suffer a medical
emergency, give them medical details, with your location. It could even
arrange an autodrive ambulance so no human needs to be involved.
The less obvious implication is that up until about 2012, it was
assumed that Internet traffic would mainly involve human-created
activities, with Internet growth projections based on the estimates of
human activities and the percentage of the world population which was
connected. Now, it is realized that in the future‚ Internet traffic demand
 2  Key Technological Disruptors    
17

will soar primarily predominantly due to a link between things, rather

than between people. McKinsey (2015) estimates that the number of
Web-linked telematic devices by 2040 will run into hundreds of bil-
lions, implying that human-sourced Web traffic will be less than 5% of
all Web traffic. Forbes (2014) found that 50–70% of home owners will
consider buying smart home devices in the next five years.
The interconnection of telematic devices, the exponential increase
in data inflow, and active analysis of the data received will be trans-
formative in nearly all industries, and thus has huge implications for
businesses. The most powerful use of telematic data will involve using
two-way feedback to predict failures before they occur, and to fine-tune
performance. There will have to be a large investment in capturing,
sorting, and analyzing the immense data flow, as well as acting on the
resultant information. The amount of work required to effectively use
the ever-increasing inflow of data will be a major issue for businesses,
and the survivors will be those who solve this problem. Businesses will
have to learn how to use the data to gain competitive advantage.
The implication for insurance is that all insured objects will give con-
stant real-time feedback to the insurer’s data center. Examples would be
the minute-by-minute driving history of every driver or the minute-by-
minute blood pressure, blood chemistry and heartbeat of every patient.
Chips linked to patient can report back to the doctor’s computer, which
can call an ambulance if the metrics indicate a need before a problem
actually occurs. The data from multiple clients can then be compared,
for example, the driver behavior and car reaction of nearly every driver
on the road, in every traffic environment, or changes in body chemis-
try which leads to illness, can be compared and trends found. Seriously,
ill patients could carry an intelligent medicine injector, which can get
feedback from a specialist. Drivers can be alerted before their cars break
down while drivers who are driving erratically can have their cars turned
off or controlled remotely. A house can identify if a stove is overheating,
or can tell police where all stolen items are now located.
In insurance terms, we are currently in the third wave of telematics.
The first wave was beta experiments by Norwich Union and Progressive
in using black boxes for collecting automobile insurance data. This
saw battles over regulations and intellectual data. With these settled,
18    
M. Naylor

the second wave trial programs were expanded to a number of other

insurers, and real-time networked data explored. The current wave of
innovation will involve telematic use in Auto insurance becoming rou-
tine, especially in commercial fleets, and an expansion of trials of telem-
atic use into other insurance types, with a focus on health.
It is important to note that the limited data flow we currently receive
from the limited number of telematics installed is poorly used. This is
because up until recently companies have struggled to store the flood of
data. They are thus typically analyzing only a fraction of it. McKinsey
(2015), for example, estimates that we are currently using less than 1%
of all currently available telematic data. An example they cite is that
a typical offshore oil rig now has about 30,000 sensors, but less than
1% of the information received is ever used, with only about 300 sen-
sors actually monitored. Modern cars can have 1000 sensors, data from
which is mostly ignored, and none gathered centrally in real time. Very
few insurers have setup systems to access data from these sensors. We
are also only using telematic data for anomaly detection and control sys-
tems, instead of taking advantage of its use in areas like optimization
and prediction. The major reason for this limited use of telematic data
has been the substantial costs and difficulties of storing and analyzing it.
As will be explored below, these limitations have now been removed.

Data Storage and Cloud Computing

The arrival of mobile devices meant that IT users had to stop being tied
to the memory of a single device, otherwise they could not move seam-
lessly between devices; accessing email on the move is difficult if emails
are stored on a PC hard drive. The introduction of ‘every-where Wi-Fi’
has allowed IT to escape physical confines. A necessary part of mobility
has thus been the creation of huge external ‘cloud-based’ data storage
facilities.
This centralizing and scaling up of data storage has had the side effect
of allowing drastic continuing drops in prices, and this has reinforced
the trend toward cloud storage. For example, Amazon’s costs of stor-
ing cloud data has been falling by about 50% every three years since
 2  Key Technological Disruptors    
19

2006, and costs are forecasted to drop even faster as the scale increases
exponentially. PWC (2015) states that the cost of storing a gigabyte
of data in the cloud dropped from 25 cents in 2010 to 0.024 cents in
2014.
Cloud computing has also meant that internationalization of services
can occur without the customer being aware of it, as nobody is aware of
where their data is stored, who is managing the service, or who is ana-
lyzing that data. This has allowed the use of cheaper data storage loca-
tions, e.g., those with lower cost electricity or cold external air. It also
allows secure multiple backups.
This near-zero marginal cost of data storage is essential as insurance
companies who use telematics will find their need for data storage sky-
rocketing. Cloud computing is thus vital as it means that insurers (i)
will have no need to run their own data servers, (ii) will find that the
cost of storing the huge data quantities created by the other trends man-
ageable, and (iii) data storage is now scalable, with insurers only rent-
ing storage as they need it. Cloud computing also allows its interlinking
with mobile devices and the internationalization of data banks.
The cost of local data storage has also plummeted. In the year 2000,
one gigabyte of hard drive space costs $44, by 2012 it had dropped to 7
cents, and by 2016 to less than 1 cent. In 2000, it costs $193 per giga-
byte to stream video, by 2010 it had dropped to 3 cents, and by 2016
to fractions of a cent. The amount of data able to be transmitted on an
optical network has also been doubling every nine months for the last
decade. The marginal cost of most data storage and transmission is thus
dropping close to zero.

Big Data
Traditionally, mathematics has assumed that data on ‘everything’ is
not available, so we need to make do with a limited sample of data.
Statistical theory, surveys, and nearly everything actuaries learn is thus
based on assessing how closely this limited sample matches the entire
population. This theory will, however, soon be archaic as the internet of
things and cloud computing means that now we can very cheaply collect
20    
M. Naylor

all the data. The world’s production of data grew 2000-fold between
2000 and 2012 and is expected to double at least every 10–18 months
for the foreseeable future. Approximately, half of all the data the world
has created in its history has been created in just the last 10 months. In
2015, data was growing by 2.5 quintillion bytes a day (18 zeros) or 8
zettabytes per year (21 zeros). The use of real-time telematic data could
soon make these forecasts of data increase underestimates, with data
probably doubling every six months. Notice that this is an exponential
curve.
The issue for traditional data analysis method is that the size of the
data flow is so large that no person can usefully access it directly. A new
mathematics of big data is consequentially evolving which uses algo-
rithms and software to automatically analyze the database for trends,
and then either summarizes it‚ or alerts an operator to preset changes,
or uses it to predict adverse patterns. Examples would be manage-
ment alerted if a truck driver started driving erratically, or if the failure
rate of a particular type of part increased, or if those who buy alcohol
at 1 a.m. have more heart attacks. Patterns of data which predict house
fires or link types of food purchase to types of sickness‚ can be discov-
ered. The extent of these huge data sets has had the side effect of provid-
ing large enough training sets for software algorithms to learn to handle
non-routine cognitive tasks, hence starting the process of creating the
Artificial Intelligence (AI) software required for complex tasks.
The efficiency gains possible from effective analysis of big data, com-
bined with telematics and AI software, are huge, with McKinsey (2015)
estimating a gain of 1% additional growth in GDP per year for the next
decade. They argue that the potential profit from use of business to
business telematics is three or four times the size of consumer telemat-
ics. An example would be a global oil giant, which could link sensors at
petrol stations, which give real-time information on types of fuel sold,
to refineries - to adjust the mix produced, to trucks - to adjust deliv-
eries, to oil rig - to adjust volume pumped. Sensors would alert staff
when things are about to fail, so they can be replaced. All this can be
done with no human making a decision, only supervising. It has been
estimated that this style of change could improve the efficiency of most
sectors by 100–200%.
 2  Key Technological Disruptors    
21

Big data is analyzed using three different methods:

Machine learning: The process of training computers how to ‘learn’
interesting patterns in data and how to adapt automatically when data
changes over time. With machine learning, algorithms are trained for a
time in order to build models, which are then used to direct automated
marketing and operational decisions, adapting as more data is gathered
once the AI is ‘live.’ This learning can be supervised or unsupervised.
Supervised learning: Supervised learning algorithms take independ-
ent variable inputs‚ such as user clickstream data and dependent variable
outputs like purchase data‚ and create a model for how they relate to
each other. Once it fits, it can be used as a predictive tool.
Unsupervised learning: This is also called ‘data mining.’ It is unsuper-
vised learning which looks for patterns and relationships within the data
itself, without concern for outputs. The focus is on its useful for seg-
mentation and customer insights.
The biggest current barrier to use of big data analysis is that universi-
ties are only now starting to create postgraduate courses in big data, so
skilled staff will be in short supply for at least the next decade or so.
Since most of the IT revolution depends on these data analysts, this is
the major limiting factor to the speed of transformation. The relative
unattractiveness of insurance companies to these graduates could be a
major barrier, which the insurer able to attract the best staff likely to
gain an unbeatable competitive advantage.

Digital Natives and the No-Wait Generation

There is a generational shift in consumer attitudes taking place. GenRe
(2014) argues that the millennial population cohort group (now aged
18–34) will represent over 50% of the global workforce by 2017, and
over 75% of the global workforce by 2025, and their spending power
will soon surpass baby boomers (now aged 52–70). Millennials were
raised with the idea that the first place to learn something, share some-
thing, or buy something is the Internet.
The generation born after the millennials have used laptops from
young, so are ‘born digital’, and use their tablet as their primary contact
22    
M. Naylor

with the world. They have played with film-quality digital games from
an early age. They will increasingly grow up in a house where a talkbox
controls house functions, so that they expect to talk to software. Their
expectations about interaction with companies are substantially differ-
ent to that of baby boomers and exceed the already high expectations of
the millennials.
Zaptitude (2016) found that 70% of millennials surveyed feel dis-
connected from financial service providers because those providers do
not package products or services in a manner they feel comfortable
with. Zaptitude argues that millennials feel that financial service provid-
ers ‘speak in an alien language, which makes the millennial feel invis-
ible.’ Readers can imagine how digital natives will respond - Facebook
and Web sites are old techs and scanning documents is dinosaur age.
They will expect to interact verbally with an insurer’s AI system, getting
instant answers. An insurer who thinks ‘being digital’ is adding a Web
page will suffer the fate of the dinosaurs.
Bain (2015) found that even among the non-digital natives, the pro-
portion that plan to use a Web channel for buying insurance will rise
from 35 to 79% over the next two years. They also plan to interact with
insurers in quite different ways. For business, a key aspect is that these
consumers will increasingly interact with Web sites via social networks
and will increasingly review the goods and services provided. Nearly all
purchases are first researched online, including service quality.
McKinsey (2012) found that 80% of users interact with social net-
works at least 5 times a week. They also find that the speed and scale of
adoption of social technology by consumers have exceeded that of pre-
vious technologies, yet estimate that only about 5% of possible uses of
social networks have yet been discovered.
Shirky (2010) argued that social networks are a unique invention for
business as they encourage consumers to cooperate together to create
rich new forms of interaction. This collective effort feeds back in a way
which ensures that individuals work together in a way that they could
not achieve alone. This ensures that the joint output exceeds individual
efforts by a multiple. Businesses cannot control this, only encourage it.
People now rely on their social networks for everything from advice
on movies, on relationships, on product reviews to decide what is
 2  Key Technological Disruptors    
23

interesting. The use of social networks is thus transformational and has

changed how people think about customer relationships. The implica-
tions of this generational change in favor of social networks for busi-
nesses are huge. Company Web sites will have to do more than act as
one-way portals for corporate communications, product promotions,
customer services, and e-commerce, and instead become interactive
two-way communication channels and lock into social activities.
While these new generations value face-to-face contact with provid-
ers, their base expectation is an extremely high level of Internet and
personal service, and a seamless integrated experience. Their primary
contact with business will be Web-based, and their expectations about
the Website will be extreme. They will expect your network to interact
with their devices automatically and your Web page to integrate with
your customer service. Every possible device must be supported. Your
company will have to integrate into online social networks and from
within the ecosystem, provide useful and active feedback to enhance
lifestyles. Quotes, service, and feedback must occur within minutes.
In particular, companies have to respond to feedback in a two-way
conversation, rather than treating the Web as a push mechanism for
the company’s views only. Complaints must be responded to within an
hour and responded to publicly. This can be very positive if handled in
a way which the crowd approves off; mistakes are accepted, and it is the
response to mistakes which increases or decreases reputations. The abil-
ity to enhance ratings via customer feedback then becomes a sales tool
vital for the new generations. The ability to respond to public feedback
from customers in a manner which enhances reputation is a skill most
insurers have yet to master.
If your business doesn’t meet these expectations, you may as well not
exist. For example, these generations won’t phone; they expect to Skype
via your Web page. Expectations around the visual presentation of your
Web service are extreme, similar to an advanced-level digital game plat-
form. Film-quality video tutorials or news blogs are expected. Very few
current insurers even come close to these expectations.
Digital natives are correctly labeled as the ‘no-wait generation’. The
essence of this is that consumer expectations around level of customer
service are raising rapidly, with declining patience for any delays in
24    
M. Naylor

response or administrative errors. Phone centers or text messaging

which takes 30 seconds to answer are regarded as archaic. Satisfactory
customer feedback and respond to complaints/issues is expected in
terms of hours rather than days. A question must be asked only once,
and then be available to all other staff, regardless of platform. These
extreme response attitudes are rapidly migrating to other generations.
Only a keyword software-based AI response system can respond with
this degree of rapidity and scale to wide fluctuations in demand. Call
centers will always struggle.
Millennials also eagerly swap consumer experiences online‚ with over
70% discussing possible purchases with friends online and 75% brows-
ing online customer feedback. Unmet expectations will be immediately
discussed online. The main forums are social networks especially video
outlets, as fewer than 25% watch more than 10 hours of TV a week,
and very few read print newspapers. Insurer reputations can be irretriev-
able damaged within a week, meaning that insurers will not have time
to fully discuss response strategy, and instead must have protocols pre-
arranged. It is surprising that in 2017 insurers still regard setting up a
Facebook site as leading edge.
It is also a mistake to assume that the older baby boomer clients
aren’t also technologically literate, don’t desire digital contact opportu-
nities, and don’t have changing expectations. CEB (2013) found that
the wealthier segment of the 55+ age population has high technologi-
cal expectations, and that wealth management company customer rela-
tionship management (CRM) systems are persistently lagging behind
the dynamics of older customer preferences. Baby boomers do, however,
have lower response expectations.
In a fast-paced world where bad experiences are immediately
uploaded, insurers will either evolve excellent customer service or die.
These extreme expectations and tendency to communally review experi-
ences online mean that social capital is now the key currency for com-
panies. Mistakes in customer response are accepted, but the company
response to issues must be excellent. Because of changing generations,
firms which do not engage the way digital natives expect will be out of
business by 2025.
 2  Key Technological Disruptors    
25

Software, not Hardware

Some experts have predicted future disruption being based on robots.
This is unlikely to be true. The initial drivers of the IT revolution were
indeed physical; e.g., reduction in transistor size, worldwide optical
fibers, and mobile devices. While these advances will continue, their
impact will lessen hereafter, as advances in physical technology have
opened up the space for software innovation.
A reason is that in a number of key areas, like automated cars or
general-purpose robots, we have reached the stage where the key fac-
tor inhibiting most progress is now software limitations rather than
hardware. A second reason is that current redesign of computer chips
to incorporate RAM will substantially enhance speed. A third reason is
that computer speed will not be an issue, with a predication that the
first useable quantum computer will be available by 2020 and in general
use by 2030 and will hundreds of times faster than today’s best super-
computer. The result of this is that past limitations due to physical pro-
cessing capacity will disappear. Analyzing huge data sets will not be a
problem.
The implication of this is that solving the software issues will lead to
exploding uses for physical technology. From that point in time soft-
ware innovation will be the key driver and will start to dominate. It has
been speculated by experts that so far, we have only discovered 1% of
all the possible uses of the Internet. Uses for software will explode as
fast we can code them. The more processing intensive uses like AI will
particularly explode. Insurers need to pay close attention to leading edge
software innovations in all sectors.

Artificial Intelligence and Complexity Science

Predicted improvement in the areas of artificial intelligence and com-
plex adaptive systems mean that software could, within 5 years, han-
dle problems regarded today as ‘impossible’. Traditionally, software has
been created using a large set of simple ‘yes/no’ rules. This required the
26    
M. Naylor

problem to be solved to be broken down into a long series of simple

steps. The data set was assumed to be stable and the relationships to be
unchanging. Forecasting systems were thus created by using ‘predictive
analytical systems’ which assumes fixed relationships between variables
based on prior data. This then postulates a single future by forecasting
this relationship forward. This meant that the range of problems which
could be solved was limited by (i) those areas where the data meet the
stability requirements, (ii) those areas which the programmer thor-
oughly understood prior to starting, and (iii) those areas which can be
solved by iteration within a reasonable time.
These restrictions are about to be largely removed by an alternative
approach, ‘complexity science’, which‚ in contrast‚ structures the prob-
lem by using a number of elements which interact with each other in a
dynamic network of ‘nodes’ whereby the weightings on links between
nodes change as the system learns. Outcomes are not just yes/no but a
complex array. The programmer does not have to understand how to
solve the issue or impose solution frameworks. Complex adaptive sys-
tems do not pre-specify logical sequences but create a system of rules,
which software ‘agents’ work within but are not constrained by. These
agents then examine data sets and look for patterns and learn as they go.
Therefore, there is a multiple of future paths dependent on exactly how
the elements evolve. There is no assumption that relationships between
elements are stable or that what happened in the past will reoccur. Nor
is there an assumption that all elements will react in the same way to
another variable. Diversity in response is allowed. A typical node net-
work is shown in Fig. 2.1.
Complex software programs can thus find patterns not apparent to
humans and the systems get better as the data sets increase in size and as
the number of runs increases. The software is thus ‘learning’ rather than
having answers preset. The key element is discovering how the elements
interact and how the relationships evolve over time. Important here is
finding paths which lead to stability rather than to chaos.
In 2007, Fei-Fei Li, the head of Stanford’s Artificial Intelligence Lab,
gave up trying to program computers to recognize objects and began
labeling the millions of raw images that a child might encounter by age
three and feeding them to computers. By being shown thousands and
 2  Key Technological Disruptors    
27

Fig. 2.1  A dynamic node network. Source Author

thousands of labeled data sets with instances of, say, a cat, the machine
could shape its own rules for deciding whether a particular set of digi-
tal pixels was, in fact, a cat. In 2015, Dr. Li’s team unveiled a program
that identifies the visual elements of any picture with a high degree of
accuracy. An image recognition challenge called ‘ImageNet’ achieved
their breakthrough in 2012, when software successfully recognized
85% of images by using AI methods. In the 2015 imaging challenge,
the winning team achieved 96%, surpassing human achievement. IBM’s
Watson machine relied on a similar self-generated scoring system among
hundreds of potential answers in its Jeopardy victory. During its Go vic-
tory, the AlphaGo AI, created a strategic move never seen before, which
involved substantial short-term sacrifice, for long-term strategic domina-
tion.2 One of the reasons for this move was that AlphaGo assessed that
it would surprise its human opponent so much that they would respond
poorly. That move was true innovation by an AI program.
Predictions of complex adaptive systems are thus dynamic and can
solve problems in areas where the relationship cannot be carefully

2https://www.technologyreview.com/s/602094/ais-language-problem/?mc_cid=c6c808bdae&mc_

eid=e1c76a51f6.
28    
M. Naylor

predefined. They are particularly useful in (i) finding patterns in data

which are too obscure to be visible to humans and (ii) finding pat-
terns in overwhelmingly large amounts of data. In particular, an area
of AI called ‘deep learning’, which involves systems finding patterns by
crunching big data sets, is now being widely exploited to power Internet
search engines, translate Web pages, recognize voices, and detect fraud.
When combined with the other disruptive technologies like big data,
telematics, cloud computing, etc., AI will be the key disruptive technol-
ogy for any industry which relies on data. Insurance, being based on
data, is particularly vulnerable.
Research into and application of complex adaptive systems was
restricted and limited up to 2005 by limited computer power and small
data sets. Progress was slow. It was overhyped and went into the disap-
pointment phrase.
Since 2012, however, the arrival of big data and corresponding com-
puting power able to run a huge number of scenarios has transformed
the field. Modern computers can run billions of scenarios to find the few
scenarios which offer optimal results in a way humans cannot match.
An important AI technique is called ‘artificial neutral networks’, which
has an input layer, an output layer, and a number of hidden processing
layers (see Fig. 2.2). Data is fed into the input layer and the problem is
Neural Network

Input Layer

Hidden
layers

Output layer

Fig. 2.2  Neutral Network. Source Author

 2  Key Technological Disruptors    
29

then separated into aspects and each hidden layer pathway attacks an
aspect of the problem. Each ‘node’ or ‘neuron’ tries to learn from past
mistakes by adjusting the weight it puts on a particular facet of the prob-
lem. By running a large data sample through enough layers, multiple
times, the system adjusts the weights until a sufficient level of accuracy is
obtained. This is called ‘training’ the system. Advanced systems can have
over 30 layers, which is only possible with very powerful computers.
Learning can be ‘supervised’, which involves labeled data and instruc-
tions of what to look for, or ‘unsupervised’, which involves unlabeled
data and allowing the program to find its own correlations. The latter is
particularly useful for insurers who want to discover new links between
claim rates and customer attributes, or to find undetected types of claim
fraud.
The more complex AI applications are still limited by current pro-
cessing power, but new chip advances will soon limit these areas to only
the largest problems. These deep searches can find relationships which
are complex in ways which could only be discovered once big data sets
are used. The bigger the data set, the more can be found, so the capacity
of these systems will improve on the exponential scale, as telematic data
becomes available.
The importance of very large data sets for training means that
Internet search companies have a nearly unsurpassable advantage in
creating very accurate AI systems. The only way insurers will be able
to compete is by joining an ecosystem with a firm which has access to
huge data sets, as well as expertise in the most advanced AI systems. The
newness of the field means that advanced level expertise is a major con-
straining issue, with insurers forced to compete for scarce Ph.Ds with
far more exciting employers. Charan (2015) argues that any organiza-
tion that is not already math and AI-based, or is unable to become one
soon, is already a legacy company. Insurers need to set up AI research
centers and retrain actuaries in AI and big data methodologies.
It needs to be noted that AI is really a portfolio of technologies, and
for this type of software to become autonomous a number of related
technologies need to be sufficiently advanced at the same time. Progress,
however, has not been even across all fields; with AI researchers making
30    
M. Naylor

much greater progress in some fields such as natural language process-

ing and data analysis, and far less progress in other fields of AI such as
decision-making and deductive reasoning. This is due to scientists in the
mid-to-late 2000s achieving a breakthrough in the way they thought
about neural networks, or the systems that allow AI to interpret data.
Along with the explosion of raw data made possible by the Internet, this
discovery allowed machine learning in related fields to take off at a near
exponential rate. Other fields of AI research, in contrast, are plodding
along merely at a linear pace.
This doesn’t mean scientists won’t make breakthroughs in lagging AI
fields, but it does mean that, for now, there are limits on fully auton-
omous software. This is vital, because in the meantime, the uneven
advances made are pushing programmers toward creating very specific
kinds of artificial intelligence. For instance, consumers are already see-
ing our machine-learning research reflected in the sudden explosion
of digital personal assistants like Siri, Alexa, and Google-Now—tech-
nologies that are very good at interpreting voice-based requests but
aren’t capable of much more than that. These ‘narrow AI’ have been
designed with a specific purpose in mind: to help people do the things
regular people do, whether it’s looking up the weather or sending a text
message.
Narrow, specialized AI is also what companies like IBM have been
pursuing. It includes, for example, algorithms to help radiologists pick
out tumors much more accurately by ‘learning’ all the cancer research
we’ve ever done and by ‘seeing’ millions of sample X-rays and MRIs.
These AI systems act much more like glorified calculators: They can
ingest way more data than a single person could hope to do with his
or her own brain, but they still operate within the confines of a specific
task like cancer diagnosis.
The limitations of this unbalanced AI progress are however minor, as
current IT capacity vastly exceeds what insurers currently use IT for, so
it will take a decade for insurers to reach any processing limitations and
by then AI advances will offer insurers virtually unlimited opportuni-
ties. Within a 5- to 10-year time frame, more even progress in AI will
result in products which will transform our lives. The expected pace of
 2  Key Technological Disruptors    
31

change is such that insurers who are not currently racing to understand
and implement current advances will soon be too far behind to ever
catch up.
This differential progress also has major implications as to what kinds
of tasks can be automated, and thus what job activities will be at risk of
being replaced by software. This means that loss of specific job activi-
ties to software will not occur all at once, but will occur one by one.
However, as this process will probably be exponential the impact on
jobs will seem like an increasing wave, giving workers and employers lit-
tle time to react to one change before the next change occurs.

Hyperscaling
BCG (2015) argues that within the digital revolution there have been
two prior secondary waves of disruption. The first wave was the dot-
com era, which was characterized by falling transaction costs altering
the trade-off between information richness and reach, so that suddenly
rich information could be broadcast widely and cheaply altering how
products were sold. Company value chains were open to attack by out-
siders. An example of this was Microsoft giving away Encarta to pro-
mote PC sales and incidentally destroying Encyclopedia Britannica.
The second wave was the explosion of Web businesses, which was
characterized by the disappearance of scales of economy for many
products. This allowed long production tails and collaborate pro-
duction, called ‘economics of community’. An example was Encarta
being replaced by Wikipedia. Successful companies actively adapted
these waves. Unsuccessful companies tried to restrict the trend and
disappeared.
BCG argues that we are now entering a third wave, which is char-
acterized by what they call ‘hyperscaling’. This involves creating huge
networks based on a common AI system because of the importance of
obtaining as wide a range of data as possible to train complex systems.
This means that competitive scale will be beyond any individual busi-
ness and gives an unbeatable competitive advantage to businesses who
can cooperate with other related businesses to standardize and share
32    
M. Naylor

data and systems. Data has to be regarded as infrastructure, and the col-
lection and use of huge data sets across a business ecosystem will be a
defining characteristic of insurance survivors.
Note that it is the combination of Wi-fi-linked devices, big data,
cloud computing, and advanced software, which is now enabling pro-
gress at a far faster rate than the previously the previous stand-alone sys-
tems could achieve. This third wave will allow us to harvest the fruit of
computerized administration systems in a way which has so far not been
possible. Insurers have to make strategy decisions increasingly based on
data access.

Voice and Visual Recognition

Hard things for humans - like calculus, financial market strategy, and
language translation—are easy for a computer, while easy things for
humans - like vision, motion, movement, and perception - are hard for
a computer.
Voice recognition, for example, is hard for computers, and accurate
recognition took far longer than analysts predicted. A key problem was
that 95% or even 98% accuracy was not considered good enough by
consumers. The largest initial hindrance was lack of large data sets in
multiple dialects and of everyday speech. A key breakthrough was the
release of the Justice Department’s store of millions of Enron emails
which gave analysts examples of real-world communication. This data
set enabled the creation of respectable voice recognition and translation
systems and the linking of these to central data stores is now generating
huge data sets for algorithms to work on.
The ah-ha moment, however, came in 2011 when IBM’s computer
beat the two top all-time champions on the game show Jeopardy. This is
impressive as Jeopardy is a complex verbal game requiring a deep under-
standing of verbal word plays and complex spoken language with hid-
den meanings, as well as learning from past mistakes. Involvement in
Jeopardy was thus long considered a complex task impossible for a com-
puter. These achievements mean that voice recognition has passed the
breakthrough stage, the point in the graph where exponential techno-
logical progress breaks through linear expectations.
 2  Key Technological Disruptors    
33

The implications of accurate voice recognition are large. Already

most large corporations have either replaced telephone receptionists
with software or are considering it. While Siri was initially crude, it has
recently become workable. As frustrating as some of these current sys-
tems are, improvements at an exponential scale indicate that your next
receptionist will probably be software. Call center operators will follow.
In the future, most interactions with software can be expected to be
voice-focused.
Similarly, visual recognition is at the breakthrough stage, so that face
recognition systems will soon be routine to handle Skype calls and visi-
tor reception. Nearly, all the physical engineering issues which hindered
visual recognition are solved, meaning that progress from this point is
software-related. Autonomous cars can now meet 99% of the visual
demands on them to recognize obstacles, and within 5 years will match
or exceed the capacity of human vision, especially in situations where
spectrums like infrared are more useful, for example, at night or during
storms. Recently, an app was released which allows a person to take a
photograph of a stranger on their mobile and then instantly search the
Web for information on that person, with about an 80% success rate at
establishing their identity.
Advances in animation are linked to this and are also disruptive.
Initial work on animated characters revealed that while humans were
happy to accept a 90% accurate human animation as a cute copy, they
strongly rejected any animation which was more than 90% but less
than 100%, as it became ‘creepy’. This is called the ‘valley of death’ or
‘uncanny valley’, in the industry, as revealed by the adverse reaction to
the movie Polar Express. Work by movie companies like Weta have,
however, since advanced animation to the level where animated heads
are now accepted by consumers.
Soul Machines is working on producing talking heads for chatbots
which are eerily realistic. These have expressive faces which convey emo-
tion in line with the spoken content - so it can express surprise or frown
when asking customers about uninsured areas. They can also read cus-
tomer emotions via the Webcam or smart phone app. Current versions
seem stilted, but it would be an unwise insurer who assumed that, given
exponential rates of improvement, these systems will not dominate
future CRM systems.
34    
M. Naylor

For companies, the key issue is not the ability of animated heads
to perform in a movie setting where thousands of hours of program-
mer time can be invested on a limited preset sequence of moves, but
the ability of an animated head to respond in an ad hoc manner to a
diverse range of customer queries. While this area is less advanced than
voice recognition, within 5–10 years animated heads will be available to
answer simple Skype queries and redirect more complex ones. Within
10–20 years, nearly all customer queries and advice, whether physical or
online, will probably be handled by animations.
Insurers need to be aware that once combined with image recogni-
tion and voice recognition, AI systems are currently being designed
which will be able to answer 90% of all text, phone, or skype queries
without the customer being aware that they are not talking to a human.
These systems are able to recognize which questions or situations require
them to refer to a human supervisor.
There is a strong tendency for humans to respond to chatbots as if they
were humans and therefore expect an appropriate human response. A
difficult area which is currently being focused on is the ability to handle
jargon, irony, sarcasm, and exaggeration. For example, if a customer says
‘I’ve been waiting on the claim payment forever’, AI systems can work out
that they just mean a very long time. AI systems can be programmed to
make jokes and to simulate ‘social intelligence’, especially tact, wit, and
charm. Apple has put a lot of effort into making Siri likable, recognizing
that occasionally this requires imperfections or kookiness.
Importantly, these systems will soon be able to recognize emotions
and provide appropriate emotional responses. The increasing ability of
software to pick up human facial clues on Skype calls will make these
systems seem emphatic, with customers finding it very hard to tell if a
human is speaking to them.
Research with psychology chatbots has shown that if the right style of
questions or responses is programmed in, humans trust these digital agents
and provide intimate financial or personal information. There is some
evidence that humans prefer admitting failing to a digital agent in prefer-
ence to a human. Given that these systems will collect data on customer
reactions, insurers will than be able to trial different emotional or facial
 2  Key Technological Disruptors    
35

reactions to customers and see which generate the highest probability of a

desired customer response. Differing facial types can also be trialled.
An important aspect of this is that software can be scaled to large
or small demand, at very low marginal costs, and can be fine-tuned to
accurately follow company policy. The software will never get bored or
irritated or take leave and can be scaled infinitely if demand soars due to
an emergency.

Digital Agents
One of the basic problems for users of the Internet is how to cope
with the overwhelming size of choices. Every year, 8 million new songs
are produced, so how does a consumer decide which few to listen
to? If a keyword search reveals 3 million possible Web pages, how to
rank the top 10? Platforms like Google use algorithms based on your
own past Web searches and match these to what people with similar
search histories clicked on next. Because this ‘people-like-you’ recom-
mendation so closely matches your preferences, these have proved to
have very high click-through and sales conversion rates. Relying on
platform algorithms, however, places power in the hands of the plat-
form. They control what you see next and what you buy. They also
control the balance between repeating past patterns and experienc-
ing new ideas/products outside your immediate comfort zone. This is
‘people-like-you-generally-dislike-these-but-learned-to-like-this-one’.
The leading edge of Web users is therefore increasingly interacting
with the digital world via the use of AI ‘digital agents’. These are pro-
grams which contain useful personal information and carry out Internet
tasks on behalf of its owner, and search the Web widely based on your
preferences, rather than what a platform allows you to see. They can be
customized to look past the answers which platform algorithms provide.
They are increasingly being used to search for bargains, arrange activi-
ties, and autonomously make simple payments. They will liase with
other household systems and create shopping lists, and keep track of
appointments. For example, they will be aware of upcoming dinner par-
ties, check the fridge contents, and arrange buying, book services, and
36    
M. Naylor

ensure other attendee’s digital agents are aware of the event. They can
thus be visualized like a digital butler.
The agents are frequently voice-activated and likely to be the heart of
future human contact with computers. Increasingly, they will use ani-
mated talking heads to communicate with their owners. Within insur-
ance, these agents will offer financial/insurance planning, search for the
best price/condition deals, and arrange cover via communication with
insurer Web sites. Customers will set the buying parameters and then
let the agent explore the Web for the correct product. Because they
will have no time restrictions they will explore the Internet more thor-
oughly than customers would. Over time, the agent will increasingly
learn their owner’s preferences and may suggest interesting new prod-
ucts or services. An as yet undefined area is what weighting these digital
agents will place on brand or reputation when making buying decisions.
Customers can give their agent feedback about the product, which the
agent will then upload to product review Web sites. Therefore, product
review may become increasingly agent rather than human based.
Insurers will have to interact with these agents, computer to agents,
to ensure their product ranks high within the agent’s selection algo-
rithm. This will be essential for dynamics insurance which involves
a series of micropayments. Minor insurance claims will, in general,
be handled by these agents. Insurers who do not have a system which
interacts with these agents, will not be in the market.

Block-Chain Payment Systems

Payment over the Internet has always been an issue as the identity of
the next party in the chain has to be verified. This has meant a need for
third-party vendors who can guarantee each payment. This imposes cost
and has hindered the development of micropayment systems.
Block-chain systems involve the use of a sequence of encrypted
ordered records, each with a timestamp and link to the previous step.
These are highly resistant to modification and thus the sequence can
record transactions between two parties in a verified, permanent, and
low-cost manner. This substantially removes the need for third-party
 2  Key Technological Disruptors    
37

payment systems and will be a vital part of real-time insurance as it will

allow near-zero-cost transactions. This is a vital part of smart admin sys-
tems. Outside of that role, however, block-chain technology will be less
important as a disruptor for insurance than it will be for finance.

Robotics
While current factories are becoming highly mechanized, this does not
generally involve ‘robotics’ or mobile machines. The key issue is that
current autonomous machinery is highly specialized and only able to
adapt to a restricted range of uses, and because they have limited aware-
ness of humans, for safety reasons they have been kept within set areas.
Future use will involve unspecialized machines equipped with artifi-
cial intelligence (AI), which will be capable of learning multiple jobs,
improving from experiences and mistakes, and because of multiple sen-
sors will be safe enough to move among the general public. Progress
toward this has been slow, however, with the Darpa Robotic Challenge
not showing the same rate of change as autonomous car challenge
exhibited.
This lack of momentum is about to change, however, especially once
issues around unstructured environments, sequence planning, and tac-
tile response are solved. Substantial progress can thus be expected in
the future, especially independently mobile machines. Most of the key
technology issues have been solved, from vision to manual dexterity to
balance. One of the key reasons for the current slow progress is that the
integration of the various robotic components depended on software
developments, and these are only just being developed. This makes cur-
rent humanoids jerky and not very useful. The creation of realistic robot
faces is well behind that achieved in animation.
In the immediate future, the high cost of initial models will restrict
their use to dangerous or extreme situations. Within a 10-to 20-year
framework, however, autonomous machines which are truly adaptable
and can be safely used beside humans will be developed and will fall
rapidly in price. They will become common in areas where staff is hard
to find, like elderly health care or food preparation. The best estimates
38    
M. Naylor

of the employment impact of moveable robotics are the loss of around

15% of all jobs, mostly in factories. For insurance administration, the
impact will be minimal; software and AI are likely to be a lot more
important.
Insuring the value of these robots is conversely likely to be a grow-
ing area in the medium term. MGI (2013) found that the market for
household cleaning robots is currently growing at more than 20% per
annum, though from a very small base. Most middle-income house-
holds will eventually use a robot for nearly all household jobs which
require mobility, though this is more likely in a 20-year time frame.
The market for general purpose robots is potentially larger than the car
market and is drawing in a lot of serious investment. However, because
these robots will be equipped with location sensors and control over
the Internet, their theft will be unviable, and as they get more reliable‚
claims for damage will become rare, so insurance revenue will be low.
Issues relating to data security and hacking are sources of revenue.

3D Printing
3D printing has since 2000 evolved from an idea to a hobby to a reality,
with printing now using a range of materials, including metals and bio-
material. Commentators have disagreed on the ultimate cost viability of
printing industrial components routinely, but development on an expo-
nential scale has substantial implications, for example, a drastic decrease
in the need to trade internationally. While this technology will have
limited application to the insurance industry, it will impact via the abil-
ity of suppliers to produce whatever parts are needed, rather than main-
tain large parts stores. For example, once repair shops or customers have
access to high-quality 3D printers, then car insurance firms only need
to supply repair shops with the software for a part, rather than the part.
SwissRe (2016) points out that 3D printing is starting to create a
range of insurance issues, e.g., are replacement parts being made to a
good quality, do printed medical implants create more or less risk, are
there differences in user instructions, and are there potential malicious
hacking issues?
 2  Key Technological Disruptors    
39

Energy and Material De-intensification

People have a limited capacity to consume most types of goods, that
is, most physical goods have a ‘low-income elasticity’. The main reason
for the severe shrinking of agriculture from 90% of production before
the industrial revolution to 2% today was that people did not buy
much more food as they got wealthier, so rapid increases in agricultural
productivity caused rapid decreases in agricultural employment. The
same has occurred with industrial goods, where the share of industrial
employment peaked in most Western countries by 1980. The physical
products we do buy contain less content, for example, beer cans weigh
80% less in 2016 than they did in 1973. In the future, cars will be com-
puters on wheels. Services, in contrast, tend to have high income elastic-
ity, so that people buy more services as their incomes rise, so that service
employment has expanded.
We are thus entering a ‘post-material’ world, where any increase in
income will result in more spending on software and entertainment,
rather than physical goods. This, together with the increasing efficiency
of modern manufacturing, will mean that consumption of energy and
materials have just about peaked in the developed world with a decline
probable. Emerging markets will follow. Demand for insurance for
physical products, contents insurance, will thus rise slower than GDP.
If they want to retain their cash flow, insurers have to increase their rev-
enue from non-risk insurance-related services.

References
Bain & Company. (2015). Global Digital Insurance Benchmarking Report 2015:
Pathways to success in a digital world.
Boston Consulting Group. (2015, July 24). Navigating a world of digital dis-
ruption. BCG Perspectives.
CEB. (2013). TowerGroup wealth management client experience survey.
Charan, R. (2015). The attacker’s advantage: Turning uncertainty into break-
through opportunities. Public Affairs, February.
Forbes. (2014). Internet of things by the numbers: Market estimates and
forecasts.
40    
M. Naylor

GenRe. (2014). Marketing to the millennial mindset. GenRe USA.

McKinsey. (2012). The social economy: Unlocking value and productivity through
social technologies: consumer financial services. McKinsey Global Institute.
McKinsey. (2015). The internet of things: Mapping the value beyond the hype.
McKinsey Global Institute.
MGI. (2013). Disruptive technologies: Advances that will transform life, business
and the global economy, Technical Report. McKinsey Global Institute.
Morgan-Stanley/BCG. (2013). Wearable devices: The internet of things becomes
personal. Boston, MA: Boston Consulting Group.
PWC. (2015). Insurance 2020 & beyond: Necessity is the mother of reinvention.
Shirky, C. (2010). Cognitive surplus: Creativity and generosity in a connected age.
New York: Penguin Press.
SwissRe. (2016). How 3D printing will reshape the insurance landscape, Centre
for Global Dialogue, New York.
Talla, V., Kellogg, B., Ransford, B., Naderiparizi, S., Gollakota, S., & Smith, J.
(2015). Powering the next billion devices with Wi-Fi. ArXiv:1505.06815.
Zaptitude. (2016). Investing in millennials. Research Report. www.zaptitude.
com.
 3
Types of Insurance

House & Contents Insurance

Property insurance insures your home against the risk of fire or other ­damage.
Most home insurance provides cover only up to the ‘sum insured’—a
capped amount, that is, the limit of what can be claimed. Home owners
with mortgages are normally legally required to insure their property against
loss or injury to others. It also provides some ‘third party’ cover if we damage
someone else’s belongings in the house we are living in.

Contents Insurance protects personal belongings, both in and out of the

house. Tenant’s/Renter’s insurance applies only to those who rent their
home.

Automobile Insurance
Automobile insurance will pay to repair or replace your automobile.
‘Comprehensive’ motor vehicle insurance is the most common and
it covers loss, theft, or damage to a vehicle. It also covers accidental

© The Author(s) 2017 41

M. Naylor, Insurance Transformed, Palgrave Studies in Financial Services Technology,
DOI 10.1007/978-3-319-63835-5_3
42    
M. Naylor

damage to the other car or property you damage, or liability risk.

Cheaper options are ‘third party’ and ‘third party, fire and theft’ cover.
Third party insurance covers for damage to another person’s vehicle or
property, but not yours. Extending third party insurance to fire and
theft covers the risk of your car being destroyed by fire or stolen, too.
The cost of insurance will vary depending on your age, claims his-
tory, the level of excess you are prepared to take, and the make and
type of vehicle. Most countries make some type of auto-insurance
compulsory.

Personal Insurance
Life insurance provides a lump sum of money on death. In some cases,
a portion or the entire ‘sum insured’ is paid out before death if you are
diagnosed with a terminal illness. There are different types of life insur-
ance cover. The most common one today is term life insurance, which
covers you for a fixed number of years such as the length of our mort-
gage. Investment ‘whole of life insurance’ has no maturity and is popu-
lar as an investment in countries where it is tax advantaged.

Trauma insurance (also called critical illness) provides a lump sum if you
suffer from certain illnesses or injuries such as cancer, heart disease, or
paralysis.

Income protection insurance pays a percentage of your income on an

ongoing basis if you suffer from named illnesses.

Disability insurance pays out a lump sum for permanent disablement

through sickness or accident.

Medical insurance covers private hospital and other medical bills in case
of sickness.
 3  Types of Insurance    
43

Travel Insurance
Travel insurance policies cover belongings against loss or theft, extra
costs if your flights are canceled, and medical treatment for accident or
illness while traveling. Travel insurance policies usually have exclusions
for preexisting conditions and unattended baggage. There are often
many conditions and exclusions in travel insurance policies.

Business Insurance
Commercial property insurance policies provide for property having a
business use. In addition to providing coverage for loss, damage, and
liability issues, on both the premises and contents, business owners buy
protection for the indirect loss of business costs associated with having
to suspend operations while recovering from an incident.

Business interruption/loss of earning insurance covers expenses incurred if

the business is unable to operate.

Business owner’s policy (BOP) packages all required coverage a business

owner would need. Often, BOP’s will include business interruption
insurance, property insurance, vehicle coverage, liability insurance, and
crime insurance. This can be customized to a company’s specific needs.
Typically, a business owner will save money by choosing a BOP because
the bundle of services often costs less than the total cost of all the indi-
vidual coverages.

Commercial Auto-Insurance protects a company’s vehicles. If a business

does not have company vehicles, but employees drive their own cars on
company business this provides ‘non-owned auto liability’ to protect the
company in case the employee does not have insurance or has inade-
quate coverage.

Worker’s Compensation provides insurance to employees who are

injured on the job, providing wage replacement and medical benefits.
44    
M. Naylor

In exchange for these benefits, the employee gives up his/her rights

to sue the employer for the incident. This protects the company from
legal complications. Some countries require companies to have this
insurance.

Liability Insurance
Business Liability Insurance provides protection when the policy holder
is financially responsible for injury or damage they cause to others. The
policy provides both legal defense costs and damages if employees or
products or services cause bodily injury or property damage to a third
party. This includes area like accidental pollution.

Professional Liability Insurance (Errors and Omissions Insurance) pro-

vides defense costs and damages for failure to or improperly rendering
professional services. Professional liability insurance is applicable for any
professional firm.

Directors and Officers Insurance protects the directors and officers of a

company against their actions that affect the profitability or operations
of the company. If a director or officer of the company, as a direct result
of their actions on the job, finds himself or herself in a legal situation,
this type of insurance can cover costs or damages lost as a result of a
lawsuit.

Data Breach Insurance provides cover for accidental or unauthorized

breaches of data. If the business stores sensitive or nonpublic informa-
tion about employees or clients on their computers, servers, or in paper
files they are responsible for protecting that information.

Medical Malpractice Insurance provides covers for medical institutions

and personal.
 3  Types of Insurance    
45

Marine/Shipping Insurance
Marine insurance covers the loss or damage of ships, cargo, terminals,
and any transport or cargo by which property is transferred, acquired,
or held between the points of origin and final destination. Cargo
­insurance is a sub-branch of marine insurance, though Marine also
includes onshore and offshore exposed property, hull, marine casualty,
and liability. Marine open cover is generally open-ended with no expiry
date and can be canceled without prior notice.

Shipping insurance covers goods transported by mail or courier.

Depending on transportation method, it can be divided into marine
cargo transportation insurance, continental cargo transportation insur-
ance, aviation cargo transportation insurance, and parcel transportation
insurance.
There are three internationally-used transport insurance categories:
(i) Free of particular average (FPA) is the narrowest form of cover as
the insurance company does not cover you for partial loss or damage
to the cargo. You are only covered if the entire consignment is lost or
damaged, for example, when a ship sinks, or by fire. (ii) With average
(WA) cover extends the FPA clause to include partial loss arising from
heavy weather and sea water damage. Both FPA and WA can usually
be extended to include protection from theft and pilferage. (iii) All
risks (AR) cover is the most comprehensive cargo insurance, provid-
ing protection against loss or damage from external causes. It does not
cover loss or damage arising from delay, inherent vice (deterioration or
­damage without outside help), losses as a result of inadequate packag-
ing, weight loss from drying out, or market changes. Risks from war,
strikes, riots, and civil unrest are also not covered, but can be covered at
extra cost.
Transport and marine insurance is often brokered and terms can be
extensively modified.
 4
The Impact of Disruptive Technology

Not Business As Usual

For the reasons explained in earlier chapters, the insurance industry
worldwide, which in general has been a laggard in the use of advances
in IT, is on the brink of a major technological-induced disruptive
­revolution. Because the essence of insurance is data analysis, insurance
as a sector is likely to face far greater disruption than most other sec-
tors as the combination of disruptive innovations will uniquely impact
more on data analytics than other sectors. Fundamentally, insurance is
about the pricing and selection of risk. The Internet of things, Wi-fi
mobility, and big data will fundamentally change the type of risk data
insurers use, how they assess risk, how they price risk, and how they
deal with customers. This change will be complex. For example, while
self-drive cars and other innovations could reduce claim rates drastically
(self-drive cars can even be programmed to deliver unauthorized drivers
to the nearest police station), they will also reduce premiums and cash
flow drastically, and lead to the virtual elimination of car insurance as a
viable stand-alone industry within 20 years.

© The Author(s) 2017 47

M. Naylor, Insurance Transformed, Palgrave Studies in Financial Services Technology,
DOI 10.1007/978-3-319-63835-5_4
48    
M. Naylor

While insurance has often been responsive to technology, unfortu-

nately, during the last decade, while the IT world has transformed, the
insurance industry has been focused on business-as-usual in an effort
to survive the troubled times of the GFC and associated regulation
changes. IBM (2006) explores reasons for this tendency of insurers to
lag technologically, concluding that insurers tend to focus on the opti-
mization of products and processes, and accept this incremental change
as though it were innovation. During the coming disruption, however,
insurers which display continued passivity to technology will no longer
be able to survive, as the cost advantages of technological leaders or
external disruptors will destroy the cost base of laggards. Christensen
(1997) argued that new resources plugged into old processes and values
does not constitute a change-capable organization.
Brynjolfsson et al. (2011) found that companies which were in the
top third of their industries in terms of making data-driven decisions
were, on average, 5–6% more productive, and the gap was increas-
ing. McKinsey (2016) found that the top 25% of digital adapters in
insurance had revenue growth twice the rate of the other 75%, while
achieving a 26% expense ratio vs. 32% for the rest. McKinsey argues
that digitizing the top 20 insurance processes would reduce human ser-
vice costs by between 30 and 50%. These kinds of cost reductions are
achievable within five years and would reduce non-adopter profit mar-
gins to near zero. Past five years cost will drop even faster. External dis-
ruptors are starting to enter the insurance sector with US$2.6B invested
in insurance-related start-ups in 2015, up from $135M in 2010. At the
end of 2016, there were in excess of 1300 InsurTech start-ups aiming
to transform the market, and this is rising fast as the sector is viewed
by potential disruptors as ripe for the plucking. The key issue for this
book is how insurance companies and their employees will respond to
these changes.
Middle-level insurance management is often aware of and informed
about these issues. Yet surveys of what issues senior insurance company
management are worrying about, rarely mention technological change
or external disruptors as the top issue. Insurance management is thus, in
general, unprepared for what is coming. Those insurers who are aware,
 4  The Impact of Disruptive Technology    
49

are only starting to explore the implications. Most have yet to react.
Very few understand that they have to reimagine their industry.
This is in contrast to financial markets and banking which have
undergone significant change. For example, financial markets no longer
have floor traders, with complex trading decisions being handled by the
millions by software with IT-trained staff overseeing rather than being
involved. McKinsey (2015) notes that more than a dozen European
banks have replaced older statistical-modeling approaches with
machine-learning techniques. Some have experienced 10% increases
in sales of new products, 20% savings in capital expenditures, 20%
increases in cash collections, and 20% declines in churn. The banks
have achieved these gains by devising new recommendation engines for
clients in retailing and in small- and medium-sized companies. They
have also built micro-targeted models that more accurately forecast who
will cancel service or default on their loans, and how best to intervene.

Costs of Production
The costs of producing a product are a balance between fixed, ‘one-off,’
costs and marginal, ‘per unit,’ costs. Most industrial products have sub-
stantial marginal costs so there is always a limit to how low prices can
drop. By contrast, software has substantial fixed costs of creation but
very low marginal costs of reproduction.
Figure 4.1 illustrates the difference between the traditional insurance
cost model and the new. Normally, we assume that: (i) as fixed costs are

Traditional
Future
Cost Model
Cost Model
Price
Price

AC

AVC
AFC AFC
AC
MC
MC

quantity quantity

Fig. 4.1  Fixed vs. marginal cost diagram. Source Author

50    
M. Naylor

spread over more outputs then the ‘average fixed cost’ (AFC) continu-
ally falls, (ii) marginal costs (MC), the extra cost per output, tend to rise
as we try to squeeze more and more output from our fixed system thus
MC tends to rise as output expands, (iii) thus average total cost (AC)
per output tends to fall and then rise. The high level of marginal cost
occurs because each output requires materials, workers, shipping, etc.
For software products, there are typically very little costs to reproduce
or distribute a product once it has been produced. Conversely, the costs
to initially produce a product may be high. Thus, the long-run average
cost, or the cost per product, tends to continually fall as total sales rise.
The marginal profit per product is high. For example, if it costs an IT
firm $10M to produce an App, and $1 to reproduce and distribute each
copy, and it charges $100 per copy, then once it has sold one hundred
thousand copies and has covered its costs, then it makes $99 profit per
extra copy. Alternatively, it could drop its price to $10 a copy, to shut
out new entrants, who cannot cover their fixed costs at that price. This
means once a minimum market size is reached, so that the fixed costs of
production are paid, prices can be set at a very low level.
There is thus an advantage to large international insurers who can
create an expensive but efficient and world-leading IT system and then
spread the cost over its worldwide network. It is very hard for smaller
firms to react. The implication of this for employees would seem to be
that if your occupation type is large enough worldwide to justify the
initial investment in creating the software, then, if you can be replaced
by a program, you will eventually be replaced. However, things are not
that simple. The price of many service activities will certainly be dras-
tically lowered, as automation drops the costs of individualization and
customization to pennies and makes them become routine. The key
thing to be understood is that if these services have a reasonable ‘price
elasticity of demand,’ then this drastic drop in per-service cost can easily
increase overall demand so much that more workers are required, even
if each service embodies only a fraction of the traditional amount of
human labor. For example, during the industrial revolution while hand
loom operators lost their jobs, as each mechanical loom worker could
produce 1000× the output, because the price of cloth dropped to a
fraction of what it had been, demand for cloth increased so much that
 4  The Impact of Disruptive Technology    
51

overall employment in the cloth sector increased substantially. Another

useful concept here is ‘income elasticity of demand,’ mentioned ear-
lier—does the overall increase in income mean that people want to buy
more of that good? In sectors like entertainment, overall demand will
rise as IT transformation increases incomes.

Servicization
Traditionally, companies have visualized consumer demands as being
best met by producing a product; people have a need for transport so
a car is produced, people have a need for travel accommodation so a
hotel is built, people have a need of music at home so a record is pro-
duced. This product was priced at cost plus a reasonable profit margin.
Digitization disrupts this model because it allows the creation of perfect
copies at a very low marginal cost.
Rifkin (2014) argues that this creates revenue issues for companies
who produce very low marginal cost products. This is because competi-
tion frequently drives price down to marginal cost. Customers naturally
demand that their needs are met at a price near to this low marginal
cost, rather at a price which allows the full costs of production to be
recovered. Summers and DeLong (2001) argue that when marginal cost
is near zero, this can mean that firms are unable to cover fixed costs, and
therefore are unable to sustain their position in the market. This is illus-
trated by many software apps being distributed for free, with provider
income arising from add-on services. The implications of this possible
future for insurers are profound. The main implication is that surviving
insurers will need to create a wider range of income streams.
This pricing quandary was first felt by the music and media industries
from the late 1990s when music download and news Web Sites started
to destroy demand for the physical product, yet did not provide an ade-
quate replacement revenue stream. Many producers tried to resist the
technological innovations, by imposing access fees or via legal actions.
These rearguard actions have had little impact on the waves of techno-
logical innovation. When products can be provided at near zero cost
and copied indefinitely, transformation is inevitable.
52    
M. Naylor

The only viable producer solution has been to reimagine consumer

demand in a way which is financially viable. This basically involves
providing the basic product at low cost and then making profit from
a range of add-on services which the consumer values. This transforms
products to services, which is called ‘Servicization.’ Rather than being
a product which is produced at a set time and place, there is a flow of
real-time value-added services. There is not a onetime purchase but an
ongoing relationship. There is not a sales price but a subscription. Once
reimagined this process opens up a range of previously unimagined ser-
vice-based sales opportunities.
Kelly (2016) argues that when copies are near-free, companies need
to sell flow attributes which can’t be copied. Some of these are: (i) Trust
is valuable, as it can’t be copied or stored. In a world of unlimited sup-
ply, brand reputation becomes increasingly the main factor for buy-
ers. (ii) Immediacy is valuable; customers want instant availability. (iii)
Personalization is valuable; customers want a service modified to their
exact needs. As these needs change, this requires an ongoing two-way con-
versation. (iv) Service support is valuable; a near unlimited supply of data
or information means that advice, interpretation, or guides on how-to-use
are vital. (v) Authenticness and data security are valuable. (vi) Accessibility
without ownership is valuable; this removes the need for maintenance.
Consumers do not want LP records, they want music. Trying to sell
an LP record with a fixed selection of songs, chosen and sound-edited
by an expert, is outdated. What consumers value is being able to select
their own mix of songs, being able to edit, or modify those songs to suit
their taste. They want and will pay for; an expert they trust, or a group
of friends, providing recommended selections from the overwhelming
mass of songs; software which allows them to stretch or reduce song
length without the sound essence being affected; being able to access
sound tracks for each instrument, so they can produce their own modi-
fied copies; being able to share, discuss, review, or rank songs, with
people like themselves; live performances; and live streaming. Similar
disruption is occurring in travel, automobiles, media, publishing, bank-
ing, and other industries. The revenue available from all these services
was not visible to record executives when the digital revolution hap-
pened. Insurance needs to focus on the imagining how they can invent
a wide array of similarly innovative value-added services.
 4  The Impact of Disruptive Technology    
53

Administration
One of the main impacts of the looming disruption will be substantial
pressure to create completely automated, software-based internal admin-
istrative systems backed by AI, often called ‘intelligence process automa-
tion’ (IPA). This means that every activity, which can be done by software,
telematics, or drone, is done without direct human involvement.
IPA comprises five aspects: (i) robotic process automation - this handles
routine administrative tasks in software, (ii) smart workflow - this is pro-
cess management software, which integrates activities done by humans
and software, (iii) machine learning - algorithms which identify patterns
in data, providing superior analysis, (iv) natural language generation -
software that allows interactions between humans and software, based
on translation rules, (v) cognitive agents - technology which combines
machine learning and natural language generation to create a digital
workforce. It is vital to understand that the future is not a question of
software vs. humans, but rather how these two can work together to cre-
ate services which customers value.
Insurance is a very administrative process - its chief output is a prom-
ise and is thus more vulnerable than other sectors to external disrup-
tors with superior IPA skills. An insurer with a fully automated, digital,
adminstrative system would have the ability to carry out most activi-
ties at a substantially lower marginal cost, and therefore undercut its
competitors. While insurance has always been interested in the use of
computers, incumbents have recently been struggling to upgrade dispa-
rate legacy systems. Therefore, recent advances in the technology sec-
tors have left most incumbents well behind the leading edge of IPA best
practice. As will be explained later, despite increased spending in IT sys-
tems, it may be difficult for incumbents to catch up in time.

The IT Productivity Paradox

Spending on IT systems has become a major headache for most ser-
vice sector companies since 1980. Spending has become increasingly
heavy with Carr (2003), showing how IT went from 5% of capital
­expenditures in 1965 to 15% in the early 1980s, to 30% in the early
54    
M. Naylor

1990s, to 50% by 2000. The problem is that despite these increases

and the fact that IT is now the backbone of the business, productivity
improvements which can be related to this IT expenditure seems mini-
mal. This is called the ‘IT productivity paradox.’1
One explanation for this is that because IT capital expenditure alone
does not confer a competitive advantage, but merely keeps a firm in
the game, spending on IT merely reduces prices so no extra revenue is
gained. IT in a physical sense also has to be combined with the abil-
ity to use that IT capacity via related changes in management systems
and across the full range of administrative activities, so that IT becomes
integrated. IT is just a resource; it has to be used wisely.2 IT invest-
ment has so far been used to cut cost rather than allowing most firms
to differentiate themselves or expand revenue. It is thus likely that most
insurers will be forced to substantially increase IT expenditure and
revolutionize their business processes just as the cost of staying in the
industry, with minimal extra profit gained as no sustainable competitive
advantage is achieved.
An additional complication, which will be explored further below, is
that the future success will require insurers to form business networks
with related companies. A key part of these relationships will be the
integration of data systems, including data formation protocols. Thus,
insurers not only have issues with integration of their disparate legacy
internal IT systems, they will have to integrate across previously dispa-
rate industries. Insurers have little experience with this and are unlikely
to be technological leaders, and so are vulnerable to external entrants
who have these skills.

Legacy Systems
Many commentators argue that most current insurers could struggle to
cope with the changes because of issues associated with lagging legacy
IT systems, poor reputations, negligent social capital with customers,

1For a background see Schryen (2013).

2This is discussed in Wernerfelt (1984) and Breznik (2012).
 4  The Impact of Disruptive Technology    
55

and an inability to become flexible fast enough to cope with new ‘born
digital’ competitors. Large insurers, especially those born from multi-
ple mergers, are already struggling to cope with current demands for
efficient IT systems due to poorly integrated multiple legacy systems.
They are likely to be unable to cope with the future demands for the
complete recreation of their IT and management systems, in an envi-
ronment where the IT systems of external competitors are moving
forward at a rapid pace. The ability to stay up with IT system techno-
logical change will become a vital characteristic of survival, as the future
changes in IT are not only increasing but the rate of increase is increas-
ing, so the need for change is increasing - exponential change.
One of the problems insurers have is that, perversely, because their
business is data based, insurance companies have always seen the advan-
tages of computers and were early adaptors, investing heavily in the 1970s
and 1980s. At that stage, packaged software was in its infancy so insurers
developed most of their systems in-house to cover specific functions. They
are thus frequently stuck with inflexible, outdated, legacy IT systems.
This has led to a number of IT issues for current insurers: (a) many
systems were written in older code, which is less flexible than modern
code, (b) there is a mix of in-house and packaged systems, which are
often incompatible, (c) the code in many systems was not well docu-
mented, esp. later modifications, and staff involved have retired, so it is
nearly impossible to effectively optimize the code, (d) types and speci-
fication of data were created as isolated silos, so have to be extensively
modified to be compatible, (e) as new technologies have emerged, insur-
ers have tended to layer these on top of existing systems, rather than
rebuild and integrate, (f ) most data analysis takes place in batches, not
real time, and (g) existing IT processes were not created with any abil-
ity to accept real-time data. While most industries have these issues,
the early adaption of IT by insurers perversely makes their issues more
acute. Most insurers also struggle with multiple incompatible legacy sys-
tems from mergers as well as myriad sources of input data which are
non-standardized and incompatible with their IT systems.
Thus, systems and IT transformation will require insurers to invest
substantially in replacing existing IT systems, in standardizing, absorb-
ing, and utilizing the huge new data flows, in partnering with telematic
56    
M. Naylor

embedders, in new Web-based customer interaction platforms, in auto-

mated policy/claims/billing software, in changing work patterns and
mind-sets, and in retraining/restructuring staff. Product line silos need
to disappear to enable a company-wide, integrated seamless process to
occur.
The need for large investments to reinvent systems means that overall
IT costs may increase substantially in the short term, even as marginal
cost per-client falls in the medium term. Since IT systems and manage-
ment systems were also created together with differing products set up
as separate streams, re-engineering an insurer into an integrated com-
pany can be very difficult, as management processes and business mod-
els will have to be transformed as the IT system is transformed, as is
discussed in later sections.
Given that the computerized administrative systems will be the
heart of the insurer survivors and its structure will determine the way
in which an insurer can react to challenges, outsourcing of IT is prob-
lematic as decisions made by the IT consultant will impact directly on
the structure of business processes. As noted earlier, the need for change
is more about the business procedures than about the physical soft-
ware. Thus, the recreation of IT systems has to be seen as a key strategic
management issue to be decided at the highest level. Note also that for
insurers their core IT systems are the essence of their business, ‘mission-
critical,’ and cannot be turned off or allowed to fail as they are trans-
formed. Thus, IT transformation has to occur in a moving system.
Note that before insurers can exploit the exploding streams of new
data, they will need to transmit it, to store it, to create software to ana-
lyze it, and then decide which trends can be responded to by the software
versus which trends need to be brought to human attention. They will
require agreed protocols about data standardization and security, both
within the firm and across their ecosystem. All of these are complex tasks
requiring expertise in areas which incumbent staff may not be strong in.
The large size, the unstructured nature, and the diverse formats of
big data mean that extracting useable insights is not easy. For example,
many call centers now provide contact staff with all correspondence
 4  The Impact of Disruptive Technology    
57

with a customer on screen. In an era of big data, this is not possible

as information on a client will be multi-channel and overwhelming in
size. All contact employees need just the information they need, all they
need, but no more, and they need it in a way which is easy to compre-
hend at a glance.
A particular issue with big data is that its size, and complexity and
the requirement that many sources of data be combined before trends
can be found, means there is no simple way to examine data - it cannot
be ‘eyeballed.’ Instead, big data requires the creation of data visualiza-
tion systems whereby a mass of data can be distilled to key variables.
The way data is presented to humans will be critical to what kind of
insights can be achieved.
Deciding exactly what trends or detail each employee needs is a com-
plex problem, when it is still unclear exactly what insights big data will
reveal. The information will need to be presented to contact employ-
ees in a dynamic framework, with a dashboard so more is available if
required. Data visualization will thus be a key skill, a field which is only
starting to establish itself. This will require substantial investment as ini-
tially the required data visualization software will have to be customized
in-house.
These complexities led Morgan Stanley/BCG (2014) to argue that
most existing insurers are too complex and siloed to be able to change
and adapt with adequate speed and flexibility to survive against ‘born
digital’ disruptors. They argue that it is easier for external firms with
already integrated systems to add insurance as a product than it is for
insurers to transform. Zoeller Consulting Principal Kent Hopper
argued, in an Actuaries Summit in May 2015, that today’s insurers
are outdated dinosaurs who will not survive. Hopper argued that this
is because the future requires ‘companies to be data based, agile, con-
sumer orientated, mobile, smart and predictive,’ whereas current insur-
ers are ‘HQ-based, process-centric, analogue, one-size-fits-all, and too
risk-orientated.’ The result is that ‘insurers are facing increasing costs,
with declining sales and customer loyalty’. Modern companies ‘respond
to customers in seconds, not weeks.’
58    
M. Naylor

In the past, established insurers have been protected from new

entrants by the high fixed costs of developing systems and distribution
capacities. In the new digital era, however, these barriers have disap-
peared, as many IT-based companies have better existing systems than
insurers do, making it easier for new entrants to develop the required
systems than it is for existing insurers to transform their systems. The
obvious advantages to customers of the future IT-based insurance will
also remove most regulatory barriers.
Firms have to get past the idea that they can survive by adding to
existing legacy systems with a few Web site or apps. They need, instead,
to reinvent entirely how they operate. The core skills of successful future
insurers will be an ability to create and utilize extremely client friendly
software, an ability to create data networks, and an ability to build
social capital.
The refocus on IT systems also changes the scale of dynamics in
insurance, as the ability to rent IT capacity via the cloud means that
large insurers will no longer have a size advantage in term of IT capac-
ity. Effective software, not ownership of mainframes, will be the key
insurer asset.
A key element in the survival rate of existing insurers is the speed of
entry by firms from outside the current insurance sector who already
have these skills, as well as social capital. These external entrants will
find it easy to enter and destroy the market of those insurers too slow to
respond. An illustration of the ability of out-of-sector firms to disrupt is
the destruction of Kodak by firms from the electronic sector.

Customers
A key impact of the impact of Web-product-comparison has been
the growing strength of customers versus companies. As explored ear-
lier, a key to company success in the future will be the level of their
social capital. This has not traditionally been an area which insurance
companies have focused on or been successful at. Insurers have instead
focused on internal issues like pricing risk and controlling internal costs.
 4  The Impact of Disruptive Technology    
59

They rarely interact with customers directly. One of the key impacts of
digitization is that it will transform insurance from a seller’s market to a
buyer’s market.

The disconnect between life insurers’ value proposition and

today’s customers3
Imagine Susan, a 34-year-old high-school teacher who is expecting her
first child and hence decides to buy a life insurance policy. First, it is very
unlikely that a provider will proactively reach out to her, or know she is
about to undergo a life-event. If she tries to shop for a policy online, she
may be intimidated by complex products and technical terminology that is
confusing and makes her feel incompetent.
If she looks for an agent and is lucky enough to find one that she feels
she can trust, she is likely to have concerns about how much the policy
will really cost, the meaning of all the fine print, and whether the agent is
more interested in her needs or in a quick commission.
If Susan overcomes these doubts and decides to buy a policy, she will
begin what may be an arduous application process: lengthy forms with
sensitive medical questions, weeks of waiting, and little transparency on
where things stand. She may wonder why the process is so complicated
and time-consuming when many companies in other industries offer con-
venient, fast (and mostly digital) services.
If she has questions on her coverage after buying the product or wants
to change her policy, she will likely struggle with the limited self-service
options offered by the insurers and may find that her agent has moved on.

Such ‘pain points’ may make Susan abandon the process or take her
business elsewhere.
McKinsey (2016) argues that the traditional insurance product cre-
ates six ‘insurance disconnect points’ along the customer’s purchase
path, each with reasons why prospective customers fail to be converted
into loyal customers. These pain points are:

3Example is from McKinsey (2016) Transforming life insurance with design thinking.
60    
M. Naylor

1. Awareness of Product:
(a) Limited product engagement.
(b) Doubts about product value.
(c) Incomplete product understanding.
(d) Tendency to put off purchase (seen as a disagreeable chore).

2. Consideration of Purchase:
(a) Low trust in delivery.
(b) Limited rapport with broker.

3. Purchase Decision:
(a) Confusion about product features.
(b) Doubt if product is right fit.
(c) Feeling of ‘being sold to’.

4. Application Process:
(a) Frustration about delay and lack of transparency.
(b) Feeling rejected when application turned down or terms
imposed.

5. Customer Care:
(a) Limited, or no access, to quick, easy, self-service tools.
(b) Limited product flexibility.
 4  The Impact of Disruptive Technology    
61

6. Claims:
(a) Feeling vulnerable and overwhelmed.
(b) Need for emergency bridge help.
(c) Feeling taken advantage off when under emotion and financial
pressure.

These business-leakage drivers create some of the main problems with

personal risk insurance: (i) low customer engagement, (ii) limited
ability to meet the expectations of Generation Y or millennials, (iii)
large, inflexible, stock of legacy polices, and (iv) slow product change
cycles.
The movement of customer acquisition from intermediation via
agents toward direct sales will also reduce customer stickiness and
increase retention issues. IBM (2015) found that only 37% of US cus-
tomers trusted their own insurer to a reasonable degree. While cus-
tomers of agents trusted their own agents, non-customers did not, in
general, trust agents. IBM also found that 40% of customers who did
switch providers did so because their old insurer could not meet their
changing expectations. They also found that 58% of respondents think
that insurance companies are more difficult to deal with than tech-based
firms. Given that acquiring new customers tends to cost about four
times the cost of retaining existing customers, changes in customers’
stickiness can have substantial cost implications.
The insurance industry generally rates low in terms of client trust,
with Equinix (2014) stating that insurers rate below banks in terms
of trust, which is bad given the public hatred of banks after the GFC.
PWC (2014) argues that companies which do not refocus from a
product orientation to a total customer-centric orientation will not
survive past 2025. They found that the top three customer needs were:
(i) accessibility, (ii) tailored to customer, and (iii) content explanation.
62    
M. Naylor

They argue that most companies who try to respond to the challenges
of the digital era put digital innovation at the forefront of their strat-
egy. This is a mistake as it should be the customer who is the focus; IT
developments are the means to that end, rather than the end. Many
existing insurance companies, however, may find that their legacy of
poor service in customers’ minds and the resultant low social capital
is a hurdle impossible to reverse. Companies with existing expertise in
building social capital in other fields, and high social capital, can thus
be expected to enter the insurance market as they will be viewed posi-
tively by customers.
McKinsey & Company (2016) argue that for generations, insurers have
delivered their promise of financial security with the help of strong actu-
arial functions, intermediated distribution channels, complex products,
limited service, rigid processes, and cumbersome customer interactions. In
the future, customers will reward transparency, speed, and flexibility.
Another IT-based change has been the evolution of groups of cus-
tomers with similar risk profiles bought together by Web-based soft-
ware. Current products are aimed at an amorphous average, yet many
customers differ from that, resulting in them not being offered cover
or feeling disconnected from the product and difficult to market to.
Traditionally, groups with similar differing needs have had to find each
other and mutually identify themselves, an information problem which
was difficult and expensive. Now, use of deep-social-data-and-web-
search-algorithms enables specialist companies to identify these differing
customers and offer them to as a group to insurers who are seeking that
risk profile. Groups of customers can also create their own groupings via
software and bulk-buy from insurers. The insurer can then create a cus-
tomized communication strategy and additional sales opportunities. An
example of this would be cancer patients who are seeking travel insur-
ance for short restorative trips.
A key question is how accepting customers will be of the new tech-
nology. The answer at this stage is that customers are happily accept-
ing elements like adaptive-cruise-control, use of Siri, software telephone
receptionists, online shopping, at a fast pace. The willingness of
 4  The Impact of Disruptive Technology    
63

customers to accept new approaches seems to be increasing over time,

e.g.; customers accepted smartphones far faster than they accepted the
initial mobile phones.
An example of this is the Japanese Web-marketplace, Rakuten, which
purchased an underwriter and now offers insurance to its wide network
of existing customers. In New Zealand, the main online shopping Web
site, TradeMe, is already offering the most popular insurance premium
comparison engine and has taken the first small step to selling policies
on behalf of insurers. It is only another small step away from selling its
own policies, underwritten by reinsurers. The fact that these disruptors
have not yet entered a particular market is not a good reason for the
delay.
Legislators in some countries may be unable to keep up, but the pos-
sibility of online insurers or advisers operating from a foreign jurisdic-
tion which has state-of-the-art legislation will give some impulse to
politicians to reform. The lagging elements at this stage are insurance
companies and advisers. They will either update or lose business.
Digitization will transform insurance from a seller’s market to a
buyer’s market - the customer will be a king. Insurers do not currently
treat them as such. While traditionally insurance customers have been
static, accepting poor service and finding switching difficult, the new
breed of customers are becoming restless and looking for alternatives, as
switching difficulties have been slashed. Capgemini/Efma (2016) details
comprehensively how Generations Y and younger have drastically dif-
ferent and more demanding customer expectations than baby boomers
and argues that insurers will have to adopt. They found that, in general,
customer expectations are evolving faster than insurers’ ability to inno-
vatively address these expectations. Customers’ disappointing insurance
experiences led them to be open to external disruptors. Customer sticki-
ness is falling.
It is vital to note that modern consumers do not judge their customer
experience of an insurance company against that of other insurance
companies, but instead judge it against their experiences with digital
leaders such as Amazon. Yet insurers still tend to benchmark themselves
64    
M. Naylor

against other insurers instead of benchmarking against possible external

disruptors. In terms of a quality real-time omni-channel interface and
a quality customer experience, insurers, in general, perform so poorly
that they are seen as ‘in a different world’ from the digital leaders, and
therefore vulnerable. This means that being the ‘best insurer’ in an area
is meaningless as a metric for long-term survival.
Morgan Stanley/BCG (2014) argues that, so far, insurance custom-
ers have found their insurance online satisfaction in the bottom 10%
of their overall online experiences, perceiving their needs as significantly
unmet and the products as expensive and inflexible. Customer expecta-
tions have changed faster than insurers have realized, with expectations
that they should be able to engage with insurers via multiple channels,4
especially mobile devices, without being bogged down in lengthy and
annoying phone conversations.
Morgan Stanley/BCG (2015), in a worldwide survey, found that
50% of insurance customers indicated they would switch to any insurer
offering superior mixed-channel service, with 80% of all customers will-
ing to consider new digital insurance products, and this peaked at over
85% in the 25–35 age group, and at over 90% in the highest income
group. About 85% were willing to divulge personal information, if sig-
nificant premium reductions were offered. It is interesting that insurers
were seen as the 3rd most trust-worthy industry in terms of handling
personal information, which gives a basis to work from.
The most important change that social media has created for insur-
ers is that they can no longer create their reputation in isolation from
customers - the value of a company’s reputation is now co-created with
customers. It is no longer a one-way process, from the PR department
to the customer, it is now inherently two-way. Most insurers have not
yet recognized this, nor understood what this means for reputation
value creation.
Castriotta et al. (2014) argue that insurers need to use the DART
model (Dialogue, Access, Risk assessment, Transparency), and that the

4For omni-channels customer expectations, see Target Group (2015).

 4  The Impact of Disruptive Technology    
65

relative absence of insurers from social media platforms has meant that
they are not visible to digital natives and when they are present they
tend to talk to customers, rather than talk with them, which is worse
than being absent. This played a major part in the declining sales to this
generation. Instead, insurers need to build high-quality digital spaces
where customers can interact and feel deeply involved in activities
which are not strictly related to the world of insurance. The key future
insurer social media skill will be able to create dialogue to jointly solve
problems. Digital laggards are rapidly disappearing from customer
awareness.
Digital natives, in particular, will switch to more in-tune newer
entrants, who will meet their expectations of substantially lower costs
and superior online service. Surveys show that digital natives exhibit
little insurer brand loyalty, but have robust service expectations.
Experience with CRM systems is cumulative so first movers in this area
will enjoy significant competitive advantages. GEICO is an example of
a new entrant to auto-insurance which has used heavy investment in IT
to cut admin and claim costs while pleasing customers. This has meant
that it has grown its market share significantly faster than older com-
petitors, who are now struggling to catch up.
Note that there cannot be one-answer solutions. Most insurers use
CRM systems which are reactive, and not proactively in-tune with cus-
tomer needs. Yet research shows that customers, in general, prefer low-
touch self-service options when browsing or researching, but when
something goes wrong or they are confused, they want high-touch,
personalized service. The younger generations will not put up with the
current low level of insurer service and will actively switch. Consumer
technology firms which already offer this style of customer service, and
have established social capital with the digital generation, are prime can-
didates for disruptive entry into the insurance sector.
Another key element is that customers are becoming more diverse,
demanding that companies engage via multiple channels. This is due
partly to growing ethnic diversity, but is also due to differing capacities
of customers to react to the accelerating technological changes. Thus, a
key skill for survivors will be the ability to retain and simplify e­ xisting
66    
M. Naylor

channels to retain low-tech clients while expanding new channels to

attract high-tech clients. Any indication from insurers that they regard
users who are reluctant to move to new channels as illogical would
be harmful to their reputation.
A better way to define customers than by birth is in terms of those
who embrace technology, ‘Digitals,’ versus those who do not. BCG
(2016) found that 5% of over 55s embrace leading edge technology
vs. 25% of those aged 31–54 and 33% of millennials. This makes digi-
tals about 20% of the current US auto-insurance market and growing.
When compared to non-digitals, digitals switch more frequently, (14%
switching within the first 9 months vs. 7%), do more product research
(70% vs. 47%), use multiple channels (30%), are more likely to use
price comparison sites (35% vs. 25%), are more likely to take an agent
or broker recommendation (21% vs. 6%) or employer recommenda-
tion (10% vs. 3%), are less motivated solely by price (69% vs. 84%),
are more motivated by loyalty programs (34% vs. 18%), are more likely
to use banks (42% vs. 24%) or Internet providers (38% vs. 17%) or
household service firms (18% vs. 8%), and are less likely to purchase
direct from an insurer (52% vs. 77%). The implication of this cultural
change is that Internet-based firms are more likely to survive in a world
of digital insurance than current insurers are.
Yoder and Rao (2015) found that 71% of insurance CEO’s sur-
veyed regarded the increasing diversity of customers as a major dis-
ruptive trend. Product suppliers and distributors will have to create a
more diversified range of products, which are simultaneously simpler
yet broader, and offer an ability to be customized to customer needs
at minimal additional cost. This may require multiple brands with dif-
ferent images or customization. The younger generations, in particu-
lar, will not be satisfied with a few standardized choices. Customers
will demand seamless and efficient feedback and a positive response
to their request for customization within a consumer-centric culture.
They have become used to this from online retailers and will not com-
prehend why insurance companies can’t tailor products. In a world
where McDonalds offers the choice of ‘build-your-own burger’ via a
software interface, insurers need to offer customized products. Mass-
customization is the future.
 4  The Impact of Disruptive Technology    
67

An essential reason for product simplification is that, as Equinix

(2014) argues, with technological complexity increasing and custom-
ers’ available time decreasing, it is no longer reasonable for companies to
take a ‘buyer-beware’ stance. Society will increasingly expect the burden
of protection to fall on those creating the product or service. Providers
who do not fulfill this expectation will face a devastating backlash from
society. Producers whose products or services are seen to be ‘unfair’ via
confusion will lose client trust. Insurers will have to reduce the per-
ceived complexity of their products while being completely transparent
and compliant with regulation.

Disintermediatization
Two big themes in other areas of financial services have been the auto-
mation of administration and the separation of distribution from
production. The growth of software in administration, the ease of inter-
national transfer of information, and the vital nature of social capital
mean that insurance companies no longer need to do be involved in all
areas of the insurance business. Theory argues that integrated companies
only exist because internal operations are cheaper than external contrac-
tors. The Web and explosion of data is changing these cost structures.
An example is banks and the finance sector, which disintermediatized
in the 1990s, e.g.; mortgage originators developed who originated and
then on-sold loans to merchant banks who then spiced and diced and
then on-sold to investors.
There is amble proof from other sectors that category killers often can
offer one area far cheaper than an integrated provider can. McKinsey
(2016) points out that every part of the insurance value chain is cur-
rently under attack by an external disruptor who can potentially do that
part of the chain better than the incumbents. Why should one com-
pany find insurance clients, as well as underwrite clients, manage big
data, process claims, and hold capital to fund claims? Business functions
like creating a brand name, dealing with customers, underwriting, hold-
ing risk, handling claims could all be handled by different companies
in a way which is seamless for the customer. Insurers can outsource
68    
M. Naylor

areas which can be handled by a specialist suppliers of each of these

services at a lower cost. Insurers have resisted disintermediatization so
far, but may in the future be forced to go down this route. This may
occur because current new entrants are focusing on part of the insur-
ance chain.
Most administrative aspects of insurance can be replaced by software,
so external firms with a high level of customer skills can easily enter by
creating or buying the appropriate software,‘an insurance company in a
box,’ adding this to their existing products, and outsourcing areas like
claims. An IT-focused firm could provide a Web-interface, backed by
data analytics, with non-programmable areas outsourced, or by selling
to existing customers as a franchise, or by distributing via an outside
chain. KMPG (2016) argues that technology has evolved to the point
where new insurance organizations can be virtually ‘stood up’ in a mat-
ter of weeks by renting everything from servers through to customer
interfaces. World Economic Forum (2015) argues that all links in the
insurance value chain are at risk from a new entrant aggressively moving
into one link, using contractors to supply other links and utilizing new
sources of capital, like cat bonds. It is now relatively easy for an external
disruptor to enter.
These developments could lead to product development, distribution,
underwriting, claims, and risk capital supply, being the area of niche
players. As telematics allows a finer pricing of risk, insurance may be
sold as a loss-leader by customer product providers in other industries.
Insurers thus need to assess which aspects of their operations are vital
and need to be in-house and which aren’t. Insurers may, for example,
decide to withdraw from customer contact and focus on risk analytics.
Insurers need, however, to retain core business units as history is full
of examples of companys which sliced off what seemed less important
areas only to find that they ended up losing a key future business skill.
The value of these core units is not just the profit but the business cul-
tural learning. This type of disaggregation is very likely a strong future
trend for insurers
 4  The Impact of Disruptive Technology    
69

Big Data and Analytics

Traditionally insurers have had very infrequent contact with the cus-
tomer, only at policy inception, renewal and claim time and therefore
know little about their clients. KPMG (2014) states that current insurer
client data is historical, shallow, and based on very few interactions. The
new sources of data will give insurers substantially larger amounts of far
higher quality data about their clients and they will thus be able to use
software to assess each client individually. Combining this with auto-
mated administrative, customer contact, and underwriting, systems will
allow insurers to offer customized policies at low additional cost, as well
as substantially reduce client churn and increase satisfaction. KPMG
(2015) thus argues that effective use of big data will be one of the most
important sources of future insurer competitive advantage. It is not just
a temporary trend but a massive and permanent change in the data used
by insurers. It will change the way insurers interact with their clients,
the way they underwrite, and the way they structure their administra-
tion. It will be as much about software and systems as data. Big data as
a competitive tool involves mainly the way data is used and analyzed
rather than the actual data. CII (2015) argues that big data analytics is a
game-changer and insurers who do not fully integrate the new methods
into their structures will not survive.5
On the up-side, insurers will gain a greater depth of understanding of
personality profiles, buying trends and behavior, vastly exceeding what
could be previously imagined. Yoder and Rao (2015) argue that the
combination of big data, telematics, and complexity analytics, will allow
insurers to move away from simply descriptive (what happened) and
diagnostic (why it happened) analysis toward predictive (what is likely
to happen next) and prescriptive (determining how to ensure the right
outcome) analysis. Multiple scenarios can be run and likely outcomes
role-played. This will create more robust supply chains and stronger
conversion rates. Insurers will gain an in-depth understanding of why
customers insure or don’t insure, and be able to individualize customer

5​Boobier (2016) provides a good summary of data analytic issues in insurance.

70    
M. Naylor

approaches so as to maximize sales. The key here to the use of increased

data is automated administration, as this will ensure there is a very low
marginal cost to what would have previously been an expensive chore if
done by workers. Underwriting will also be more accurate as it will be
based on a rich depth and breath of detailed data on causal variables.
Current analytical underwriting and claims systems have been shown
to have the ability to increase the speed and cut cost of underwriting and
claims processing by up to 100x, often reducing claims processing times
from months to minutes. Similar techniques can be applied to customer
relations, or sales data, pinpointing what issues or styles are important for
each segment of the market. Probably, more importantly, successful insur-
ers will use detailed underwriting analytics to allow them to offer person-
alized products, with products broken down into blocks, and customers
choosing the mix of products that suit their needs and their budget, with
the insurer able to provide detailed and personalized advice at each step.
This will allow them to construct tailored and personalized products,
rather than the traditional one-fits-all model. How, for example, do
higher risk members of the white, female, suburban, SUV-driving group
differ from higher risk members of the white, female, city-center, non-car-
owning group? Does gym attendance 2x a week differ in health outcomes
from 4x a week? Customers will feel the insurer has been listening to their
needs and differences. Special risk or small group customers, now deemed
uninsurable, will be easier to handle. Groups of customers with desirable
characteristics will be able to interact and seek group discounts.
Michal Kosinski showed in 2012 that on the basis of 70 Facebook
‘likes’ they could know a persons’ social and purchase preferences bet-
ter than the person’s friends could, by 150 ‘likes’ they could out-per-
form what the person’s parents knew, and by 300 ‘likes’ predict better
than their life partner. Adding in additional social media and mobile
phone data allowed them to predict a person’s preferences better than
the person themselves could. Firms have built on this style of analysis6
to develop very detailed psychological profiles of nearly every person in

6Facebook has tried to prevent use of its data for this, but multiple other sources of data are

available.
 4  The Impact of Disruptive Technology    
71

selected countries, showing how to target the emotionally appropriate

appeal. This involves variation in multiple areas like message, graph-
ics, style, color, font. Rather than one TV ad, an insurer could send out
100,000 differing targeted digital ads. Feedback from response to this
can then be used to target the next round of ads.
Analytics will also allow customized client contracts, for example it
may reveal if a customer belongs to a group likely to lapse, or to change
providers, and at what point in time or in what circumstances this lapse
is most likely to occur, and will then give the insurer the capacity to
provide a personalized insurance experience or policy contact structured
in a way most likely to retain that customer.
Big data is also starting to revolutionize risk assessment, which because
of the growing size of data sets and their very intensive nature, means
that patterns can be found which are not visible when assessing a sample
data set. For example, in health care, the digitalization of millions of cli-
ent medical records has allowed software to compare each client’s indi-
vidual symptoms with their genetics, their family background, their gut
bacteria, their environmental factors, and other factors to create optimal
and personalized treatment plans, and individualized medicines, all with
minimal human oversight. New patterns have emerged which have dras-
tically changed some traditional ideas about optimal treatment. Patients
no longer have to be put into categories and treated as groups.
Complexity arises because data needs to be combined from as wide
a range of sources as possible, many of which can be quite different;
for example, minute-by-minute purchase data, location data, text,
online comments, blogs, and call center communications. Combining
these diverse sources can yield unsuspected insights. A useful source of
data for insurers is retail data from loyalty cards. KPMG (2014) points
out that current insurer client data is historical, shallow, and based on
very few interactions, whereas loyalty card data is deep, rich, and real
time. An example of the possible gains is the discovery by Woolworths
Australia that customers who drank lots of milk and ate lots of red meat
had a significantly lower auto-insurance risk than customers who drank
spirits, ate lots of pasta, and filled their petrol tanks at night.
72    
M. Naylor

Gathering the data to create these types of insights will require pro-
active collaboration with other companies in the industry and in unre-
lated industries. This will require sector-wide data standardization
agreements as well as a transformation in insurer attitudes to data.
Insurers currently tend to treat data as private property and handle it
in siloed, furtive, and undisclosed ways. Yet the larger the database, the
more the insights, so that in the future, insurers who refuse to collabo-
rate with a wide ecosystem will suffer a competitive disadvantage large
enough to bankrupt them.
The major issue facing insurers, who have always been fundamen-
tally in the data business, is that the overwhelming size of the future
stream of data means that they will struggle to cope unless they funda-
mentally change the way they analyze data. Software, not people, has to
scrutinize the data for meaning and then present limited aspects of it to
humans. Data analysts will not visually purview data but will build AI
systems to mine it for useful insights. Skills in leading edge data visuali-
zation are rare, yet unless managers get access to data in a form which
can be used for strategic decisions, they may as well not have the data.
Commentators thus argue that if you’re a huge, data-driven, business,
you will either have cutting-edge AI driving your company products,
or you’re out of business. Key skills will be big data analytic skills and
data presentation skills, as well as deep knowledge of customer behavior.
Workers with outstanding skill sets in all these three areas are limited
and will be in high demand. Workers with only IT skills will need to
work within multi-skill teams.
Existing data software suppliers may not survive the transition if they
do not have skills in AI mining of large and disparate data volumes.
Instead, social media firms or consumer Internet players, who do have
experience in AI mining, will probably take over supply of data soft-
ware. Many new players in this area already exist, e.g., Captora uses data
science to personalize and evolve customer-specific marketing material.
UK payday loan firm Wonga used social data to cut its default rate from
50% to less than 10%. Insurers thus need to be proactive in using a
range of software suppliers, and even ceding control of this area.
CII (2015) argues that the two key customer issues with big data are:
(i) the value which customers experience or perceive from its use and
 4  The Impact of Disruptive Technology    
73

(ii) the extent to which insurers’ handling of big data builds or under-

mines trust. Insurers need to have an intensive and ongoing conversation
with customers about data collection and use, so that customers expe-
rience a shared realization that the changes will be mutually beneficial.
These two issues are linked as insurers currently have very little contact
with customers and engage in no trust building exercises. Within this
scenario, the use of increasingly inquisitive data thus has myriad issues.
This means that data decisions cannot just be taken according to the law,
but must be carried according to what the public considers is right, so
the insurer can be actively and publicly proud of their changes.
The Chartered Insurance Institute, CII (2015), argues big data is
the most important issue for insurers for the current century and that
successfully handling big data will be the defining characteristic of suc-
cessful survivors. They argue that two issues in particular require profes-
sional judgment: (i) social sorting and the issue of how to handle those
customers who will be worse off, and (ii) the right balance between per-
sonalization and pooling of risk. There will be a number of issues which
will involve complex managerial or ethics decisions and these will need
to be proactively dealt with. An example here is how to deal with elderly
drivers whose telematics indicates they are a risk to other road users, or
how to deal with families who have an inherited cancer risk gene?
Other issues CII (2015) identify include: (i) given real-time risk data,
how much can the rate/premium fluctuate before the customer starts to
regard insurers as unstable institutions, (ii) how will customers deal with
knowing that the insurer knows a lot about them which the customer
did not actively tell insurers, especially if sourced from informal social
media, (like photos of the claimant attending a big party before driv-
ing, or trying to sell a customer a change in their life cover because the
client searched online for baby products), (iii) can individual clients set
differentiated rules on what data can be collected on them, (iv) will data
trading across firms lead to customer unease, especially if the same piece
of data can mean different things in different contexts (like increased
purchases of plus-sized clothes), (v) will the public trust of insurers drop
if software dominates in decision making, (vi) can premiums decisions
based on correlations between not-obviously related data be justified to
clients, in terms of causation, (vii) will personalized underwriting and
74    
M. Naylor

the related increase in pricing complexity lead to confused and disgrun-

tled customers, (viii) what happens when automated pricing or claim
software fails to know an important fact which makes the decision
unethical from a human viewpoint, and (ix) how to embed ethics and
professional principles into automated decisions.
It is vital that automation does not decrease the ability of insurers
to think critically about the issues, and the ability to reflect upon the
implications which flow from that decision. Human oversight of auto-
mated administrative systems will be required as software can make
decisions which seem incredibly stupid to humans. This requires insur-
ers to run as wide a range of scenario tests on software as possible and
to carefully set limits whereby the software brings an issue to humans
for checking or decision. Software, especially AI programs, needs to be
structured so humans can oversee it and actively intervene if the ‘right
decision’ is not being taken.
All these challenges mean that existing insurers will have to create
entirely new IT, data flow, and decision systems based on new manage-
ment processes. KPMG (2015) argues that the issue for current insurers is
that in large organizations data is normally trapped into silos and locked
behind access controls so it is virtually impossible to gain an integrated
overview of reality. Data governance and ownership is also normally
spread across an organization so that there is no centralized view of what
data they have and how it is used, so that it is impossible for any creative
new uses of the data to be envisioned. Some of the most useful insights
will come from comparing diverse data sets which seem to have no con-
nection. KPMG argues that the management culture shock of moving
from legacy systems to what will be required is profound. This implies
that external companies who already use big data effectively will find it
easier to move into insurance than insurers will find it to transform their
legacy systems. To survive, insurers will need to see the changes as a big
opportunity for self-transformation rather than as a big challenge they
have to overcome. Management needs a proactive enthusiastic response.
McKinsey (2015) argues that companies embarking on machine
learning should: (i) Firstly, investigate all feasible alternatives; (ii) sec-
ondly, pursue the strategy wholeheartedly at the top executive level; and
(iii) thirdly, use existing executive expertise and knowledge to guide the
 4  The Impact of Disruptive Technology    
75

application of that strategy. The people charged with creating the stra-
tegic vision may well be (or have been) data scientists, who will need
guidance from top executives overseeing other crucial strategic initia-
tives. More broadly, McKinsey argues that companies must have two
types of people to unleash the potential of machine learning; ‘Quants’
who are schooled in its language and methods, and ‘Translators’ who
can bridge the disciplines of data, machine learning, and decision mak-
ing by reframing the quants’ complex results as actionable insights that
generalist managers can execute.
McKinsey argues that behavioral change will be critical, and one of top
management’s key roles will be to influence and encourage it. Traditional
managers, for example, will have to get comfortable with their own soft-
ware variations. Frontline managers, armed with insights from increasingly
powerful computers, must learn to make more decisions on their own, with
top management setting the overall direction and zeroing in only when
exceptions surface. The front line will have to be provided with the neces-
sary skills and setting appropriate incentives to encourage data sharing.
Executives should think about applied machine learning in three
stages: description, prediction, and prescription.

• The description stage is about collecting, developing, and analyzing

databases. This online analytical processing is now a routine activity.
• The prediction stage is more difficult as current technology already
allows businesses not only to look at their historical data but also to
predict behavior or outcomes in the future. While data quality can
be an issue, McKinsey argues that past IT investments have equipped
most companies with sufficient information to obtain new insights
even from incomplete, messy data sets, provided of course that those
companies choose the right algorithm. Adding exotic new data
sources can in fact be of only marginal benefit compared with what
can be mined from existing data warehouses.
• The prescription stage is the opportunity for the future and therefore
must be where top executive attention is focused. The aim is to pre-
dict what customers are going to do; only by understanding why they
are going to do it, can companies encourage or deter that behavior in
the future.
76    
M. Naylor

While some current executives may see a software driven future as threat-
ening, over time, AI software will become culturally invisible in the same
way technological inventions of the twentieth century disappeared into the
background. The role of humans will be to direct and guide the algorithms
as they attempt to achieve the objectives that they are given.

Telematics and Dynamic Insurance

The linking of built-smart connected products with Wi-fi networks has
profound implications. Porter and Hepplemann (2015) point out that
this is more than the linking of products by the Internet, but is a trans-
formation in the nature of products and the way we use them. The data
processing innovations of the 1970s on, and the Internet innovations
of the 1990s on, left products largely unchanged. Now products, which
are everywhere, will be transformed. The resultant cultural change will
be profound.
This telematic connectivity can be: (i) one-to-one - one product join-
ing to one central unit, (ii) one-to-many - one central unit joined to
many sensors, or (iii) many-to-many - each sensor joined in a network to
many others. It is the last which is the most transformative, as a smart
connected product has qualitatively different implications than just a
smart product.
Insurers are, in general, not up-with-the-play in the area of the
Internet of things. They are currently either ignoring the Internet of
things or viewing it negatively. What insurers need to realize is that tele-
matics offers them a qualitatively different way to handle claims and
relate to clients.
The impact on insurance claims could be immense. Claims will
be a lot easier - as insurers will get real-time feedback from their tele­
matics. For example, the networked chips in the car will inform the
insurer’s computer instantly of the crash, call emergency services, and
software can instantly assess who was to blame by comparing the two
cars’ records, inform the tow truck and book the car repair, and make all
payments, all without human intervention. Smoke alarms can already
phone both you and the fire brigade, notify the repair service, and then
 4  The Impact of Disruptive Technology    
77

talk to the oven or fridge or TV to ascertain the cause. A biochip can

sense the results of changes in diet, exercise patterns, or health indica-
tors, before the customer or their health professional is aware of any
health issue.
Once real-time data feedback from telematics is added, the conse-
quences for rate-making are revolutionary. The feedback from embedded
devices will enable premiums to be set on an individual basis by software,
and adjusted in real-time as insured behavior changes - what is called
‘dynamic insurance.’ Clients will be underwritten individually, with pre-
miums dynamically set based on feedback from the telematic devices. For
example, the analytics may reveal that a particular activity is risky for that
kind of customer, and rather than excluding it, the insurer could offer
cover at a different rate during that time, as long as the customer uses
appropriate telematics to indicate to the software when to charge a dif-
ferentiated premium. Clients who engage in an unauthorized temporary
pattern of risky behavior can have their insurance immediately canceled.
The use of networking will allow the creation of real-time dynamic
risk-rating, which will transform the insurer relationship with clients.
Thus, for a networked linked car, a client who changes from a pattern of
limited driving to intensive commuting for a short while may find their
premiums jump for that period, whereas a client who drives more safely
will find their premiums drop. Thus, premiums may be set weekly or
even shorter, based on usage. Since insurers will get real-time feedback
on when cars are used, it will be possible to offer innovative products.
For example, several insurers have started using telematics to trial asking
car owners to pay a ‘driving premium’ when they actually drive, with a
‘residual premium’ if the car is parked in the garage. Given that cars are
parked for an average of 95% of their lives, it would have a huge impact
on the distribution of premiums, with heavy users of cars paying multi-
ples of the premiums which light users pay. Higher rates could be set for
evening or holiday driving.
Retailers are already putting tiny Wi-fi-linked spots on goods, so if
they are stolen they can be tracked by a mobile phone. In trials, this
has led to rapid multiple arrests and decimated the shoplifting commu-
nity. The flow-on effect of theft becoming non-viable is that premium
income for contents insurers will drop sharply.
78    
M. Naylor

It is a common mistake in insurance circles to think of telematics as

impacting solely in the area of general insurance. Morgan-Stanley/BCG
(2013) argues, however, that personal risk areas like medicine are rapidly
developing a range of Wi-fi devices to monitor those with medical condi-
tions, including wrist devices which can monitor blood components, in
real time. Ralph Lauren’s Polo-Tech shirts stream various athletic health
data to the wearer’s mobile phone. Health providers could issue acute
patients with this kind of clothing so software can track key indicators and
alert staff if preset levels occur. Stick-on patches (bio-stamps) are being
developed which read blood contents. The patient can then be warned
to seek help or their mobile phone can confirm location and contact an
emergency health service. Linked to a mobile phone these can send a con-
tinuous stream of updates to a hospital computer so that unusual trends
can be spotted before the patient may be aware of them. Treatment plans
and drugs can be modified based on real-time feedback, especially if an
auto-drug-dispenser is integrated into the overall system. This will lead to
preventative treatments likely to cut health costs. With the computeriza-
tion of hospital administration, all tests, all services, all patient health met-
rics can be linked to the insurer so immediate payment can be offered and
trends in the effectiveness of drugs or procedures analyzed.
Taken a step further, these devices could be installed in the healthy
clients, to record exercises, heart pressure, calorie or alcohol consump-
tion, etc., with large premium discounts given to clients who follow
agreed risk reducing behavior. Changes in blood chemistry would reveal
to the software what the client is eating and could send the client feed-
back, offering discounts or increases. Life insurance can then be under-
written based on these indicators.
The sharing of personal information is part of everyday life for mil-
lennials and any privacy concerns will be quickly outweighed by the
substantially lower premiums. Algorithms will quickly work out what
combinations of behavior led to less claims and offer discounts to
selected clients if they improve their habits. Health researchers will have
an unprecedented database of information to establish definite cause
and effect relationships which will allow illnesses to be identified at an
early stage. The focus will shift from sickness to health. This is only pos-
sible cost-wise if customers and their telematic devices interact via soft-
ware, so the marginal cost of altering premiums is near to zero.
 4  The Impact of Disruptive Technology    
79

Individualization
Digital natives will not accept a standardized product from a firm which
does not seem to be paying them much attention. Expectations are that
the firm should work hard to gain their business, that the firm must
listen to the needs of the digital natives, and that the firm must person-
alize both advice and products to fit. This response and personalization
should be universal and rapid. This kind of customer service cannot
happen at a reasonable cost and within the expected time frame without
the majority of the administrative process being automated and analy­
sis provided by AI software. Increasingly all insurance customers will
be underwritten, and dynamically, rather than just at policy inception,
or renewal. The integration of that pricing software with big data will
allow insurers to very accurately price policies.
Insurance staff will oversee the creation of products via testing and
improving the software, rather be directly involved in product creation.
This will demand staff who have a deep understanding of the interac-
tion between software and customers, as well a heavy investment in IT
systems.
Client discrimination will no longer have to be on the basis of gross
characteristics like gender or age, but on actual causal factors. The obvi-
ous issue that individualization creates is how to deal with customers
who have genetic or other unchangeable issues which put them into
high-risk groups. Legislators will have to come to arrangements with
insurers on how to provide these groups with required insurance cover.
On the up-side, individuals from groups who currently face under-
writing issues, like young drivers, will be able to prove that they do
not individually possess the behavior patterns which make that group
high risk. Note also that for the EU, use of personalized rates based on
dynamic data derived via telematics will enable the impact of gender
neutrality to be limited.
A likely future trend is the creation of personal data agents. This is
an App which holds your personal Internet information, all data about
you, including your real-time Web surfing history. Some firms now pay
to obtain surfing history, and this can be handled and paid for by the
App. It is likely that clients will provide permission for insurers to use
their data via an agent and will be able to set data access boundaries.
80    
M. Naylor

Data-Focused
The current focus of insurance management is on strict cost control, so
that premiums can be kept competitive. These costs, however, are diffi-
cult to control and estimate, as underwriting estimates are created prior
to a policy issue, and are only updated at renewal time. Even then, if
clients are found to be higher risk than estimated, there can be legal or
regulation issues around substantial premium increases or policy refusal.
Client data is currently sparse and fuzzy, with contact only at renewal
or claim time, and clients are grouped into broad categories. In general,
despite being data focused, insurers currently know very little about
their clients.
The new era of real-time big data via telematics means that insurers
will have a qualitatively different quality of data on clients. They will
be able to price far more accurately, and dynamically change prices, if
initially-set risk parameters are exceeded. The big data sets will enable
insurers to more accurately understand causal factors and create predic-
tive models of risk. They will be able to offer clients attractive premiums
if they meet certain behavioral conditions.
The impact of telematic big data on pricing is a revolution. Currently,
insurers use actuarial-based statistical algorithms based on compiling
past data and event occurrences to forecast annual event probabilities.
The new approach will be based on the structural drivers behind events
as well as any conceivably related data. By being able to examine real-
time structural data on casual factors, insurers will be able to price risk
very finely, as well as customize risk to individual clients and adjust
premiums on a real-time basis. Insurers will have to create systems to
track risk in real time and adjusting premiums or reserves - for example,
tracking the spread of infectious diseases, tracking spending on health
or unhealthy food products by customers, tracking weather risk, or
tracking trends in burglary by street and house type. The sheer size of
this data means that the job of tracking and responses to clients need to
be primarily software based, with human staff overseeing. SAS (2013)
argues that insurers need to move to a predictive claims process rather
than a reactive, thus enabling smaller liquid reserves and higher profits.
 4  The Impact of Disruptive Technology    
81

This requires insurers to reimagine themselves, to change their vision

of their business from being one of providing insurance products to
being one where they are data companies with real-time links to custom-
ers, specializing in personal products and services. This will enable insur-
ers to expand their product range to provide real-time data-based services
to clients. Because big data will provide highly accurate predictive indica-
tors, insurers will move from compensation after an event to advice and
warnings on how to prevent an event. Examples would be data on traffic
congestion, parking availability, weather, relevant retail discounts, health
events, lifestyle tips, and using that to communicate alerts and helpful
tips to clients. The aim of these may seem initially to be to reduce client
risk, but the real aim would be to integrate the company into customer
lives. An insurer could then brand itself as handling a customer’s lifecycle
management, with the aim to minimize the events which lead to claims.
Another area is reducing client churn. Currently, insurers have lit-
tle idea what occurs inside client’s heads when they decide to renew or
not renew. Does sending a client an incentive to renew reduce churn
because the client is attracted to the incentive or increase churn because
the client is reminded that they can switch? What parts of the brochure
does the client who renews read? What behavioral characteristics distin-
guish those who renew from those who don’t? What types of interac-
tions with friends or what online reviews decrease policy churn? What
style of channel contact most appeals to that type of client? Which style
of contact repeals a type of client? What type of non-clients could be
influenced by an e-newsletter? Which members of a particular social
network are most influential in the buying decisions of other members?
These persuasive modelings have proven success in diverse fields, but are
little used in insurance due to inadequate data.
It needs to be remembered that by itself big data is useless and over-
whelming - the key is the ability to extract useful customer insights.
Often this can mean creating new methods of sourcing data, of stand-
ardizing data, of understanding what outcomes are required, of analyz-
ing the data, of integrating data from multiple sources, and of using the
results to guide decision making. There are two main reasons for neg-
ative outcomes when using big data: (i) unreliable or unsuitable data,
often provided by external sources and not properly incorporated into
82    
M. Naylor

the organization’s own environment and (ii) the analytics team missing
an important component of data because they didn’t fully understand
the business situation. Thus, managers are needed who both under-
stand the new data analytical methodologies and who understand the
­insurance business.
This will require heavy investment in physical IT capacity, in big
data analytical staff, and in creating new management decision-making
procedures to make best use of the new insights. There will need to be
a willingness to source expertise and software from outside the com-
pany. The overwhelming mass of data also means that high-level data
presentation skills, via the use of dashboards, are a vital new area.
Improving the quality and usefulness of data is vital. BCG (2016)
argues that the most important steps are;

1. Be clear about the question: Insurers need to be clear about the poten-
tial new uses for data, assess the benefits or costs of the new use, and
then determine exactly what issues arise. For example, how can exist-
ing high-value-customer retention rates be increased? What deter-
mines client stickiness? Then assess what issues arise, like regulatory,
and consider the uses for the data in terms of technical, organiza-
tional, and data stewardship feasibility.
2. Determine the necessary types and quality of data: Seizing an opportu-
nity may require multiple sources of data, both internal and external,
which may be in differing formats and quality. This has to be assessed
along multiple dimensions: (a) Validity - the degree to which the data
confirms to logical criteria, (b) Completeness - the degree to which the
required data is available, (c) Consistency - the degree to which the
data is the same in definition, rules, format, and time, (d) Accuracy -
the degree to which the data reflects reality, (e) Timeliness - the degree
to which the data reflects the most recent information.
3. Define clear targets for improvement: Normally, some aspect of the
data types will be missing or suffer quality issues. A gap analysis
needs to be undertaken to determine the baseline and the target for
each type of data. The benefit vs. cost of improvements needs to be
analyzed. Real-time dashboards can determine quality gaps.
 4  The Impact of Disruptive Technology    
83

4. Determine the causes of bad data: Poor quality data can arise from
many causes, whether gathering, processes, staff management, or
technology. Management incentives may have to be changed as col-
lection or processing staff may have more incentive to input speedily
rather than checking quality. The cost of correcting these needs to be
weighed against the lost opportunities elsewhere in the business due
to faulty data.
5. Assign an internal business owner to data sources: Data must be ‘owned’
to become high quality. Someone needs to be assigned who can over-
see all stages of the data flow and creating end-to-end information
models. These models will include the master data, the transaction
data standards, and the metadata. The data owner is not the same
role as a data quality officer, as they focus on business deliverables
and benefits rather than technology.
6. Scale what works: It is important to start small and quality check new
systems. However, too many data projects cherry-pick high quality data
and the best staff for trials, leading to projects floundering when they
are applied systems-wide on normal data by normal staff. The checking
will be as much about changing management processes and incentives
as about technology. Flexibility and innovativeness need to be built in
so the company can quickly move onto the next transformation.

CII (2012) points out that two potentially large problems exist with the
use of large complex data analytical systems. One is that unseen insta-
bilities can occur within large interconnected systems and led to insta-
bility and sudden crashes. The second is that the systems only predict
based on existing data, they cannot predict scenarios which have not
yet occurred within the data period - the Black Swan problem. The lat-
ter has the prime cause of Lloyds near bankruptcy in the 1980s and the
recent Global Financial Crisis. To combat this, managers of complex sys-
tems need to build in stability control break points, run extensive sce-
nario analyses, and understand client psychology. It is well known, for
example, that insured people behalf differently than non-insured, and
therefore loss rates for new products may well be higher than expected.
Data managers will need to be twinned with managers who have exten-
sive client experience and a grounding in insurance behavioral theory.
84    
M. Naylor

Insurance Rates
The use of big data to set individualized rates and premiums with the
associated improvement in customer’s risk behavior will have the effect
of reducing and spreading insurance risk. This is shown in Fig. 4.2.
For the majority of customers, the change will be beneficial as their
insurance premiums will drop and they receive useful feedback on risk
reducing activities. There will, however, be an increased group whose
premiums will rise. This may be because they: (i) engage in risk increas-
ing activities, or because (ii) they have genetics or lifestyles disadvan-
tages which are outside their control. Insurers or insurance regulators
will need to ensure this latter group have arrangements made to guaran-
tee that they receive appropriate insurance cover. This is discussed in a
later chapter.

Safer Lifestyles
The combination of many of these changes, together with advances in
medicine in areas like genetics and organ printing, means that lifestyles
will be a lot safer. Car accidents will largely disappear, houses will not

Number Current
of
customers
Future

area of
unaffordable
insurance

Low risk Average risk High risk

Fig. 4.2  Changes in the spread of insurance rates. Source Author

 4  The Impact of Disruptive Technology    
85

burn or be burglarable, people will not die prematurely as often, etc. As

a result, many of key pillars of the current insurance industry will either
disappear or face sharply reduced premium flows. This is complex, how-
ever, as reduction in premiums and increased flexibility will potentially
expand markets, bringing in currently un- or underinsured clients, as
will be explored later.

Robo-Advice
A new but growing trend is using software to provide financial and
insurance advice based on a limited financial questionnaire. This advice
is nicknamed ‘robo-advice’, is already provided by firms like Future
Adviser, and is being introduced by the major index fund provid-
ers. Using the example of the rate of increase in the sophistication of
accounting software, indicates that within 5 years it will be increasingly
hard to tell the difference between humans and software in terms of
advice for routine clients. While current financial robo-advice is crude
it is improving fast and the sophistication of robo-advice will rapidly
improve once artificial intelligence programs are applied. One of the
key aspects is thus how fast robo-advice software can include details like
Federal and state taxation, and region-specific investment advice as well
as up-to-the minute news.
This type of personalized financial and insurance advice will suit sit-
uations where the clients have no unusual or complex features so that
standardized solutions can be offered, which suits the least wealthy 60%
of clients. The advantage of robo-advice is that because of its very low
per unit cost, it can be offered to run-of-the-mill-clients for a very low
fee. Asset managers like Charles Schwab are already introducing these
systems to deal very cheaply with clients with as little as $5000 in funds.
Currently, the combined assets under management in the USA via
robo-advisers are less than US$20B, against US$17T for traditional
managers, and no individual robo-adviser has yet to reach the size
required for sustained profitability. Given the capacity of robo-advice
to continually improve, however, while continually and substantially
cutting cost to clients, there is no reason, however, to use the current
86    
M. Naylor

size of the market as an indicator of the future size. US fund managers

who use robo-adviser are currently doubling their funds under manage-
ment every few months. Even if this slowed to an annual doubling of
demand this would see robo-funds-under-management exceed human-
advised funds within a decade. In 2014, US$290M was invested by
venture capitalists into robo-advice start-ups, which has also been more
than doubling every year. Robo-advice is not a fad; it is the future of the
majority of financial advice.
Most firms offering robo-advice include it within a Web site which
also offers educational videos, trade-execution, portfolio management,
investment news, as well as tax advice and statement filing. Once ava-
tars are added, which can offer video advice and respond to standard-
ized questions, it is easy to envisage that average income retail clients
will see little reason to seek out a human adviser for value-added advice.
Financial advisers are unlikely to be concerned about the migration of
the average income client to robo-advice as they offer too little potential
fee income to make it profitable to spend much time advising. Banks
and other large financial institutions may offer robo-advice to these cli-
ents as a loss-leader, to attract clients to take out mortgages or transfer
investment funds.
The structural issue is that while financial advisers may retain a high
share of baby boomers who are used to the personal touch, the mar-
ket share of robo-advise fund managers is particularly strong amongst
the under 30s and the millennials. While these younger, computer
savvy clients are currently unlikely to be wealthy enough to interest an
adviser, the danger is that when they do acquire wealth in a decade or
so, advances of robo-advice technology, combined with their decade of
using robo-advisers, will mean that it will be difficult for human advis-
ers to attract them.
Another likely market which advisers may be worried about los-
ing are affluent individuals who are highly computer literate and feel
uncomfortable discussing financial affairs with a person. Given that
these may include high-net-worth individuals, it will be difficult for the
average adviser to attract them unless they set up a very advanced Web
site, with the added capacity for personal advice for those with complex
needs.
 4  The Impact of Disruptive Technology    
87

Given that the key skills involved would be customer service, brand
trustworthiness, and quality software, it is also likely that software or
Web-firms like Google could easily add these robo-advice sites to their
existing services and have a competitive advantage over those existing
financial service providers who suffer from low brand trust.
Another key skill to success in robo-advice will be analysis of the
data collected from clients. Over time data analysts will be able to spot
trends which distinguish the financially successful from the less suc-
cessful and help all clients achieve their goals. The ability to analyze
big data then becomes vital, especially when trying to ascertain which
behavioral finance characteristics determine financial success, or which
types of advice best promotes the right customer response. There will be
an advantage to financial firms like banks which have a more complete
set of client financial data, including income and spending patterns.
Small firms which create smart software, however, can also be success-
ful if they can obtain enough base data to fine-tune their models. Cloud
computing means that there is no longer an inherent advantage to large
firms simply because they can afford mainframe computers.
It is the careful analysis of this big data set which will allow robo-
advice to be tailored to more detailed client profiles so that advice is
not generalized but specific to each client. Clients can be offered advice
relating to what the software identifies as that client’s behavioral weak-
nesses. This can then be tied into insurance advice and individualized
underwriting. This can be complimented by mining of social media so
that current concerns can be identified and advice offered. Algorithms
can then be used to create predictive models of financial needs, product
preferences, and customer interactions.
Carefully constructed robo-adviser financial Web sites will thus be
able to offer education and news which is specific to a narrow set of cli-
ents and make clients feel as if the robo-adviser understands their needs,
without the cost factor of human involvement. There is psychological
evidence that investors are more willing to share personal data with an
anonymous data warehouse, as the software will make no value judg-
ments about life choices.
While robo-advice will not be able to match human advisers in
the next decade in terms of understanding the myriad financial and
88    
M. Naylor

emotional reasons why individuals either succeed or fail in achieving

their financial goals, agent-based learning software is currently being
built which will, within five years, start to approach human adviser
capacities at an exponential rate. Asset managers or dealer groups who
do not respond adequately will be left behind. Note that retaining client
trust in the privacy and the integrity of the process is vital and any pro-
viders who exhibit abuse of this trust, or suffer data security breaches,
will lose irretrievable market share.
Rao (2015) argues that there will be three generations of robo-advice
software:

1. First generation: These are currently available and are stand-alone

programs targeted at the self-directed end users. They allow users
to aggregate their financial data, create a unified portfolio, obtain
generic advice, determine portfolio optimization based on life-stage
and execute trades. These programs use simple segmentation and
portfolio optimization assumptions.
2. Second generation: These currently being trialed and will be aimed at
both end users and advisers. These allow integration between inves-
tors, advisers, and clients and are based on wrap programs. The
increased information available allows more customization of advice
and allows advisers and institutional traders to interact dynamically
with clients, offering real-time news and trading advice.
3. Third generation: These are being planned and will be available within
a five to ten-year period. These will build off advances in Artificial
Intelligence learning software, and will have the ability to improve
advice based on feedback from client behavior. These will offer life-
long personal financial planning advice by combining the mining of
big data and social media to inferring investment behavior, spending
and saving patterns, behavioral mistakes, and risk preferences, using
learning agent modeling to create a model of ‘someone-like-you’ per-
sona for the end user.

This third generation will watch how the clients react to advice and
modify advice based on this. It will monitor economic and market data
and offer investment and insurance advice based on current conditions.
 4  The Impact of Disruptive Technology    
89

By adding a Web-avatar, this advice can be delivered in a personal way

using techniques designed to be most persuasive. The software will run
scenarios and illustrate how differing portfolios will react to possible
futures. This can then easily be combined with banking and tax advice.
A key element of the third generation will be client feedback on how
they feel about the advice. The collection and use of increasingly large
databases on client reaction will enable algorithms to be created which
will improve the sophistication of the advice, making it pleasing and
relevant to clients at an exponential rate. Advances in computer voice
modulation will fine-tune how to sound the most persuasive, as will
advances in avatar face movements. The integration of the lessons from
psychology and behavioral finance will be vital to this feedback process
and is the aspect which robo-advisers will probably struggle with the
most.
It is going to be difficult for small advice firms to match the IT
investment and data warehouses of large mutual fund or banks. The key
to their survival will be the availability of high-quality software from
3rd-party vendors.

Conclusion
The insurance industry has been built on its ability to detect, analyze,
and manage risk. Capgemini/Efma (2016) argues that the insurance
sector is changing in four fundamentals ways: (i) most key risk variables,
the determinates of risk, are changing and becoming more individual-
ized, less random, (ii) in general, risk is decreasing and becoming more
avoidable, (iii) the nature of risk ownership is changing and becoming
more communal, and (iv) new techniques are causing shifts in basic
insurance business principles and models as connected technology and
risk mitigation reduce loss.
The insurance industry is in general a laggard in terms of both its
adoption of technology and in the extent to which it has been impacted
by disruptive innovators. These, of course, go together, as insurers have
not been pushed to innovate. Insurance, however, has now reached a
tipping point, where a perfect storm of technological innovations will
90    
M. Naylor

coincide to generate in the immediate future a step-up in both the

impact of disruptive innovators and the requirement for insurers to
adopt technological solutions.
The insurance industry needs to urgently start making strategic deci-
sions on how to survive. It needs to reimagine its product into an array
of currently unoffered value-added services. In most cases, this transfor-
mation will involve extensive use of external change consultants.

Dynamic Insurance
Insurance has traditionally been a static product: sold once, with a risk-
rating set largely for the life of the client or product. Clients did not think
much about insurance and insurers did not know much about their cus-
tomers. Neither the insurer nor the client engaged much with the other.
The future of insurance has to be as a dynamic product: sold in mod-
ules, with real-time risk-pricing dependent on the client activities, with the
ability of either side to add or subtract aspects of the product and respond
to temporary specials. The insurer and the client will engage on a continu-
ous basis, with the insurer having a rich and deep knowledge of the cli-
ent, and the client choosing an insurer on the basis of the warmth of the
engagement, the worthwhileness of their value-added services, and the
attractiveness of their pricing. This involves combining both pay-as-you-
use and pay-how-you-use data with real-time feedback.

References
Boobier, T. (2016). Analytics for Insurance: The real business of big data,
Wiley, UK.
Boston Consulting Group. (2016). How to avoid the Big Bad Data Trap, BCG
Perspectives, October 2.
Breznik, L. (2012). Can information technology be a source of competitive
advantage? Economic and Business Review, 14(3), 251–269.
Brynjolfsson, E. R., Hitt, L. & Kim, H. (2011). Strength in numbers: How does
data-driven decision-making affect firm performance? Social Science Review
Network, April.
Capgemini/Efma. (2016). World Insurance report.
Carr, N. (2003, May). IT doesn’t matter. Harvard Business Review, 5–12.
 4  The Impact of Disruptive Technology    
91

Castriotta, M., Floreddu, P. B., Di Guardo, M. C., & Cabiddu, F. (2014).

Disentangling the Strategic Use of Social Media in the Insurance Industry:
A value co-creation perspective, chpt 4 in Social Media in Strategic
Management: Advanced Series in Management, 63–86.
Christensen, C. M. (1997 & 2011). The innovator’s dilemma (4th Ed.). NY,
USA: Harper Business/Harvard Business School.
CII. (2012). Future risk: How technology could make or break our world.
London, UK: Chartered Insurance Institute.
CII. (2015). Big data and insurance: A conversation. London, UK: Chartered
Insurance Institute.
Equinix. (2014). Challenge to change: three-part series, Acord/Global Futures &
Foresight.
IBM. (2006). Insurance 2020: Innovating beyond old models. New York, USA:
IBM Institute for Business Value.
IBM. (2015). Capturing hearts, minds and market share: How connected insur-
ers are improving customer retention. New York, USA: IBM Institute for
Business Value.
Kelly, K. (2016). The inevitable: Understanding the 12 technological forces that
will shape our future. New York, USA: Viking.
KPMG. (2014). An industry transforming: insurers and intermediaries in a fast
changing, digital world, white paper.
KPMG. (2015). Big Data: The big factor driving competitive advantage.
Frontiers in Finance.
KPMG. (2016). Empowered for the future: Insurance reinvented, KPMG
International.
McKinsey. (2015). An executive’s guide to machine learning. McKinsey
Quarterly, June.
McKinsey. (2016). Transforming life insurance with design thinking, May.
Morgan-Stanley/B.C.G. (2013). Wearable devices: The internet of things becomes
personal. Boston, USA: Boston Consulting Group.
Morgan-Stanley/B.C.G. (2014). Insurance and technology: Evolution and revolu-
tion in a digital world. Boston, USA: Boston Consulting Group.
Morgan-Stanley/B.C.G. (2015). Insurance and technology: The emerging role of
ecosystems in insurance. Boston, USA: Boston Consulting Group.
Porter, M. & Heppelmann, J. (2015, November). How smart connected prod-
ucts are transforming competition. Harvard Business Review.
PWC. (2014). Insurance 2020: The digital prize—taking customer connection to
a new level.
92    
M. Naylor

Rao, A. (2015). The rise of robo-advisers, Insurance Thought Leadership, PWC,

June.
Rifkin, J. (2014). The zero marginal cost society: The internet of things,
the collaborative commons, and the eclipse of capitalism. London, UK:
Palgrave-MacMillian.
SAS. (2013). Predictive Claims Processing: Transforming the insurance claims life-
cycle using analytics, SAS White Paper.
Schryen, G. (2013). Revisiting IS business value research: What we already
know, what we still need to know, and how we can get there. European
Journal of Information Systems, 22, 139–169.
Summers, L., & DeLong, J. B. (2001). The ‘new economy’: Background, his-
torical perspective, questions, and speculations. Economic Review of Federal
Reserve Bank of Kansas, USA.
Target Group (2015) ‘Distribution Channels in Insurance: Understanding the
Customer Viewpoint’.
Wernerfelt, B. (1984). A resource-based view of the firm. Strategic Management
Journal, 5(2), 171–180.
World Economic Forum. (2015). The future of financial services: How disruptive
innovations are reshaping the way financial services are structured, provisioned
and consumed.
Yoder, J., & Rao, A. (2015). Insurance at a Tipping Point, Insurance Thought
Leadership, PWC, June 20.
 5
What Insurance Companies Need to Do

Management during Disruptive Waves

Christensen (1997) in his classic book points out that technological
disruption can blindside even successful companies. Commentators
often argue that only weak or badly run firms fail when faced with a
market challenge. In fact, Christensen argues that even well-managed
companies, those who scan for looming market challenges, listen
astutely to customers, and invest aggressively in new technologies, can
fail to foresee and meet disruptive innovations.
Christensen distinguishes between ‘incremental innovations’ and ‘dis-
ruption innovations,’ with the differentiating factor being that incre-
mental innovation can be accommodated within existing business
models and disruptive innovations cannot. In that sense, ‘disruptive
innovation’ is wider than technology, as non-technological areas can also
involve disruptive innovations.
Markides (2006) divides disruptive innovation into three types:
(a) disruptive technological innovations, (b) disruptive business model
innovations, and (c) disruptive product innovations. These will have differ-
ing disruption pathways. This book is has so far focused on the first type,

© The Author(s) 2017 93

M. Naylor, Insurance Transformed, Palgrave Studies in Financial Services Technology,
DOI 10.1007/978-3-319-63835-5_5
94    
M. Naylor

but disruption in the other two areas will arise as a consequence. For
example, the use of real-time data is technologically based but will rede-
fine the insurance business model from a product to a service. This will
expand the insurance market to new customers and into new products.
This implies that there will be waves of disruption of differing types.
Christensen (1997) points out that an aspect of disruptive innovation
is that it occurs in waves, with each wave having distinct characteris-
tics. The management style which suits each wave can be quite distinct,
so that companies which were successful in one wave fail to handle the
next. For example, IBM dominated the mainframe market, but stum-
bled after the emergence of minicomputers. In fact, no mainframe man-
ufacturer managed to handle that disruptive wave. An emergent firm,
Digital Equipment Corp, dominated the minicomputer market, along-
side Data General, Prime, Wang, Hewlett-Packard, and Nixdorf. These
emergent firms, in turn, failed to handle the change to desktop personal
computers, whose market was dominated by the new emerging firms:
Apple, Commodore, Tandy, and a revived IBM. Of these new firms,
only Apple managed to survive the shift to portable laptops and recently
to tablets and phones. Firms whose expertise was in very different sec-
tors, like Samsung, have entered the mobile market.
This disruption outcome is similar whatever the pattern of the dis-
ruptive waves, whether the change arrives in a rush or is slow, whether
the changes were complex or simple, or whether the innovation is an
extension of existing technologies or new.
Christensen notes that the blindsided firms were, in the main, well
managed, with Digital Equipment in particular praised in the Peters
and Waterman (1982) classic ‘In Search of Excellence,’ just two years
before it ran into trouble. Christensen argues that the basic problem
is that well-managed firms follow patterns of behavior which are best
practice only in terms of the current environment, not in terms of the
new environment; each innovative stage has differing best-practice man-
agement requirements, and it is often the best managed firms, those
who are sure of their excellence, which have trouble recognizing the
need for transformation. It is precisely their excellence at adapting to
the old environment which causes issues handling disruptive waves.
 5  What Insurance Companies Need to Do    
95

Christensen thus argues that when facing an uncertain future, with

multiple possible waves, most of which will end up insignificant but
several transformational, a different style of management is required.
Given that customers will have even more limited knowledge of the
future, management may be justified in following counterintuitive
methods like not listening to customers, in investing in lower perfor-
mance products with lower margins, and in aggressively pursuing small
rather than large markets.
Christensen argues that these aspects occur because disruptive tech-
nologies are typically cheaper, simpler, smaller, and easier to use than
existing technology. They are initially aimed to lower-end or more for-
giving customers. Any survey of existing customers tends to lead to the
new product being scorned and rejected. Imagine asking mainframe
users in the early 1960s about their impressions of the new-fangled
minicomputers!
This dichotomy occurs because the disruptive innovation is typically
evaluated by its new users by a different metric than customers of the
existing technology use. For example, existing market customers may
prize speed, while new market customers prize flexibility. To both cus-
tomer sets, switching to the other product reduces usability. This means
that customer surveys of new innovation product attributes will have to
involve customers who not currently are in the market or have not yet
considered what they want.
In terms of existing metrics, disruptive technologies, therefore, tend
to underperform compared to existing technologies, so are ignored by
market leaders who are focused on fine-tuning their market-leading
products. However, because the disruptive technologies provide more
customers with an experience closer to what they actually want, even if
they are unaware of it, disruptive technologies progress at a faster rate
and rapidly end up more performance competitive. Indeed, the new
product often creates new markets, which could not have been imag-
ined prior to its introduction. It needs to be noted that the initial cus-
tomers are commonly technophiles who may value different attributes
to latter users; they are less price sensitive, more fault tolerant, and more
demanding of advanced features.
96    
M. Naylor

Adner (2002) shows how differing structures of demand along two or

three preference attributes create three possible outcomes: (i) The two
technologies exist in isolation in different markets, (ii) the two tech-
nologies converge in the same market and compete, and (iii) the newer
technology disrupts the old. All three of these outcomes are likely in dif-
ferent parts of the insurance sector.
Christensen (1997) distills his ideas about the problems of waves
of disruption into five principles. These apply to each wave of disrup-
tion, so that yesterday’s emergent innovator becomes today’s struggling
incumbent.

1. Customers and investors control resources: Within successful compa-

nies, managers are often under the illusion that they control com-
pany resources. They normally have well-developed, rapidly working,
systems for killing ideas which customers don’t want. They thus find
it difficult to allocate resources to new technologies which custom-
ers currently don’t seem to want, and then find it hard to reallocate
resources when customers suddenly decide they do want the technol-
ogy, especially if these are customers from outside the existing profile,
and the market is seemingly small and low margin.
2. Large companies find small markets difficult: Emerging technologies
start small and offer few enticements to existing large market lead-
ers. Thus, incumbents typically decide to ‘wait-and-see.’ Given, how-
ever, that there are typically significant first-mover advantages, lack of
initial engagement can condemn incumbents to playing eternal catch
up; especially, if the disruption involves a new type of customer who
the incumbent does not understand.
3. Markets that don’t exist can’t be analyzed: Sound market research and
good planning are embedded parts of market leaders. These tech-
niques fail when faced with disruptive innovations as the ultimate
use and extent of products are not established and customers cannot
know if they appeal. Who would have guessed, pre-mobile phones,
that people would prefer to play games or take photos with their
phones?
4. Company processes are more inflexible than people: Good companies
often assume that staff who are capable within current structures
 5  What Insurance Companies Need to Do    
97

will deal with a new task. Yet, the disruptive innovation often can’t
be dealt with until company processes drastically change, and then,
the staff may not be as capable within the new environment. A new
business model is required, which may cannibalize the existing profit
stream.
5. Disruptive innovations typically exceed customer capacity: Within dis-
ruptive waves, typically technologies will include multiple features
which will be ignored, until the market can absorb previous changes.
These new features are often so innovative that the market fails to
understand that a product can be used within an entirely different
setting. Therefore, products can seem to initially under-perform, fol-
lowed by an extremely rapid change, as customers’ mind-set changes.
Companies often inadvertently ‘over-satisfy,’ and end up too up-mar-
ket, opening up a space at the lower end.

Christensen argues that what these principles result in is the ‘innovators’

dilemma’—good companies begin their descent into failure by listening
to their best customers and then aggressively investing in the products
and services which their most profitable customers want. The innova-
tion maximizes a different metric, one which does not appeal to existing
customers, but does to ill-serviced or non-serviced customers, customers
who do not suit the incumbent’s existing business model. Managing the
right way ends up being wrong.
It is common for individual teams within incumbents to be aware
of the disruptive innovation and to have developed their own version
before the disruptor brought their product to market. Yet, it is uncom-
mon for the senior managers of incumbents to approve the transition in
time.
For innovative unproven technologies, an issue is that disrup-
tive technologies will at the start typically seem unprofitable, have no
demand from existing customers, be technologically crude, and gener-
ally not make sense as a rational place to invest resources. Financial and
marketing analyses will thus show the innovation as a near certain fail-
ure. Given that 90% of possible innovations will come to nothing, it
is near impossible for a manager to successfully argue that an incum-
bent should transform itself because of a possible disruptive innovation
98    
M. Naylor

before that innovation has established itself in the market. Research

shows that predicting which 10% of innovations will succeed is near
impossible.
Christensen argues that incumbent delay also occurs for innovations
which are the obvious next step, innovations where incumbents may
have a performance advantage. This is mainly due to the fact that the
innovation is inferior at the start, so the most profitable existing cus-
tomers do not typically want the new innovation. The current product
may be quite profitable and have a strong growth path. Feedback from
suppliers and customers within the wider network will thus confirm the
existing business model. It is therefore hard to visualize a new product
possessing a superior growth path.
Disruptive innovations thus enter by servicing a market the incum-
bent handles badly, who value a different metric to the core customers.
The disruptor then improves its performance on all metrics and starts
to nibble at the lower-end incumbent customers, those over-serviced by
the main metric. The incumbent finds this hard to respond to, as alter-
ing their product to service lower-end customers would alienate core
customers. Eventually, the disruptor ends up servicing the mainstream
market, with the incumbent left with a minority. The recommended
response is for the incumbent to create its own separate group to take
on the disruptor. This, however, creates a series of new issues, which will
subsequently be discussed.
McGill (2008) argues that financial sector companies have survived
prior disruptive shocks better than those from other sectors because of
their sector breadth and the advantages of utilizing these within a large
vertical corporation. I would argue, however, that because the looming
disruptive waves involve the transformation of data use, the core skill,
nearly every sector of insurance will be impacted, so breadth will not be
a defense.
Management during disruptive changes requires a different culture
and staff than during incremental changes. Because disruptive technol-
ogies create new markets, they tend to find their main uses and new
customers over time as they evolve by trial and error rather than as the
innovators initially visualized. Therefore, with waves of disruption the
new norm, flexibility to follow startling new changes is becoming more
 5  What Insurance Companies Need to Do    
99

important than careful management or efficiency. Improving the ability

to assess possible disruptive innovations is critical.
Johne and Davies (2000) examined the upheaval in the UK general
insurance market caused by online distributors and found that CEOs at
surviving firms used more aggressive responses to the disruption threat
than CEOs at non-survivors. In particular, there was more destruction
of existing control structures and the removal of middle managers who
were unable to adjust fast enough. This was followed by support of staff
involved in innovation and experimentation.
McKinsey (2017) shows that the digitization reaction has barely
started in most industries and yet has already reduced revenue by
an average of 7.3%. Importantly, though, the digital leaders are three
times more likely to generate market-leading profit margins than digi-
tal laggards are. The main impact of the looming waves of disruption is
therefore likely to require survivors to position themselves among the
technological leaders or at least the fast followers.

Existing Insurer Capacities

Despite the looming issues around disruptive technology which I have
described, KPMG (2015a) found that 51% of US insurance execu-
tives knew little about the issues surrounding IT and were not predict-
ing major changes. Only 22% of insurers had thought about the issues
or were making plans. Company CIOs are significantly more worried
about the lack of action than CEOs. PWC (2015b) found that change
management was only the sixth top concern, and that most respondents
had a very limited understanding of the issues.
Accenture (2013) found that 95% of insurance CEOs surveyed were
not sure that they had the right operating model for today’s world, let
alone the future. Bain (2013) argues that while insurers are rolling out a
range of digital initiatives, in most cases these involve a string of uncon-
nected initiatives layered onto a legacy IT system and a traditional
mind-set. Very few current insurers even understand the coming scale of
change required, let alone are starting to restructure their businesses to
the extent which will be required.
100    
M. Naylor

While this ignorance started to change around 2015, with KPMG

(2016) finding that insurance CEOs now regarded technological dis-
ruption as their major future issue, Bain (2015) found that about 50%
do not have a plan for digital transformation and 60% of those with
a plan are missing key elements. Few of the surveyed insurance CEOs
had confidence that they had a workable transformation plan. None
of the companies who have achieved some success have managed to
achieve best practice in more than in a few dimensions. You cannot just
add a Web site. Very few insurers have tried to reimagine their future.
Analysis, thus, shows that the insurance industry is fundamentally
unprepared for the future, despite the fact that the insurance landscape
of 2026 will look very different to what it looks like now.
KPMG (2016) argues that success in the insurance industry in the
future will not come from simply tweaking the Status Quo; insurers
will need to change virtually every part of their business if they hope to
not just survive, but thrive. They point out that insurers have been try-
ing to transform themselves for decades, firstly to respond to big enter-
prise systems and then to the internet. Despite this effort and expense,
KPMG concluded that little has changed, especially when compared to
sectors like banking or financial markets. KPMG does not expect this
inability to react to change. KPMG argues that this is because insurers
require fundamental transformation and this is very difficult.
KPMG (2016) found that while insurance CEOs are now aware of
the threat, most are still not aware of the fundamental extent of the
changes required. For example, KPMG found that while 70% of CEOs
saw the need for fundamental change, less than a quarter expected their
operating model to be disrupted by changes in customer behavior. This
is either ignorance or self-delusion, because it is clear that a customer-
centric model is required. KPMG found that respondents had little idea
of what a ‘customer-led business model’ meant, despite many consult-
ants pointing this out. Customer behavior should actually be the inspi-
ration behind insurers’ efforts to reinvent themselves.
KPMG argues that while a quarter of respondents said they have
started a major transformation in the past 2 years, nearly all of these
efforts have been inadequate and superficial. They note that there is
a big difference between conducting a transformation initiative and
 5  What Insurance Companies Need to Do    
101

actually achieving a transformational outcome. The reality is that insur-

ers will need to achieve change at a far more fundamental level if they
hope to reinvent themselves for the future. To truly ‘reimagine,’ their
business insurers will need to transform business and operating mod-
els from the ground up. Any attempt to survive by ad hoc changes or
minor tweaks will fail in the face of external disruptors. It is the purpose
of this book to make it clear the extent to which fundamental reimagin-
ing is required.

Services instead of Products

Insurers currently offer what is inherently a service, a promise of future
compensation. Yet, insurers have traditionally visualized themselves as
selling a fixed, highly specified product, with customers only offered a
limited range, with no opportunity for modification, or day-to-day
interaction between insurers and customers. Policies were delivered in
fancy brochures, and insurers built impressive buildings, in an effort to
create an impression of solidness. This packaging of a promise as a phys-
ical product was useful when nearly all other types of consumption were
physical products, so that consumers found it hard to visualize buying a
flow of non-physical services.
By 2010, customers started to move on, and getting used to buying
formally physical products like books or music as flows. Borders Group
saw themselves as selling physical books within physical stores and were
pioneers of software-based inventory control and sales forecasting. They
saw the new entrant, Amazon, as merely selling physical books from a
physical warehouse aided by a sales Web site. Border thus forecast mini-
mal disruption, which they could cope with by opening their own Web
site.1 They failed to understand that because Amazon could make profit
from sales of other products once customers were on their Web site and
could add on services via e-book feedback, as well as attracting readers
via user-generated reviews or blogs, Amazon had little need to make a

1Surprisingly, they outsourced the Web site running and book shipping to Amazon.
102    
M. Naylor

profit on physical sales. Amazon thus rapidly undercut Border’s prices

and drove them bankrupt. Reading via an e-reader which allows annota-
tion and commenting becomes a social activity, in a way physical book
can’t.2
Insurers are at threat of being left behind in the same way.
Incumbents are tending to approach the IT challenge by modifying
their existing product-based model and by merely adding on digital
sales avenues. Both of these fail to understand the utter transformation
of their environment and fail to ignite the imagination of their work
forces to start creating a range of valuable flow-based services. Instead,
insurers need to reimagine insurance as a series of related, customer-
modifiable, real-time flows. Given that Amazon is a potential insurance
disruptor, insurers need to stop reacting like Borders did.
This transformation could be done by arranging insurance as a
series of modular policies, which customers can pick and choose from,
with adjustable attributes. Each module can come with omni-channel
advice on use as well as customer-created reviews in ‘what a user-like-
you’ thinks. Policies will be linked to telematics, and users will receive
real-time feedback on how their activities impact on their risk level.
For example, health insurance customers could receive free genetic
tests, which when added to telematic feedback will then aid in person-
alizing medication, as well as creating personalized health articles and
tips. Users can be linked into groups of similar users, who can share tips
on optimizing health outcomes. Higher-risk customers can have their
health metrics tracked in real time so they can be notified of danger-
ous trends. This will transform insurance from a boring, depressing,
one-off purchase, to an engaging series of directly useful value-added
services, aimed at optimizing a customer’s life cycle healthcare man-
agement. During this process, insurers will gain an extremely rich data
series which links client activity to health outcomes, as well as a detailed
understanding of the best kind of feedback and inducements required
to generate desired health outcomes.

2E-books have, however, been hobbled by publisher restrictions on extraction and rearranging, in

a way which music has not.

 5  What Insurance Companies Need to Do    
103

Automated Underwriting
The first visible internal change will be within the underwriting sector as this
is already transforming. SwissRe (2013) found that 56% of both US and
European insurers were started to use automated underwriting. However,
very few have moved very far - most current underwriting is still paper and
human based, and automated systems are still based on old business models.
Only a few are thinking about dynamic underwriting. This lack of urgency
will have to dissappear as the waves of disruptive technology build.
It is argued by some commentators that incumbents will have a
substantial advantage because they have decades of experience in run-
ning risk assessment models based on historical data. External firms
have neither the data nor the analysis capacities, so will have to learn
these. I would argue that this conclusion is faulty, for two reasons.
Firstly, the basis of underwriting will transform to a new paradigm of
big data algorithms, and technology firms have more experience with
these than insurers do. Secondly, data is not an issue, historical data can
be purchased and telematic data will overwhelm it. Unless incumbents
immerse their underwriting teams in the new paradigm, they will fall
irretrievably behind within 3 or 4 years. They need to start by trans-
forming existing data from all areas of the firm into a common format,
which can be data mined.
The best industry expert estimates are that by 2018 up to 90% of
all underwriting by insurers will be handled by automated underwriting
software. This will occur in conjunction with software-based applica-
tion/renewal forms which feeds answers into the adminstrative as digital
data. They will also be integrated with the collection of client data from
as many alternative sources as possible and increasingly integrated with
client telematics. This can then be integrated into all other parts of the
business.
This will eventually have the impacts of: (i) allowing premiums to be
set on a client-specific basis, (ii) allowing premiums to be dynamically
adjusted as telematics indicate client behavior is differing, (iii) give the
insurer a far deeper data set and a far deeper understanding of client
characteristics, and (iv) give the insurer a far deeper understanding of
104    
M. Naylor

risk causation. This deeper understanding can be fed back into improv-
ing the client questionnaire and underwriting modeling. The Geneva
Association back in 2013 argued that a paradigm shift from historic to
predictive risk assessment was necessary.
However, the most important result of automated underwriting will
be to reduce the marginal cost to near zero, as explained earlier. This
reduction in cost and the resultant surge in demand will allow a large
expansion of underwriting, so that most clients are individually under-
written using a wide array of variables. Until this first automation
underwriting step is undertaken, the rest of insurance can’t transform
itself, as dynamic insurance depends on near-zero cost underwriting
joined to automated straight-through processing. The key element will
be rules around when human oversight is required. Note that the rapid
expansion of underwriting means that jobs are unlikely to be lost over-
all. However, surviving underwriters will need the IT skills to under-
stand the algorithms and usefully build on the insights provided.
The grouping of clients into risk pools has been the basis of under-
writing. In the future, there will not be explicit pools, but an overall risk
algorithm. This will involve a range of risk factors, with the weighting
of each differing for each client. Client grouping is an administrative
technique to make it easier to set rates. Within a world where the cost
of individual risk rating is cents, there is no need to group. The algo-
rithm may group clients but as a consequence, not a management tool.
This will also an impact on capital reserving and reinsurance, as scenario
testing with the algorithm will determine risk capital, rather than the
creation of preset client pools.
PWC (2015a) argues that a side effect of this automated underwrit-
ing will be the commoditization of insurance provision, as software-
based firms enter. In the long run, this will drive down profit levels to
the low levels normal in retail sectors. PWC argues that insurers then
will be forced to gain their profits from the value-added service activities
which they can package alongside the low-margin insurance.
There are many issues to be sorted out. The most crucial is to trans-
fer the skill and experience of existing underwriters into programmable
rules, via true/false flow diagrams. The industry saying that ‘underwrit-
ing is both an art and a science’ needs to be carefully considered as only
 5  What Insurance Companies Need to Do    
105

the science can be programmed, and thus areas where exemptions exist
need to be defined so decisions can be defaulted to humans.
The use of online applications means that questions can be interac-
tive, with question areas expanded or contracted dependent on prior
answers. Thus, there should be little need for additional questions, and
fewer questions can be asked of clients who have no current conditions.
These questions can be backed by explanations, either in text or short
videos, as well as popup Skype boxes, so a staff member can be con-
tacted within the Web page. This will provide data feedback on which
terms or areas cause most concern, which will add in the Web site and
policy redesign.
The use of big data will bring up many new correlations and useful
insights into client behavior, which go well past current insurance score
systems. These new insights need to be thought through and clarified
as they may have to be explained and justified to clients, to regulators,
and to courts. For example, it is well known that credit scores correlate
with insurance risk, but an early insight from data is that clients who
are more likely to default on their consumer loans are also more likely
to die prematurely. Reasons for this can only be speculated, so whether
it is causal or just a temporary aberration is unsure, and it is unlikely to
hold for all individual cases.
The greater depth and width of data used in new-style underwrit-
ing means that new areas can be explored. Examples: (i) which sales
agents or sales practices are better at attracting higher-quality clients or
encouraging clients to disclose, (iii) which questions on the application
form add value, or which questions not currently asked need to be, (iii)
whether different channels have differing claim or non-disclosure rates,
(iv) what types of social media data are useful, (v) what is the compara-
tive reliability of conditions clients disclosed vs medical evidence for dif-
ferent claim types, (vi) how do differing types of financial data relate
to claim rates, (vii) which types of clients non-disclose which types of
conditions, in which channel, in which question format.
KPMG (2015b) argues that insurers need to: (i) clean up existing
data; (ii) develop a data governance model which allows flexibility and
innovation; (iii) create an enterprise data management function whose
focus should be on improving data flow and usage; (iv) create a culture
106    
M. Naylor

of experimentation and enterprise, as the correct answer will not be

immediately evident; (v) focus on new ideas rather than what is already
known about data, especially new areas of value creation; (vi) focus on
profit-driven analytics by trying to eliminate the irrelevant 99% of data;
and (vii) focus on asking interesting questions so useful new insights
can be discovered, particularly questions which can’t be answered by
existing data.
The application and underwriting process will be different under a
dynamic setting. It needs to be noted that the digital generations will
expect instant premiums on application. Insurers will not be able to ask
clients to wait, while decisions are required from managers. This can
be solved in several ways: (i) The dynamic ability of online question-
naires means that precise questions can be asked, and higher-risk appli-
cants identified; (ii) client disclosures at the time of application are not
as important, as insurers can check these subsequently by telematics, or
via other data sources; (iii) multiple channels can be integrated, so that
if applicants answer a question indicating higher risk, then a Skype or
WhatsApp box can open allowing a human underwriter to ask addi-
tional questions and visually assess the applicant; (iv) insurers can link
in real time to medical or doctor databases; and (v) applicants can use
such additional channels to ask about questions. It is well established
that the reason for 95% of client non-disclosure is not fraud but igno-
rance, or not understanding what information is required. Thus, being
able to open a box and ask a direct question will clear up the majority
of current confusion. Insurers can use innovative new ideas like an App
which uses a mobile phone camera to analyze blood.
This dynamic process will then allow non-disclosure laws to be
amended to ensure that the onus is on insurers to ask the correct ques-
tions at application time and therefore allow insurance to be guaranteed
at claims time for all cases except for clear and deliberate client fraud.
It will also allow a simplified and rapid application process for 90% of
clients, in-depth questioning of the difficult 10%, while giving insurers
a more complete risk profile and allow individualized underwriting.
An automated data-driven approach to underwriting will allow insur-
ers to differentiate premiums in real time. Examples of this would be
offering different rates when a car is being driven to when it is parked,
 5  What Insurance Companies Need to Do    
107

or a differing rate if a house is occupied than when the owners are

on vacation, or varying rates by a customer’s adherence to an exercise
regime. Privacy issues would be a concern but could be overcome by the
insurer establishing a high level of trust.
The extra dimensions added to insurer information on clients will
allow deeper insight into client risk and a qualitatively different level of
management reporting and decision making and therefore better profit-
ability. The ability to risk-rate clients in real time by activity will allow
insurers to proactively engage with clients on a regular basis and pro-
vide suggestions to clients on how to reduce risk or lead healthier lives.
Insurers could, for example, reduce health insurance premiums for cli-
ents whose exercise telematic records more than a set level of consist-
ent physical activity over the last 3 months, and could conversely warn
clients whose activity level drops.
A vital part of dynamic insurance will be the creation of a client-indi-
vidualized ‘mobile risk profile.’ This is an algorithm which defines a cli-
ent’s risk characteristics based on telematic feedback. This moves with
a client, so regardless of the car they drive, health activity they under-
take, country they are in, or insurer they use, the profile moves with
them. It is thus tied to the person and not any physical object. It could
link to the person’s digital agent. The profile can interact with telemat-
ics or their phone so the client gets feedback on how to improve their
risk profile, listing helpful activities, and then, dynamic changes in risk-
based premiums can be displayed if those activities are undertaken. I
call this ‘real-time dynamic risk assessment.’
These types of mobile risk profiles will lead to the possibility of ‘life-
style insurance,’ whereby insurers actively and positively engage with cli-
ents about their life choices and additional activities like incentives from
related companies within their ecosystem, for example, discounts from
sports clothing stores, or articles from health writers. This will help to
solve the existing issues of non-communication with clients and trans-
form the public’s attitude to insurers. Insurers will focus on providing
useful services which reduce risk.
The power of the information provided by these new analyti-
cal systems will allow a movement away from the current insurer
focus on using structured data to make tactical decisions toward using
108    
M. Naylor

unstructured data to make strategic decisions. This may require the crea-
tion of artificial intelligence systems so proactive predictive modeling
can be used, including extensive scenario analysis. These will form the
heart of future insurance management and will offer a major competi-
tive advantage to those insurers able to integrate their insights into man-
agement decision making. Use of real-time dynamic risk assessment will
allow management to get weekly or even hourly trends in loss likeli-
hoods. Note this type of dynamic risk assessment is not possible unless
underwriting is as automatic as possible so that marginal cost is near zero.
EY (2014) argues that to do this, however, underwriting has to shift
away from a focus on internal cost efficiency and refocus on improv-
ing the client experience. Thus, the skill set for underwriters will need
less computation, which will be automated, and more understanding
of client relationship skills. They will be involved in higher-value activi-
ties like account planning, solution development, agent partnering, and
nuanced risk assessment and decision making. They will be able to, for
example, use their new analytics to help agents and advisers select more
profitable clients and achieve higher prospect conversion rates.
EY (2014) notes that before 1990 underwriters were the heart of
insurance companies and exhibited a broad range of skills, and were the
main source of senior managers. However, since then finance executives
have provided most senior managers. One reason for this is that as core
finance functions were computerized, finance steadily expanded its skills
to include planning, analysis, and strategic thinking. Underwriting,
conversely, narrowed down its skill set to focus mainly on the binary
transactional process. This has to reverse. EY (2014) argues that the
risk focus of underwriters and their organizational position make them
better placed to manage the required transformation of insurers than
finance executives as underwriters are naturally better suited to coordi-
nating the different insurance sectors to ensure that the right clients are
selected and managed through their insurance lifecycle. To fulfill this
role, the culture of underwriting has to undergo a similar change to that
which finance went through in the 1990s. I would argue that insurers
who do not rapidly get near the leading edge of innovation of under-
writing innovation will fall irretrievably behind.
 5  What Insurance Companies Need to Do    
109

The ideal future underwriter will need to possess technical skills:

with risk assessment, with big data, with software structure, with cus-
tomer relationship management, with customer profit segmentation,
with optimal omni-channel presentation, with using the relevant eco-
system to enhance the relationship between dynamic premiums and
incentives provided to clients to induce them to make appropriate deci-
sions, with innovation creation, and with strategic development. This
needs to be intuitive as well as technical. The basis of all of these areas
will be the new sources of information provided by big data analytics
and telematics.
Ideal underwriters will need to leverage their access to this infor-
mation as well as their understanding of desirable customer types to
improve the quality of lead identification and quantification processes,
to improve promotional campaigns and cross-selling, to improve
account retention, to assess the risk-premium elasticity trade-off, to
assess the comparable profitability of client risk segments, to create cli-
ent social media interest groups, and to assess performance in innovative
new ways. They will operate integrated sales-underwriting-propensity
models to obtain ideal profit profiles and assess customer channels to
improve client satisfaction. Note that a larger focus will be on external
interactions with customers and the ecosystems rather than internal
efficiency.
EY (2014) argues that future underwriters will need: (i) the ability to
synthesize and analyze a wide range of data, (ii) the cognitive skills to
recognize a broader and more dynamic business rules context, (iii) the
technical skills to leverage a high level of automation to gather insight-
ful insights, and (iv) the communications skills to engage and convince
a wider variety of stakeholders. The actual risk pricing of individual cli-
ents will be automated in 90% of cases and thus will not be an essential
skill.
EY (2014) thus argues that the skill set required of an underwriter
in the future will be so different from the traditional skill set that many
current underwriters may not be able to sufficiently transform them-
selves, and that new talent from non-traditional areas will be required.
Newer underwriters will need big data analytical-based degrees and
probably MBAs to obtain the strategic insight. This increased demand
110    
M. Naylor

for strategic skills is occurring against a backdrop of an increasing short-

age of experienced underwriters. Therefore, EY (2014) argues that one
of the keys to insurer survival will be the smart organizational change
management and human resource development of the underwriting sec-
tor, so that older underwriters are retained to oversee the younger ones.
The future underwriter will be as much a data scientist as a risk asses-
sor. Multiple statistics-based models will be joined to codified, heuristic,
underwriting decision tools. Interactive analytical tools will offer ‘what
if ’ scenarios and information visualization techniques like heat maps
and density illustrations. For commercial clients, a far more in-depth
experience can be offered with risk minimization advice based on inno-
vative data and subject-expert advice.
These types of deep-data-analysis systems are already being used in
other industries, which has the advantage for existing insurers that they
merely need to adopt them, and the disadvantage that external firms
which already skilled in these systems are entry threats.
Insurers who are successful will need an underwriting team which has
a culture of continuous innovation and experimentation. It is vital to
note that the pace of innovation will increase at an exponential rate, as
explained in the introduction - that is, the rate of increase of innovation
will increase. Thus, insurers who do not rapidly get near the leading
edge of innovation will rapidly fall behind. As underwriting innovations
will both exponentially cut the cost of client acquisition and exponen-
tially increase client stickiness, insurers will face increasing pressure to
innovate.
Traditionally, companies have followed a three-step new product cre-
ation process of (a) design a product, (b) make the product, and then
(c) sell the product. A low-marginal-cost automated underwriting pro-
cess enables this to be reversed as (a) sell the product, (b) design the
product, and then (c) build the product. Customization to the needs of
customers is the heart of this, even if those needs change as customers
demand new features as they gradually adjust to the new consumption
framework. This is thus not a one-off change but a cultural shift to a
differing style of underwriting.
Therefore, I would argue that instead of being a dying sec-
tor the future is bright for underwriters who see this technological
 5  What Insurance Companies Need to Do    
111

transformation as a challenge worth seizing. Given that a number

of major insurers will fail to respond and thus face bankruptcy, smart
underwriters need to ensure their employer is one of those who are
proactively engaging or prepare to move to an employer who is. Smart
underwriters can use the new techniques and information to enhance
their key role in insurance and join the strategic decision-making pro-
cess. Note that unlike most other areas discussed in this report, the
disruptive impact of IT change on underwriting has already started to
arrive - the future is now.

Automated Claims
While the insurance industry worldwide has made intensive efforts to
speed up claims payments, there is a growing gap between the speed
at which insurers think claims can be paid and the speed at which the
digital generations now expect. Accenture (2013) found that 74% of
policyholders said that their top source of ‘extreme frustration’ with
insurers was having to contact them multiple times for the same rea-
son. Clearly, a large number of pain issues arise with current insurer
systems. In particular, millennials think in terms of days or even inter-
day, while insurers think in terms of weeks or months. Millennials will
experience growing frustration unless a transformation occurs in claims
management. Claims administration needs to be real time and near-
zero cost.
The only solution is automation via telematics. The intensive use of
telematics and administrative software will mean that most activities in
the claims process will be automated, with little direct human involve-
ment. For example, if two network-linked cars crash, the telematics will
provide a detailed history of each car’s actions prior to the crash, and the
two insurers’ computers will thus be able to determine who is at fault,
agree on a sequence of actions, and then contact bank computers to
arrange payments, all within microseconds. They can also inform emer-
gency services, tow trucks, crash repair firms, friends of the car drivers,
change appointments, arrange auto-taxis, etc., before the dust has set-
tled, without asking the clients many, if any, questions.
112    
M. Naylor

If a network-linked TV is stolen, the TV itself can inform the insurer

of this and provide its new address, and inform the police. If the TV
breaks down, it can provide a full report on why and give details of
the part required to the repair shop’s computer. If a health telematic
detects an impending illness, it can send the required medical details
to the insurer and contact local emergency health providers, as well as
the clients’ friends, possibly before the patient is aware of any issue. The
hospital’s computer will keep the insurer’s computer updated with all
current results and procedures. This means that there will be no papers
shuffled or manual administration. Nearly, all contacts will be made
by software, including writing e-mails or text alerts and making phone
calls. Managers can oversee claims in real time and get in-depth details.
The role of the claims staff will then be to oversee the software sys-
tems, to ensure that procedures are working as required, to ensure that
jobs are actually undertaken to the specified quality, to intervene if the
complexity of issues exceeds preset parameters, to ensure the client is
informed and happy, and to create innovative new products or proce-
dures. Note that these are not jobs for IT experts but for claims experts
as they are concerned with claims procedures and client relationships.
The huge increase in data will allow detailed analysis of issues with
claims and liaison with underwriters to determine if particular types
of clients claim in unexpected ways. Claims data can be automatically
linked back to underwriting algorithms.
This automation process can also be in terms of smart software. An
example of this is CoreLogic & Symbility Solutions, who have jointly
created tablet-based software which allows a house assessor to quickly
map a house, note damage, estimate repair cost, and send the required
forms via the internet, to be handled by admin staff anywhere in the
world. This has cut the time to assess a house by 90%, eliminated nearly
all the paper work, and allows scaling in an emergency. There will be
no need for skilled engineering professionals to be physically present, as
they can watch the assessment from their office computer.
Automation has huge advantages for customers and insurer–customer
relationships. Currently, it is difficult for customers to make insurance
clams, as they have to positively contact the insurer, make the circum-
stances of their claim clear, including providing required evidence,
 5  What Insurance Companies Need to Do    
113

understand if their claim fits complex policy terms, work with assessors
and claims staff, and then contact the insurer multiple times to ascer-
tain if each stage of the claims process has been undertaken. The cus-
tomer generally finds the process, even if efficiently run, to be lengthy,
confusing, and alienating. This human and step-based process creates
high costs for the insurer and has no place in a modern digital society.
Digital natives expect same day results with easy access and are often
profoundly shocked by the slowness and obliqueness of current insurer
claims.
The current claims process is antique - slow and excruciating to cus-
tomers. An automated process in contrast is speedy and efficient, as evi-
dence can be gathered by telematics, or imputed electronically by claims
staff, processed in seconds, and made clear to clients. Client interaction
can predominately electronic, with each client able to access an App
which shows where their claim is in real time. Because this informa-
tion links to software, there is no need for humans to update it. There is
no need for large call centers, or even Web sites as information will be
pushed to clients’ mobile phones on a real-time basis. Standard claims
should be able to be processed same day. Because loss assessors and
adjusters, as well as repairs contractors/health systems, and advisers/bro-
kers, are all linked together into an integrated system, the client can get
a complete picture of the claims process.
Research has shown that the customer claim experience is a vital
determinate of perceptions of insurer reputation. Use of telematics will
enable valid claims to be paid a lot faster and provide strong evidence
for decisions when claims are rejected. Aviva Insurance’s intensive use
of data analytics has reduced fraudulent claims by half a million UK
pounds a month and enhanced its reputation.
Insurers can use the integrated data to assess where delays or exces-
sive costs occur and make changes. They can link the claims data to that
gathered from their underwriting or wider ecosystem data to ascertain
which types of clients prefer which type of contact at differing stages of
the claims process or rank them in terms of ease of claims processing, or
which styles increase client happiness. They can link the claims process
into their social media system or wider business ecosystem and provide
the client with useful information or offer cross-selling opportunities.
114    
M. Naylor

They can market the speed and ease of their claims process as a competi-
tive advantage. Claims data analysis can then be integrated with under-
writing data and client response data to create personalized life cycle
advice to minimize adverse events.
The need to undertake a cultural and technological change in claims
is, however, hindered by the decline in experienced claims adjusters,
with over 70% of US adjusters aged over 45. A world of experience is
being lost at a time when a younger generation of programmers needs
to access that skill to be able to create robust automated claims systems.

Cultural Change
The biggest concept which insurers need to understand about digital
transformation is that the transformation is not about the technology
or the data - it is fundamentally about transforming the organiza-
tional culture so the company can integrate the new technology-based
opportunities. McKinsey (2016c) argues that the biggest challenge
­
in adopting an analytics approach and making the evolution from a
‘­knowing culture’ to a ‘learning culture’ is not cost, but is largely lack of
imagination and excess of inertia.
The major mistake companies trying to cope with the digital chal-
lenge make is to regard it as a IT project to be handled by the chief
technology officer, instead of an organizational culture challenge which
has to embedded into all parts of the company.
As explained earlier, IBM (2006) argues that insurers have tradition-
ally focused on product and process optimization, rather than innova-
tion. This has created a point-solution mentality which causes insurers
to treat the symptoms of persistent problems while ignoring root causes.
This has to change. Instead of working hard to improve current pro-
cesses, insurers need to drastically transform their businesses.
The insurance industry does not have a technology problem; it has
a problem in the use of technology due to corporate culture and inter-
nal politics. Bain (2015) argues that insurers need to: (i) build advanced
analytical systems, (ii) create effective digital distribution, (iii) create
customer centricity, and (iv) digitize internal operations and sharply
 5  What Insurance Companies Need to Do    
115

reduce cost. One of the major issues associated with this transformation
is inhibitions caused by the traditional culture and organization of the
insurance companies. KPMG (2014) argues that the incompatibility of
multiple legacy systems is a symptom not a root problem. Insurers tend
to be policy-centric with all the systems geared to a specific product
with a minimal view of customers. This creates siloes between business
lines and an unwillingness to share data or internal functions.
Insurers need to reorganize and move away from current practices.
They need to move from traditional lengthy business planning to more
flexible, less exact, more data-led, and more iterative approaches. They
need to learn how to launch new products or services while testing and
obtaining feedback and responding to that feedback in a continuous
process. Time to market for new products needs to be shortened and
made more flexible, so multiple combinations can be trialed. They need
to simplify products while also customizing and diversifying product
ranges. They need to be prepared to offer trial products to customers
and service them through a multitude of service channels.
Murli Buluswar, in McKinsey (2016c), argues that the effective use of
analytics is dependent on a cultural change and that most insurers fail
this challenge due to fear of the unknown. He argues that it’s a mind-
set change from an ‘expert-based mind-set’ to a ‘learning-based, flexible
mind-set’. The issue with experts in the future world is that knowledge
is likely to change at a deep level, so that the paradigm which the expert
has used to arrange their thinking is no longer valid. Given that it took
time and effort to learn the old one, it is hard for most people to accept
that they have to adopt a new paradigm. Knowing how to rethink and
reimagine will be more important than knowing stuff. The company
will have to refocus company culture toward creating a dynamic learning
environment. Yet, how many incumbent insurers currently have a ‘cul-
ture executive’ or a ‘learning executive’? The Facebook manual famously
notes that ‘If we don’t create the thing that kills Facebook, someone else
will. “Embracing change” isn’t enough. It has to be so hardwired into
who we are so that even talking about it seems redundant.’
New entrants from sectors which already have flexible product
cycles are more likely to be able to manage a more flexible approach
to production than traditional insurers stuck within legacy organization
116    
M. Naylor

systems. These changes mean that traditional insurers’ focus on exacti-

tude needs to be replaced by a focus on speed and flexibility. Insurers
who follow the traditional focus on cost control and taking the time to
get a product just right before it is released run the danger of the market
having changed substantially in the meantime.
KPMG (2016) notes that large business organizations tend to be
good at minor, tweaking, changes which focus on improving existing
business and operating models. Insurers are keen on optimizing the
sub-optimal. However, they do poorly at fundamental transformation,
the type of enterprise-wide strategic change which truly reinvents the
company. There is a big difference between the skills required to encour-
age digitization and those required to guide a successful transformation.
You need a really strong change management capacity, and that is rarely
available from existing managers, as they have skills in achieving effi-
ciency in a stable environment, not in disrupting that environment.
This is because there are typically far too many embedded incum-
bent managers who fear they will lose from a change and because trans-
formation requires a much broader view, and a much more strategic
approach, than most insurers have needed for past initiatives. It is also
very hard to see the value of benefits from an exponential process during
the initial stages - all that can be seen are the immediate losses. Given
the inherent fuzziness of transformation, many of the initial changes
will fail to produce sufficient advantages and therefore will give ammu-
nition to those who argue against the need for change. KPMG argues
that for these reasons most incumbent transformation efforts will fail.
What may be required is the introduction of new senior managers with
transformation experience, or skilled consultants.
If an insurer wants to reinvent itself, they can’t just look at what
insurer competitors are doing, but instead need to look at what external
firms are doing in disruptive sectors. Many companies also fail because
they have tended to appoint employees like themselves. This means that
even a transformation team will all have the same world view and there-
fore will be unable to visualize the broad range of possible futures. An
insurer will need to create a diverse company culture with many dif-
ferent viewpoints, focused especially on employees who can break
 5  What Insurance Companies Need to Do    
117

incumbent employees out of their existing mind-sets. Discomfort has to

be maximized.
This transformation process has to be embedded in every aspect of
the insurer’s operations, not just underwriting or customer response. All
administration areas need to be consulted on requirements so all infor-
mation and decisions are available to all who need it. Marketing needs
to be comfortable with big data and social network analytics, and be
able to communicate closely with the data analysts so that the correct
metrics can be created and found in the huge mass of data. Customer
contact staff needs to be at ease across multi-channels and have instant
access to whatever they require to answer questions. Brokers and inter-
mediaries need access to all the information and tools they require in a
manner which enhances the customer experience. These tools need to
be flexible so they can be instantly changed based on customer reaction.
Pricing needs to be instant and dynamic.
Product siloes need to be replaced by a customer-orientated divisions
so that each segment of customers can receive the type of experience
which makes them enthusiastic and feel engaged. All decisions need to
be seen through the viewpoint of the fairness for the customer - whether
they understand the product, whether they can afford it, and whether
it is likely to deliver value for money. Customer experiences need to
be mapped so opportunities to improve the customer experience can
be found and enhanced by additional products. Products and services
need to be developed and tested with customers and social networks.
Social media trends need to be tracked and used for customer debate.
Metrics need to be developed which track customer satisfaction, level
of understanding, product usefulness, and likelihood to recommend or
rate highly.
Flexibility and speed of response are essential. Experimentation
is required where new products or services are trialed and modified.
Separate divisions focused on new products or new business processes
could be created. It can be useful to initially focus on quick wins so
enthusiasm and trust in the process can be developed.
It is also important to note that while the data and analytics should
be very complex, the IT interface to staff or customers has to be as
118    
M. Naylor

simple as possible, so that information is presented to staff in a high-

quality fashion, and so it can be easily comprehended and acted on.
Interfaces need to be developed by experts in human–computer interac-
tions and in graphical displays. This way every member of a team can
avoid feeling overwhelmed.
The initial problem is that substantial reorganizations are very costly
and disruptive of current productivity, without immediately generat-
ing additional revenue. It is hard to argue for costly change now, for
the sake of future survival, unless there is an immediate threat. Leaders
are needed who have the capacity to understand a longer-term strategic
vision and the ability to implement it. Boards and CEOs need to under-
stand exactly how far their current culture is from the culture required
for future survival, the ‘cultural gap’, and be willing to push for the
required changes. If boards wait until the gap is large enough to allow
external entrants to disrupt the market, then they have probably waited
too long.
Fortunately, there are a number of data integration opportunities
which can be more easily trialed and which can help to signpost the
transformation path. One is to create a data visualization system which
sits on top of existing disparate databases and can be used to extract key
data and create a data analytics system. These can trial different data sets
or models so that real value insights can be found to be shown to the
board. Another immediate priority should be to create an experimen-
tal company, where new ideas can be trialed and problems discovered
before they are spread company wide. The main danger here is these
short-term measures reinforce the mind-set that the solution is techno-
logical based rather than cultural, so their purpose should primarily be
used to create buy-in from the senior leadership team.

Transformation Pathways
EY (2013) suggests that the path to transformation in the near term
involves:
 5  What Insurance Companies Need to Do    
119

Step 1: the now 

1. Creating a clearly defined digital strategy: This involves identifying the
ideal future insurer model and mapping how to get there. This pre-
dominately involves looking at examples outside the industry and
recognizing that it may not be possible to create sector leadership in
every dimension.
2. Identifying initiatives which deliver the most: This involves creating
data and analytical capacities so that insurers can identify customer
expectations, matching these with current inadequacies and therefore
identifying where to focus management effort and spending.
3. Helping distribution partners develop digital capacities: This involves
examination of all the channels customers may choose to use and
ensuring a high-quality, omni-channel experience. Therefore, inter-
mediaries and agents need to have the capacity to deliver the required
customer experience as part of an overall process.
4. Frame the investment argument: This involves shifting management
focus from short-term to long-term decision making and from inter-
nal efficiency to customer-centric. The conflict between the need for
transformation and the CFO agenda needs to be made explicit, as
well as the conflict between customer retention and acquisition.
5. Build analytical capacities: This involves developing physical and staff
resources, as well as reorienting rate-making and marketing staff to
a big data focus. The focus has to be on the best use of the analyt-
ics rather just its creation, particularly how to use analytics to create
deep insights into customer priorities.
6. Develop mobile functionality: Customer interaction via mobile phone
Apps is the new norm.
7. Take social media seriously: Any insurer who is not constantly on
social media will not exist in the eyes of digital natives.
8. Embed continual innovation into the company culture: Change is the
norm, so the transformation cannot be seen as a once-off but contin-
uous. This involves identifying cultural constraints, utilizing external
influences to overcome them, and aligning KPIs with innovation.
120    
M. Naylor

Step 2: near term 

1. Combine cultural change with technological change: The response to
the technological challenge is more about company culture than
technology.
2. Embrace cloud computing: Digital information needs to be always and
everywhere available. Given the skyrocketing demands of big data,
insurers need to be able to rapidly upscale capacity. It is mission-criti-
cal that data outrages do not occur.
3. Create a sustainable culture of collaboration and innovation: It is
impossible to impose significant cultural change by top-down edict.
Staff needs to be brought through an evolutionary process, so cul-
tural change is embedded. The idea that flexibility and experimenta-
tion are more important than accuracy or cost control needs to be
embedded into management. Company-wide calls for innovative
possibilities need to be made.

KPMG (2016) argues that insurers can define their future business
and operating models by breaking the transformation analysis into lev-
els of value and examining these over differing time frames; 3–5 years,
5–10 years, longer than 10 years:

1. Financial ambitions: What are the future financial outcomes, struc-

turing, investment, and capital allocations.
2. Markets: What are the best future markets and do the current portfo-
lios of business maximize the chance of achieving those targets.
3. Propositions and brands: How should the portfolio of propositions
and brands evolve over time.
4. Clients and channels: What changes to the operating model puts the
client at the heart of the business.
5. Core business operations: What business processes will deliver the
financial outcomes required within a high-performance model.
6. Operational and technology infrastructure: What are the required pri-
ority technological structures.
7. Organizational structures: Are there the proper oversight structures to
drive change and ensure responsibilities are met.
8. People and culture: What leadership is required at various levels to
drive the transformation with a high level of passion.
 5  What Insurance Companies Need to Do    
121

9. Measures and Incentives: How will change be measured, in terms of

what issues.

It is vital to note that while regulatory issues have delayed change in the
insurance sector, this will not save insurers. This is because the insur-
ance sector is inherently internationalizing. Therefore, while regula-
tion remains national, external disruptors will find it very easy to cross
­borders - on the Web there are no borders. This means that if regula-
tions in a particular country incentivize national insurers to delay
transformation, those insurers will fall more and more behind the
international leaders and the increasing productivity gap will make it
increasingly attractive for international disruptors to enter via Web sites.
Incumbents will not be in a position to respond.
It is thus in the long-term interests of insurers to induce national
regulators to proactively respond with state-of-the-art regulation, so that
national insurers become the disruptors rather than the victims.

Legacy Systems
Many of the top financial service firms are large, complex, organizations
which were created from multiple mergers and acquisitions. They thus
have siloed personal structures, fragmented administrative processes,
and mashed together IT systems, relying on incompatible data types.
This ensures that correct information is hard to obtain and decision
making is protracted.
One key to future success insurer is thus to get rid of legacy systems
and the complexity and bureaucracy associated with the traditional sys-
tems. They must create simplified, scalable, technology based on big
data, real-time feedback, and dynamic responses to customer expecta-
tions. They have to support the kind of real-time multi-channel inter-
actions, predictive analytics, and mobile service delivery options which
the millennials and subsequent generations expect.
Insurers have to be aware that competition will not just come from
existing insurers but will also come from firms which have already cre-
ated those innovative and socially linked characteristics in other sectors.
122    
M. Naylor

These firms are not burdened by legacy IT systems and business pro-
cesses, are tech-savvy, and have substantial social capital, as well as
existing intimate knowledge of customer characteristics. For example,
Google has already entered the US general insurance market. It is, thus,
vital for insurance managers to not dismiss the threat of a disruptive
technology just because no other insurer has yet trialed it.
The new entrant could have the bulk of its staff based in foreign
country. Youi, for example, has recently entered the NZ market with a
minimal local presence and has shown that a firm does not have to be
locally based to be effective. Any insurer CEO who doubts the possible
threat only needs to contemplate the example of Kodak dying because
of the entry of firms from a previously unrelated sector. Could the firm
which destroys your insurance business be an airline based in Malaysia
or a Chinese Web platform?
Thus, a key to future success is to work with firms from other sec-
tors which have existing superior software skills, especially in customer
contact, and to supply insurance products via these firms’ Web systems.
While most insurers are moving into direct online sales, their attempts
have in general so far been weak and have not exhibited the attention to
customization and rapid feedback which is now expected. Direct Web
sales are also hardly leading edge. Very few insurers are exhibiting any
indication that they have the skills to move into a qualitatively different
era in terms of client servicing.
It is important for insurers to understand how disruption has occurred
in different sectors. For example, Amazon created the first wave of its digi-
tal disruption by using a Web site to offer a catalogue 10x larger than the
largest physical store and 10% cheaper. They created the second wave by
enabling a community via allowing bloggers to post reviews and encour-
aging the public to post ratings. It then used the behavior of this com-
munity to gain insights and develop collaborate algorithms. It understood
that the bigger its network, the richer its data, the better is the customer
experience. It is currently creating the third, larger, wave by broadening
its product range to include anything deliverable by truck or net, by dis-
tributing for smaller merchants, and by building the world’s largest cloud
computing service. Amazon has worked on building scale in data and ana-
lytical techniques. It has been proactive, not waiting for change, and not
 5  What Insurance Companies Need to Do    
123

protected existing income streams from new opportunities. It has under-

stood that it is not an online book seller, but a business based on data
analysis.
Amazon beat traditional bricks-and-mortar stores because it built
digital from the ground up, rather than try to add bits of digital onto
traditional systems. Insurers will struggle because they generally fail to
understand that going digital involves changing the entire structure of
their business processes, and not just adding a shiny front-end Web
site. It is more about internal management than physical technology.
Insurers need to bring in outside expertise, but they cannot just insert
them into current structures or their talent will be stifled and they will
leave. There needs to be management support or the appointment, of
say, a ‘digital manager’ will achieve nothing as the rest of the staff may
think it’s that person’s job and they don’t also have to change.
One of the main issues with IT system improvements is that gener-
ally IT departments are already overwhelmed. Gartner surveys show
that typically 80–85% of an IT department’s budgets are spent on
essential maintenance and that requests for new services or changes sub-
stantially exceed what can be done in the remaining 15–20%. Gartner
shows that the ‘new request’ IT backlog for US companies typically
increases by 10–20% per year, and since resources are not there to han-
dle this, it is creating a compounding backlog. Gartner estimates that
each IT application cost about 40% of its purchase cost to maintain.
The budget allocation for new projects has typically been decreasing
because the time required for maintenance has increased faster than
budget allocations. This has resulted in management being disappointed
with IT department performance and has discouraged any plans for
radical change in IT services.
Management tends to response to demands for urgent changes in IT
systems by one of four methods:

1. Buy and customize a packaged application: This is often the fastest

solution. However, typically, these packages deliver standardized pro-
cesses which are normally different to what the firm requires. Thus,
customization is typically required, and the packages are not built
for change so customizing is difficult and expensive. Adding this
124    
M. Naylor

requirement onto an already overworked IT department is a recipe

for further failure, adds to the routine maintenance workload, and
multiplies incompatible systems.
2. Build a custom solution in-house: Building a large and complex sys-
tem is often very expensive and complex given current techniques
and tools. It also tends to be prone to increasing failure rates as the
size increases. The in-house systems are typically not built for change
so amendments are difficult, and because it is non-standard, the
company is hostage to the original systems’ integrators. Over time
ad-hoc expansion multiplies maintenance workloads and adds over-
whelming complexity.
3. Select and rent a cloud-based solution: Software-as-a-service companies
now offer partly customizable software and supply external hardware.
However, this does not address issues with interfaces, workflows, and
data repositories.
4. Do nothing: This is the typical response of an overworked IT depart-
ment, whereby new requests are logged in and put into an endless
waiting list.

The real reason for this growing IT mess is that it is very expensive and
time-consuming to create and amend large software systems, and this
increases nonlinearly with size. Some reasons for this in a typical com-
pany are as follows:

1. Reverse-engineering hundreds of thousands of lines of code to find

the exact place to make changes is complex and requires experienced
staff.
2. Validating these changes across all applications and sub-systems is
extremely complex, and dependences grow exponentially with size.
3. After changes have been made and validated, deploying a new version
into production is complex and error prone.
4. Integrating existing and new diverse systems is complex and
problematic.
5. Under deadline pressure, non-functional requirements are often for-
gotten or ignored.
6. After going live, no audit or troubleshooting capacities are in place.
 5  What Insurance Companies Need to Do    
125

The solution to these issues requires a substantial budget and personal

increase, external expert support, and an integrated agile software engi-
neering process aimed at gradually streamlining and integrating systems
so maintenance reduces proportionally over time. The end objective has
to be an IT process aimed at innovative, user-friendly, products which
transform insurer productivity. None of this is easy.

Integration with Customers
Insurer customer engagement is regarded by experts as one of the top
crisis issues, with 80%+ insurers having stated programs of investments
in improvements. The customer disconnect crisis, however, is not due to
insurer failure, but due to the changing expectations of customers and
generational change. Insurance customers are becoming disconnected
from the insurer’s value proposition, confused, and tend not show
understanding of any aspects of the value proposition apart from price.
They are becoming increasingly hard to sell to, with only a limited pro-
portion actively seeking cover. Profit margins are dropping as insurers
compete on price, yet find that customer stickiness is dropping, so that
expenses on client acquisition are not recovered. What modern custom-
ers want is simplicity, accessibility, platform neutrality, social connected-
ness, and a clear explanation of a product’s value. Very few insurers are
meeting these demands.
PWC (2014) argues that insurers’ focus on risk, ratings, and prod-
ucts mean that their understanding of customers lags substantially
behind that of other industries and leaves them highly vulnerable to dis-
ruptive entry by techno-literate firms with high social capital. Morgan
Stanley/BCG (2014) argues that unless insurers drastically increase the
frequency and quality of their interactions with clients, insurers will be
very vulnerable to disruptive entry by service providers who have an
established high level of client interactivity.
For the ‘no-wait-generation,’ the speed of response needs to be quali-
tatively changed. Millennials expect a response to a simple query within
one hour and to a complex query within 3 or 4 hours. Any phone not
answered within 2 minutes is an aggravation. Any insurer who talks
126    
M. Naylor

longer than one day to respond to a query will not get that customer
coming back. The expectations of digital natives are in minutes, not
hours. Why can’t a claim be paid in 10 minutes? Customers expect the
person or system they have contact with, to have the answers, and not
respond ‘I’ll get back to you’ - any respectable customer relations IT sys-
tem will give each staff member all the information and authority they
need to satisfy a customer.
Responses to customers and changes to online information sys-
tems need to occur inter-day, yet many large insurers take months to make
simple changes. Morgan Stanley/BCG (2014) cites an example of an
insurer taking 180 days and 35 signatures to change a file on the company
Web page. Inflexible companies will be at a severe competitive disadvan-
tage, as the speed of response is starting to become more important than
accuracy. Many insurers still have policy application processes which take
weeks to complete rather than minutes. A dangerous customer expectation
gap is opening up, which can be filled by external firms from sectors who
are used to fulfilling the new generations’ higher customer expectations.
Modern customers also expect multiple channels to be available, Web
pages, branches, phone advisers, etc., and expect these to integrate, so
whatever information is given to a Web site is available to any other
staff/adviser; information should only be given once. While the use of
physical branches is dropping, customers expect these to be available
for when they need them. Capgemini/Efma (2016) found that while
baby boomers were happy with insurer contact via e-mail and Web
sites, younger generations expected multi-channel access and lack of this
was a major factor in switching decisions. They also found that while
younger generations were engaging with insurers up to 2 or 3 times as
often as baby boomers, they were unhappier than baby boomers with
the level of engagement. This means that even though insurers world-
wide are trying to improve their customer responsiveness, customer
expectations are rising faster than insurers can change. Technology gives
insurers the ability to proactively service customers. If a customer has to
call an insurer to find out what is going on with any part of their policy
that should be considered a failure by the insurer.
In general, it is still difficult for a customer to get information on
insurance products, on policy quality issues, on how other customers
 5  What Insurance Companies Need to Do    
127

liked the product, on claims processes, and on the process of their

claim. Typically, customers have to use call centers and often experience
delays of three or four minutes. The call center staff member often can’t
answer issues and has no real-time information as to where in the pro-
cess the claim is. Customers compare this to other sectors, like Amazon,
where obtaining information on products is simple, independent prod-
uct reviews are available, payment is simple, each stage of delivery is
tracked, and returns are assured if the product is not to their liking.
They can get very aggrieved that insurance seems to be from a by-gone
age. How many insurer Web sites offer advice on what insurance prod-
ucts the customer does not need?
Customer loyalty is also dropping, with MyCustomer (2015) find-
ing that 57% of UK respondents consider switching their car or house
insurance every year. They also found that only 32% of customers
reported positive customer experiences. Thus, very few customers are
excited or engaged with their insurer and are very open to considering
an external disruptor.
Insurers are aware of their customer disenchantment issues, yet,
despite active investment, seem unable to close the engagement gap.
Adding an App or two won’t cut it; insurers need to drastically trans-
form the customer experience. Yet, many insurers still seem impressed
with themselves for having a Web site! Insurers need to move into the
twenty-first century. In particular, insurers cannot expect customers to
come to them; they need to attract the customers to them. They need
to be more interesting and relevant - show that their product enhances
customers’ lives and is not a grudge purchase. They need regular,
enhancing, and relevant contact with customers, across all channels.
This contact needs to be interactive, and adaptive, responding to cus-
tomer concerns in real time. While insurer response may involve some
customer segmentation by channel, it is vital to recognize that most cus-
tomers will expect to use multiple channels.
The customer journey from initial awareness of the use of an insur-
ance product, to purchase, to claim, needs to be systematically
reworked, so that insurers are reoriented and redesigned around the
customer, and not around reducing actuarial risk. McKinsey (2016a)
argues that this redesign requires four dimensions:
128    
M. Naylor

1. Installing customer empathy: Good design requires real empathy

that goes beyond what customers think they want, and responds to
underlying needs, not the stated needs. Detailed customer feedback is
required.
2. Getting comfortable with an iterative approach: Good design meth-
odology requires an initial release of ‘imperfect’ products, which are
iteratively improved. This enables customer feedback to be incor-
porated and drastically speeds up response time. In this framework,
products are never complete.
3. Replacing functional siloes with agile cross-functional teams: Because
insurers have been structured around competency areas (e.g., risk,
underwriting, sales), no one really owns the full customer experience
or is accountable to customers. This can only be overcome by use of
agile, cross-functional, teams which are project orientated.
4. Braiding business and IT: Traditionally, business has led IT, setting
requirements and parameters. In the future, IT will be too impor-
tant to be left to either the CTO or the COO, and thus, business
processes and IT need to be braided together. All IT solutions need
detailed feedback from all areas of the business, as well as from
potential customers.

This redesign process has to end up evolving a customer journey

which is exciting, engaging, and ongoing. This leads to four outputs:
(i) deep customer understanding, so that you can predict what they
value and respond to it; (ii) embracing empathy, so the right people
with the right skill sets are in the right place; (iii) real-time response,
which combines design, strategy, and technology; and (iv) quickness, so
that the company responds rapidly to changes.
The goal should be to create relevant, engaging, interactions at every
contact point, regardless of mode, to understand the context of each
interaction, to provide genuine two-way dialog, and to provide consist-
ent high-quality service. Each stage should be tracked by data metrics,
so that breakdown points can be found and trials conducted on how to
increase customer happiness. Low cost is far less important than engage-
ment satisfaction.
 5  What Insurance Companies Need to Do    
129

Customer engagement needs to be customized, so that the style

of engagement, reminder frequency, social network links, etc., are
amended as the particular customer desires. This could be based on
feedback metrics or could be set by the customer. This is a vital issue -
very few insurers think about letting customers themselves define the
terms of interactions. Traditionally, insurers have controlled infor-
mation, rather than allowed a two-way flow. This is contrary to the
assumptions of the younger generations.
Very few insurance incumbents have tried to examine their business
process from a customer perspective; they are still locked into a prod-
ucts-and-processes model. Most incumbents cannot even provide Apps
which allow a customer to update account information in an active
manner (so answers change subsequent questions) or cross-populate
other customer data fields. Very few offer millennial stables such as one-
click-buying, set-and-forget-reordering, open-customer-reviews, and
personalized suggestions based on prior purchases or Web serving or
social media. These are basic level functions, not advanced, but seem
beyond current incumbents.
Insurers also have minimal information on their customers. For
example, social media giants will be automatically aware if customers
undergo significant insurance trigger events (like buying a house, get-
ting married, or having a baby), while insurance incumbents have no
idea. A simple App which allows a customer to change all customer
details for all companies within a business ecosystem after a life event,
for example, a change of address, is still speculative for insurance
incumbents, yet is regarded as base standard by digital natives.
Personalized, context-specific, interactions require the combination
of customer insights, digital content creation, with delivery via omni-
channels. This gives customers the sense that insurers know who they
are, what their needs are, how they relate to the product or company,
where they are in the life cycle, and what occurred in previous interac-
tions, whatever the channel, wherever it occurred inside the ecosystem.
This is a product both of end-to-end IT systems and interactive CRM
systems.
Martinez (2016) argues that it is vital in internet advertising to
understand where the power in customer data lies: Is it with the
130    
M. Naylor

network supplier, ‘the publisher’, or is it with the provider of the click-

through ads, ‘the advertiser’? Each of these two parties collects a dif-
ferent type of information about browsers, and thus, the two data sets
have quite different monetary values. In general, value has been shift-
ing toward advertiser data, so if an eyeball provider, like an insurer, isn’t
savvy enough to create sophisticated browser targeting and tracking
software, then it will find it very hard to arbitrage any value from its
ecosystem partners. Currently, most insurers do not have Web site CPM
or the ability to track conversion rates.
The largest issue is that insurers are not generally aware of how far
behind the leading edge of customer engagement software or customer
expectations they are. They need to radically reimagine their role and
recreate their IT processes to close the engagement gap. The upside is
that experiences with active engagement InsurTech show substantially
increased sales across a range of products and a substantial increase in
customer stickiness.
Potential external disruptors have advanced customer relationship
and insight systems and therefore will find it easier to add insurance as
a product than incumbents will find it to drastically redesign their busi-
ness processes. Incumbents will struggle more in this area than others.

Big Data and Artificial Intelligence

While some large companies have tried to create effective customer
insight functions, so that it doesn’t end up as just another traditional
market research operation, most have struggled. One reason for this is
that elevating customer insight into a strategic position requires strong
and persistent efforts from top managers. Another reason is that cus-
tomer behavior is rapidly changing, with younger generations approach-
ing product purchase quite differently. The largest source of data
relating to customer changes is those derived from big data sources,
yet most insurers have yet to move from struggling to understand this
toward making mastery of big data a competitive advantage.
The availability of big data is exploding, yet the possibilities of its
use have been little explored in the insurance industry so far. A major
 5  What Insurance Companies Need to Do    
131

issue with big data is its size and unstructured nature. To the uniniti-
ated most will seem to be noise. Only inadequate standardized systems
have been developed so far to allow insurers to extract the useful aspects
of this overwhelming data, so innovators will have to invest in creating
bespoke analytical systems.
It is also useful to move from a simple closed analytical underwrit-
ing model, which has a set of equations, toward a more forward-looking
model, which uses scenarios and structured cause–effect chains to give
a deeper understanding of loss possibilities. This is because in a rapidly
changing world, the causes of major losses may be events which have
not yet occurred, rather than events which are known. Within this,
insurers need to sort out the ‘known unknowns’ from the ‘unknown
knowns.’
Using machine learning methods on big data involves a move away
from a ‘deductive approach’ - where you apply rules learned from a sam-
ple to particular clients, toward an ‘inductive approach’ - where you use
detailed data on clients to learn general rules. It also involves regarding
data as a ‘real-time flow’ rather than static, closed, pool. Siegel (2016)
argues that this requires focusing on data dynamics, the characteristics
which cause change, rather than on stationary casual factors; ‘how do
the behavioral characteristics of current clients differ from the clients
you had two years ago? Are they more or less likely to claim or cancel?’
Insurers can use big data to increase their customer base in two ways:
(i) by selling more to existing customers and (ii) by selling to new cus-
tomers. A vital aspect of this will be effective data-driven customer
insight systems. These use predictive models to anticipate evolutions
in customer behavior. Careful analysis of big data helps this process by
giving agents feedback on which types of customers and which style of
interactions result in higher prospect to sales conversion, in lower lapse
rates, in larger policy size, and in more reliable customer information.
Big data also allows companies to differentiate agents on these criteria.
The reduced cost of underwriting due to automation will enable insur-
ers to penetrate lower down the income chain, to the lower middle-
income market, which has traditionally been seen as too expensive to
service.
132    
M. Naylor

A vital part of the response by insurance survivors has to be the crea-

tion of effective AI systems using adaptable learning algorithms. One
of the key reasons for this is that possible competitors like Google have
taken the deliberate policy of transforming themselves into a ‘machine-
learning-first company.’ Although these Silicon Valley pioneers have
always used AI, they have recently become obsessed by it. For years, AI
was considered a specialty, limited to an elite few; now, IT thought lead-
ers argue that it is becoming the core of future disruptors.
The core skills required to master AI differ substantially from the
skills traditionally required from programmers. In particular, AI does
not involve deciding on work flows and output equations before the
start but setting up a system to create its own answers. Creating AI sys-
tems thus involves (i) identifying the right data, (ii) choosing the right
algorithmic approach, (iii) creating the right conditions for success, and
(iv) and then trusting the system to do the work correctly.
While AI systems won’t replace humans, they will radically change
human life. The use of superior AI systems will allow insurers to mine
big data derived from telematics and ecosystems to discover activity-risk
correlations which are not obvious to actuaries, and to cut marginal cost
substantially. AI will be a key component of internal administrative sys-
tems, especially their ability to respond to customers via auto-genera-
tion of e-mails or phone calls or social media posts (using text and vocal
recognition skills) as well as responding actively to telematic feedback.
The AI analysis systems need to be led by business and not technical
managers, so that practical applications and insights are discovered. This
role of linking the data to the business is critical and involves a rare tal-
ent of understanding both data and business.
Expertise in AI will allow insurers to integrate their products into
the telematic-based products used by other firms in their ecosystem,
enhancing these products and extracting vital data. An insurer who has
no ability to offer ecosystem partners this integration ability will not be
attractive ecosystem members and will likely lose their place. The differ-
ence in approach means that insurers’ teams of programmers may need
extensive retraining. This can be difficult as skilled and experienced AI
trainers are scarce, and once trained, data analytical staff will be offered
large inducements to move to more exciting companies.
 5  What Insurance Companies Need to Do    
133

It is vital to understand that use of big data insights within an

expanded IT administrative system does not just allow replacement of
repetitive activities, but their major gain is that new insights will allow
humans to work more efficiently. For example, (i) pricing can become
segmented and targeted, as analytics allow deeper understanding of the
value of products to different customers and the resultant price elastici-
ties; (ii) management can work out the optimal size and composition
of teams, and the optimal number of tasks particular types of workers
should be involved in at any one time; (iii) management can work out
what types of sales approaches work best for which types of clients; and
(iv) management can work out which types of data dashboards provide
staff with the critical information and how different types of customers
react to this.
Technology is going to be an important part of insurers’ ability to
capture and analyze new sources of customer data and develop deeper
relationships. Yet, the real differentiator is how effectively this informa-
tion is turned into insights and a readiness to lead the innovations in
the marketplace. Using big data analytics to develop a better under-
standing of customers also enables insurers to develop products or ser-
vices which radically enhance the customer experience.
The use of these data streams can also enhance insurers’ integration
with customers’ lives by offering data-based services. For example, car
insurers can link their in-car dynamics insurance systems to traffic sys-
tems and provide minute-by-minute feedback on traffic density condi-
tions over differing routes to the customer’s destination. Health insurers
can link the food a family buys to blood-chemistry outcomes. One of
the keys for insurers to become a customer-centric company is a focus
on the integration of big data into client-facing activities, delivered via a
radical reduction in per-service administrative costs.
It needs to be remembered that data security will be mission-critical,
as any serious breach of customer confidentiality could lead to law suits,
regulator action, and insurer bankruptcy. Cloud systems and data link-
ages are also not infallible and can suffer outrages. For a dynamic rat-
ing system, this could lead to major issues. Thus, multiple backups and
access systems will be required, as will power backups, alongside state-
of-the-art security protocols and a proactive data security team.
134    
M. Naylor

A major issue with multinationals using big data about people is that
its use is subject to regulation, and this will differ for every country the
firm operates in. Given that the only way to maximize data use is on
an international basis, across all jurisdictions, international agreements
are required. This needs to include legally enforceable guarantees for
protection of privacy and non-discriminatory application. This can be
extremely difficult if over a hundred different units within a firm over
twenty countries are constantly interacting with the data lake.

Internal Systems
Many of the issues which insurers face can be traced to either to
poor legacy IT and inflexible management systems, which fail to give
staff company-wide information on a customer, or to inadequately
fine-tuned data, or to a failure to positively engage with customers.
McKinsey (2016b) found that only 14% of companies with transforma-
tive plans had been effective in trying to achieve the aims of their data
and analytics programs. Those 14%, however, had found that their mar-
ket share and revenues had increased more than they had expected. Of
those who struggled, their major obstacles were not technological but
difficulties in securing effective internal management support and dif-
ficulties in designing an effective organizational structure. CEOs and
boards seemed more able to focus on the physical aspects of the crea-
tion of a new IT system than on either the analytical side or the need
for a new style of management. All the high-performing companies
had CEOs who were engaged and supportive of the required cultural
changes.
EY (2013) also finds that the main factors holding insurers back
from digital transformation are internal - legacy technology, slow pace
of IT system implementation, and business cultural constraints. They
found that 47% of insurers surveyed had no clear digital business case
and for those that did have plans, those plans only covered low-level
issues like integrating budgets with HR planning. The majority of insur-
ers felt they had inflexible and inappropriate operating models and
management systems. Only 5% of insurance middle management felt
 5  What Insurance Companies Need to Do    
135

that the senior management understood the change required and led
by example. While 30% envisage this improving, no insurers had sen-
ior management which seemed to comprehend the scale and urgency
of the transformation needed. McKinsey (2016b) argues that the big-
gest hurdles to an effective analytics program have not been a lack of
strategy or tools; it has been a lack of leadership support and com-
munication, ill-fitting structures, and trouble finding the right people
for the job. Unless the mind-set of the senior management of current
insurers changes, they will continue to lag further and further behind
other sectors and thus remain extremely vulnerable to entry by external
disruptors.
The siloed, product line, basis of traditional insurer management
means that insurers tend to have different IT systems for each busi-
ness line and different systems for underwriting, for customer relation
management, for claims management, and for sales. There is no IT link
with related firms like suppliers or brokers. This means that client infor-
mation often has to be inputted multiple times, staff often have only
a partial view, queries and claim cannot be handled fast, effective data
analytics are impossible, and the top management does not have a com-
prehensive overview of key metrics.
The main reason why insurers have been laggards in IT is because
most insurers have traditionally had a risk-adverse culture focused
on cost control and accuracy of data predictions. Failure in a project
has been seen as career death - there is no ‘Fail Often, Fail Fast, Fail
Forward,’ approach. However, the future will require experimentation
and a willing to allow systems and staff to fail at these trials. A ‘con-
trolled failure’ has to be seen as a career-enhancing rather that a career-
limiting move. As well, since more data is vital the traditional idea that
‘those who hold the data hold the power’ has to change to the idea that
‘those who share the data have the most power.’
Banks have been more innovative in use of digital systems, informa-
tion, and customer engagement and thus threaten to take market share
from traditional insurers. Unfortunately, internal insurer attitudes to
data sharing are inbred into insurer company culture as well as legally
regulated so change could be too slow to enable insurers to meet the
challenge of outsiders with a sharing culture.
136    
M. Naylor

A substantial increase in IT investment is required. EY (2013)

found that insurers currently only spend about 10% of their business
and IT budgets on digital developments and IT budget share is barely
enough to maintain existing systems. Very few insurers spend 2% of
their budget on R&D, yet to compete this will have to exceed 4%.
Investments in IT systems and staff retraining will have to be multiplied
by orders of magnitude, as the most common reason for the failure of
technological transformation projects is a lack of adequate investment.
Unfortunately, EY (2013) found that a majority of insurer managers
have no clear idea of where existing spend goes. Only about 15% of
insurers are trialing cultural change projects. The vast majority of insur-
ers are still focused on internal change issues rather than rapidly chang-
ing customer expectations.
One of the major obstacles is that the future is arriving so fast that
proposed threats often seem very unlikely and so are outside the scope
of any ‘reasonable’ threat review. Why would Kodak have thought
about the electronics sector when they examined threats to their film
business? Given that 90% of possible threats will prove to be busts, it
is also very hard to establish clear plans for reaction. Therefore, threat
reviews need to be open to blue-sky thinking and ‘unreasonable’
ideas. As wide and diverse a range of employees as possible should be
included, including external thought leaders. Insurers will need to trans-
form their work forces with a culturally different style of staff and a
management structure which focuses on customer response and flexibil-
ity rather than accuracy or cost. Businesses then need to maximize their
flexibility so unexpected threats can be countered.
One way to create major internal cultural change is to shift to a social
media style internal communication system. This means that, instead of
internal communications being person to person via e-mail or phone,
with a strong top-down bias, webs of information are established, with
the ability of many employees actively sought. The internal social plat-
forms are then used to access information, with a focus on collaboration
via social tools. McKinsey (2012) argues that this can improve staff pro-
ductivity by up to 25% and increases innovation and flexibility as new
ideas are discussed by a selected wider circle of experienced staff, rather
than imposed by upper management. These internal social networks
 5  What Insurance Companies Need to Do    
137

will form the exoskeleton of the company and will result in richer, more
engaged, and more flexible company structures which don’t rely on
command and control hierarchies to maintain coherence.
The key reason for slow change among insurers has not been a lack
of foresight or a lack of investment, but has been inadequate thought
about what the implications of ‘what-has-yet-to-occur’ which is difficult
as the impact of WHYTO will always be uncertain and not posing a
definite threat. To counter this requires a change to a flexible style of
management which is opposite to traditional insurance risk-reduction
modes. This has to start with the CEO, as most employees are paid to
optimize the present, not to prepare for the future.
An example of the use of internal social networks is Canada’s TD
Bank Group, which after a series of mergers found itself internally dis-
jointed. Launched in November 2001, TD’s internal ‘enterprise social
network’ has 4000 communities and thousands of blogs and wikis. It
is used to communicate within company-wide networks of teams, to
share expertise, to offer advice, to collaborate on projects, and to find
relevant knowledge from employee profiles. TD found that the social
network cut down drastically on phone calls, e-mails, and meetings. In
particular, unwanted or irrelevant e-mails or meetings were largely elim-
inated. Most importantly, the network gives management the chance to
lead and supervise in new, more flexible, ways. For example, sales man-
agers can communicate with sales staff directly without going through
branch managers. Status boards keep groups up-to-date. The collective
intelligence of nationwide groups can be accessed and internal cohesion
enhanced without gathering staff for conferences. It recognizes that the
best ideas can sometimes come from junior staff, even cleaners or ware-
house packers.
Implementing major technological, managerial, and cultural change
is extremely complex and is frequently unsuccessful. A substantial issue
is that investment in new systems has to occur while existing systems
are maintained, even though the new systems may cannibalize the old.
There are often major management hindrances. Eastman Kodak did not
fail to transform because of lack of technical capacity - they invented
the digital camera, or lack of customer demand - people still took pho-
tos. Kodak failed predominately because it did not market its new
138    
M. Naylor

technology for fear of harming its old business. This held back change
for a while but meant that when change did occur Kodak was swept
away by it, rather than leading it. Insurers will only survive the com-
ing disruption, will only lead it, if they acquire a senior leadership team
which is dedicated to internal and external transformation and has the
expertise to implement this.
Traditionally, however, insurers have a culture of confidentiality and
discretion that works against the blossoming of a free-wheeling social
network, whether external or internal. Regulations also limit the extent
and scope of insurer communications with clients. So far, most insurers
have limited social technology investments to minor areas like market-
ing functions and have not attempted to embed deep collaboration -
they are still talking to clients and staff rather than talking with clients
and staff. They still tend to view adverse customer or stakeholder feed-
back as an area to be restricted, rather than a quality growth opportu-
nity. Insurance laggards fail to recognize that in today’s world online
communities will engage in discussion and reviews of their products
regardless of what they company does - the only question is; Is this dis-
cussion happening with the company proactively involved or is it hap-
pening with the company excluded?
Given that these approaches require a substantially different style of
management, and will take time to learn, insurers who delay experi-
ments in learning the skills will find themselves rapidly falling behind
rivals.
Based on these findings, the majority of insurers will probably not
cope with the coming perfect storm of disruption, with at least half
of existing insurers likely to face rapidly shrinking market share. Only
about 10% of existing insurers are making the transformation changes
required to ensure sector leadership. The leaders of the insurance sec-
tor in the future are likely to be external firms with outstanding skills
in areas, like customer relationships, which insurance products can
be added on to. Note that adding insurance as a product to a strong
external disruptor is easier than transforming an existing weak insurer.
It is arguably likely that the leading insurer of 2030 has yet to enter
insurance.
 5  What Insurance Companies Need to Do    
139

An example of an outside entrant is the Malaysian Tune Group,

which has used its subsidiary’s, Air Asia, extensive IT-based customer
data to enable the administration and distribution of insurance. It
started in 2011 in travel insurance and then moved into life and non-
life products, selling over 7 million polices by 2013, across Malaysia,
Indonesia, Thailand, Singapore, and China. It currently has over 2000
agents, is busy opening physical branches, and is negotiating to buy
major existing regional insurers. By 2016, a mere 5 years after it started,
it is the leading SE Asian digital insurance franchise. Which SE Asian
insurer foresaw that a discount airline would become a major threat?
Note that there are large financial implications for insurers in moving
toward a more flexible, real-time feedback model. Insurers are used to
the stability of fixed 12-month premiums so the use of dynamic short-
term premiums could create uncertain fluctuations in cash flow which
could be destabilizing. This will be easier dealt with by external firms
which have existing large and uncorrelated cash flows.

End-to-End IT Systems
The heart of future insurance companies will be a high-quality inte-
grated IT system, which aggregates, analyses, and provides data to users
in a high-quality friendly format. Insurance has always been a data-
intensive business and the future world of a tidal wave of big data will
reinforce this trend. Survivors will be those who can undertake this
IT transformation faster and more effectively than their competitors,
including external disruptors.
Historically, companies have favored an incremental approach to IT
changes, addressing the immediate point of pain and then handling
subsequent issue as they arise. It has been viewed as just another shared
service, to be allocated a budget appropriate to the urgent issues need-
ing a solution. This causes issues as technology teams in different parts
of the company independently address issues in discrete systems, hence
creating islands of solutions, breeding complexity, and fragmented and
redundant systems. For example, most insurer documents and com-
munications are created via MS Word templates, yet these are difficult
140    
M. Naylor

to extract data from, do not allow flexible formatting to differing plat-

forms, do not allow easy integration with differing data formats, and
use-activity is hard to track. These templates need to be in more digital
friendly formats, which allow encryption and activity tracking. These
systems need to allow two-way interactions, so customer inputted data
or corrections are noted system-wide.
Neirotto and Paolucci (2007) show that most IT system invest-
ments in insurance have failed to deliver what was intended, with lit-
tle correlation between IT investments and resultant improvements in
competitive advantage. They argue that this has been because insurer
IT investments have not been implemented with an adequate strategic
vision. This is an issue as Breznik (2012) shows that successful strategic
IT investment is a growing source of competitive advantage.
In the future, an insurer’s IT system needs to be viewed as the com-
pany DNA, its core, around which everything else is built. This inherently
means that the IT transformation is too important to be left to the CIO
and the coders. The inputs and outputs need to be understandable intui-
tively by everyone in the company including customers. The human–com-
puter interaction and data visualization aspects are vital. A superior system
will enable workers to complete tasks faster and cheaper than competitors,
and customers will judge the quality of an insurer’s service primarily based
on the ease of use and adaptability of its omni-channel interfaces.
It needs to be remembered that that cost of a legacy system is not
the cost of the physical system, but the cost of the inefficient processes
which surround that legacy system, as well as the opportunity cost of
what the legacy system doesn’t allow you to do.
A blockchain payment systems joined to smart contracts and elec-
tronic notarization will enable the simplification of administration
chains, cut costs to pennies, and provide clear evidence and transpar-
ency of which actions have been taken, as well as providing a data evi-
dence chain to follow the payment process. This will be the basis of the
micro-payments which will underlie dynamic insurance. It will also ena-
ble mobile phone-based micro-insurance3 in emerging markets or for
younger customers.

3For example, it is common in Africa for a micro-insurer to offer a $150 life policy costing a few

cents each week added to the mobile phone bill. Why not do this for renter’s insurance?
 5  What Insurance Companies Need to Do    
141

Bain (2013) argues that the keys to successful IT transformation are:

(i) getting the basics right - so that customer access speeds and reliability
match that of social media firms; (ii) use a ‘service-orientated architec-
ture’ to extend existing systems in modules, rather than try to replace
the entire current system; (iii) focus on the highest priority areas first,
so to get some wins; (iv) integrate technology awareness and systems
into all business units, so IT is not just an issue for one department;
(v) think primarily from a customer viewpoint, by understanding what
their priorities are; (vi) track technological trends, and trial which inno-
vations could be useful, and become fast followers of trends; and (vii)
segment customers and defend high-value ones while discouraging low-
value ones, offering the latest services to high-value customers. Tim
Yorke4 argued that the way we get away from legacy is to stop it being
the monolith and putting architecture in places that allow you to plug
things into an architecture that is flexible.
The rebuild of the IT system needs to occur in an end-to-end fashion
with usability and flexibility as the core concepts. Three key strategic
steps are required: (i) defining the end IT system, (ii) deciding which
elements of the existing system have to change, and (iii) determining
the sequence and scope of the change. The end IT system has to be
defined in terms of the tasks it is required to achieve, divided into essen-
tials and nice-to-haves. Example questions would be ‘what information
is required to actively excite customers about their insurance product?’
or ‘how is information best presented to actively engaged customers?’ or
‘where and how is data best stored so it is available to those who require
it?’
The end objective of the rebuild of IT systems has to be to drastically
reduce distribution costs and cut the marginal cost of offering clients
choices to near zero. The system must have inherent flexibility as the
transformation is not a one-off threat, but a change to a faster rate of
change, a transformation which has no endpoint.

4Tim Yorke, COO, ERS, Insurance Post, Aug' 16.

142    
M. Naylor

The survivors will need a firm vision of a future of insurance which

involves dynamic, flexible, modularized, customer-friendly products.
This needs to be able to handle an ever-swelling tide of customer data,
in a way which enhances the customer experience. Customers need
to be provided with products which excite and a customer experience
which matches what they experience from their favorite social media
channel. Ad hoc amendments to existing systems will not achieve this.

Cost, Scale, and Platforms

A vital aspect of the required changes which insurance survivors need
to undergo is reducing the marginal cost of transactions close to zero.
Customization and dynamic premiums are not possible if the cost of
each transaction is measured in hundreds of dollars and not in cents.
The advantage of an end-to-end intensive IT system is that it allows this
level of marginal cost to be achievable. The issue with these intensive
IT solutions is that they have a high establishment and maintenance
cost, defined as fixed cost. This type of high-fixed cost, low-marginal-
cost environment obviously means that the larger a firm is, the lower
average cost becomes. Added to this is the requirement of massive data
sets before AI systems work efficiently. Hyperscaling is thus important
as there is a massive scale advantage.
Against this scale advantage is the issue of IT complexity. It is a well-
known problem that large IT transformation projects typically fail due
to complexity; the larger the firm is, the more complex are the inter-­
relations, the program requirements, and the reform politics. Smaller,
simpler, firms can thus create IT systems which are more efficient and
can reform these faster. Since these new systems should have costs as
low as 10% of the old systems, and given the increasing rate of disrup-
tion, flexibility and speed could be more important than scale and end-
point cost. Thus, a key survival skill for insurers will be the ability to
remake their IT system periodically.
It needs to be emphasized that in the future, the IT system will not
just a stand-alone inside-company system. Successful insurers will need
to work with a wide range of other firms in a business ecosystem, as
 5  What Insurance Companies Need to Do    
143

discussed in a subsequent section. At the heart, this ecosystem will be an

IT ‘platform,’ which links all the firms in a continual flow of data. The
insurer’s internal IT system will have to link seamlessly with this plat-
form. The firms who control these platforms will dominate the future.

Flexible Programming
The two key requirements for future insurance business systems will be
‘control’ and ‘flexibility.’ Control is needed because mistakes in insur-
ance can have major cost implications. Flexibility is needed to cope
with ongoing transformation. Because programming large systems can
be overwhelmingly complex, however, most systems of programming
traditionally use a set step process of (i) ascertaining system require-
ments, (ii) defining system parameters, (iii) defining data, (iv) defin-
ing user-interfaces, (v) coding, (vi) testing and error handling, and then
(vii) implementation.
Two major problems can occur during IT project creation to disrupt
this sequence: (i) the external environment changes, so system require-
ments change, and (ii) as the implementation process progresses, the
understanding of non-IT staff grows so they think up a series of new
innovations. These are both valid reasons for amending system specifica-
tions. The problem is that a key business lesson is that IT projects tend
to fail when the process is interrupted by changes in system require-
ments. Programming of large-scale IT projects is very complex as all
elements are linked so that changes in one area can cause unexpected
errors in another area. Programmers, therefore, prefer to complete the
project to original specifications and then subsequently make ad hoc
changes. This is done by dealing with requests one at a time, with exten-
sive testing before each change is implemented.
Yet continuing with the original project when it is obvious that
the external environment has changed can also often be a waste of
money. This is particularly true if required changes are to the under-
lying business logic, so that ad hoc changes to a now ill-specified sys-
tem open it to instability and inefficiency caused by a stable core system
being disrupted by a growing mass of add-ons. The extreme difficulty
144    
M. Naylor

of implementing large IT projects is why incumbent insurers are at a

disadvantage against IT savvy external disruptors who have an up-to-
date internal system. These new systems are based on newer inherently
agile programming approaches.
One of the main defining characteristics of insurance survivors will
therefore be expertise in IT system speed-to-market and system flexibil-
ity, combining control and flexibility.
Progress (2014) argues that one way to achieve this agility is by
(i) breaking up IT projects into modules and (ii) using a ‘business-rules
management system’ (BRMS). This is a decision system which separates
the key business decisions from the rest of the programming activities.
The essential problem is that for a programmer, key business require-
ments are no more important in coding than, say, data retrieval pro-
tocols. Progress (2014) argues that what a BRMS does is to separate
out the key business aspects and by using an intuitively easy dashboard
screen, put the ability to change these rules under the control of key
managers or the relevant line manager. This enables business users to
author, validate, test, and deploy rule sets as decision-making services
with minimal IT knowledge. This places control in the hands of busi-
ness users and sharply reduces time-to-deployment, hence creating IT
flexibility. For example, for a quote calculation, the inputs and outputs
would be agreed on by the managers and the programmers. Then, the
premium calculation is defined as a set of business rules, within the
stand-alone BRMS. This means that the managers know how to define
the rule and the programmers know how to request a decision. This
BRMS is then linked to all other areas of the IT system. The BRMS
separation means that managers can change the premium calculation
without impacting on other programming activities, as the only change
would be in the premium outcome. This allows ongoing business flex-
ibility while allowing the programmers to focus on usability, reliability,
and performance.
BRMS systems should present decision rules in standard language
to managers and allow extensive pre-deployment testing. This enables
managers to play around with systems outputs in terms which users
understand, and allows selected users from a wide range of areas to trial
 5  What Insurance Companies Need to Do    
145

changes and see if they misunderstand uses. It is thus vital that the inter-
face is user-friendly. The ability to make substantial changes to IT sys-
tems while the system is active has a transformative power for insurers.
An illustration of the problems inflexible software systems can cause
can be seen from issues encountered with the New Zealand health
insurer, Southern Cross’s employer health policy. This policy is set up so
it covers the employee and their family, with the employee as the ‘policy
holder.’ Any claim has to be made by the policy holder, and all commu-
nication has to go via the policy holder. As there can only be one pol-
icy holder, Southern Cross’ systems cannot cope with both of a couple,
husband and wife, working for the same employer. The result of this
inflexibility is a 1950s’ Kafkian nightmare world, whereby if the hus-
band is the policy holder, then the wife has to ask her husband to make
claims, OK all procedures, and handle all correspondence. The wife is
treated on the same level of competence as the children. Southern Cross
are aware of this issue, but (i) the IT issues involved in making a change
to the modern world are beyond them, and (ii) individual staff have no
incentive to take charge of an issue and ensure it is solved. Thus, the
1950s continues to live on. Compare this to Tesla, who when notified
of an issue, immediately accept responsibility, proactively work out a
software fix, and upload that to all cars before 99% of owners are even
aware the problem exists.

Digital Laggards
Most insurers are failing to meet the challenges described above. EY
(2013) finds that nearly all insurers are currently digital laggards, trail-
ing substantially behind most sectors in their use of digital channels and
are not in-the-play in terms of interactive customer relationship soft-
ware. International surveys consistently find that insurers are regarded
by customers as the most backward service sector in terms of customer
experience. KPMG (2016) thus argues that incumbent insurance
companies are now in an epic battle to transform themselves to avoid
becoming disintermediated by new digital competitors.
146    
M. Naylor

EY (2013) also finds that while insurers have aspirational digital

ambitions, these are predominately unrealistic, as insurers do not have
the flexible management systems, the staff, or the experience to com-
pete with digital leaders; especially, as those digital leaders are them-
selves enhancing their expertise at a rapid pace, insurers are chasing a
rapidly moving target and thus increasingly open to technological dis-
ruptive entry by those digital-savvy firms. In particular, expertise in
inter-active social networks and use of mobile screens are weak, even
though strength in these areas is vital to future success.
PWC (2014) argues that insurers are trying to meet the challenge by
primarily focusing on e-commerce: that is, doing what they are already
doing, but digitally. For example, nearly all incumbent insurers have
inflexible, form-based, administration procedures. Insurance forms tend
to be paper forms which have been digitalized, rather than rethought.
While there may be legal reasons for forms to go back and forth, in
90% of cases, the procedure can be simplified, and forms made inter-
active. Insurers need to rethink chains of procedures from a customer
viewpoint.
Only a few insurance leaders are taking the next step of utilizing their
data capacity to develop deeper, more personal, and longer-lasting cus-
tomer relationships, allowing them to move away from price. PWC
(2014)’s survey finds that customers are more willing to engage with
insurers than the products and Apps which insurers are currently offer-
ing allow them to. For example, 67% are willing to use telematics to cut
premiums.
Customer relationships are also lagging. Morgan Stanley/BCG
(2015) finds that currently insurers typically have one or less interac-
tions with customers per year, and typically when contact is made,
normally at claim time, the experience is disappointing. Their research
shows that in terms of customer satisfaction with the online experience,
insurance is one of the lowest ranked industries, 13th from 16 indus-
tries. The style of insurer–customer interaction makes this disappoint-
ment inevitable. This dissatisfaction is particularly high for small and
medium business owners. Customer satisfaction decreases as they move
from purchase to maintenance to claims to renewal. Therefore, the
 5  What Insurance Companies Need to Do    
147

insurance market is extremely vulnerable to disruptive entry by firms

with expertise in intensive and superior customer satisfaction.
Insurers have put substantial effort into market research and surveys.
Despite this, they are currently struggling to understand the fine detail
of the difficulties which customers are experiencing and the complex-
ity of the experience from the customer angle. Surveys persistently show
that customers rate the digital experience provided by insurers far lower
than the insurers rate the experience they think they provide - insurers
typically do not understand how backward their customer experience
is. The current approach is not working; insurers are increasing being
regarded by the younger generation as IT dinosaurs.
For example, many insurers have created a slick Web site, added a
Twitter app and a Facebook page, and then have been stuck for what
to do next. KPMG (2014) shows that while 69% of insurers say they
have a digital strategy beyond a Web page, only 37% say it is aligned
to the company’s strategic objectives; 90% have only a limited idea
about how to move beyond a ‘push product’ approach toward two-way
active interaction. The few who have tried two-way communication are
often burnt by internet trolls and stuck on how to handle this. They are
uncertain about how to create regular contacts with their mostly older
client base in a way which is useful to clients and thus welcomed by
them rather than being seen as a distraction in busy lives. This shows
that these insurers are missing the bigger picture that organizational
change is needed. They are unaware that insurers must move beyond
talking to customers to talking with customers. These are basic issues.
Transforming the company using a big data intensive approach
would be a good start to solving issues. An example of how insurers
could meet future challenges is the way the use of customer telematics
can enable insurers to access the real-time feedback to integrate them-
selves into customer’s lives, so they become irreplaceable partners and
can offer value-added services. PWC (2015a) envisages areas like insur-
ers being instantly aware if clients’ vehicles or machinery breaks down
and thus arranging repair crews, or being aware if clients are exercis-
ing and improving their health and thus reducing premiums, or being
aware of driving errors and increasing premiums.
148    
M. Naylor

Another example is the use of survey or social media comments,

where currently insurers devote substantial attention to have staff manu-
ally reading through. These can give rich insights into customer experi-
ence and is an area of future competitive advantage for the survivors. Yet
in an era of increasing online commentary and rapid response expecta-
tions, it is simply not possible to manually analyze more than a sample.
Comments need to be analyzed by language-based software, with staff
only altered to key aspects or trends. Software should be able to respond
by itself to the majority of generic issues. Yet, very few insurers world-
wide have robust social media data AI capacity.

Omni-Channel Response
Floreddu and Cabiddu (2014) define an omni-channel response as sim-
ply meaning ‘enabling a broader range of channels for establishing per-
sonalized relationships with customers.’ However, it really means more
than this, as an effective response to the threat of social media entrants
requires a revolution in the way insurance customers are treated and the
way information inside an insurer is collected and organized.
An omni-channel response does not mean new approaches simply
layered on top of old; instead, the channels have to be integrated into
a seamless fashion, so that customers can access any information or ser-
vice they need via any channel. Any staff who needs an aspect of cus-
tomer information, whether call center or actuarial or claims adjusters
or marketing, needs that information in the most appropriate manner
with key details emphasized. This means that if a customer initiates an
inquiry via e-mail, then explores via the Web page, sends a text, and
then phones a call center, all the queries and information the customer
has provided, and the topics they are interested in, are available to the
call center worker.
McKinsey (2012) notes that reliance on personal visits or phone
calls or newspapers for product sales has just about died, and radio
and TV viewing is increasingly restricted to the older generation. The
insurer response has to be more than merely setting up Facebook or
Twitter accounts; it is active engagement in a wide array of social media
 5  What Insurance Companies Need to Do    
149

platforms. For example, if a customer is on the Web, they may prefer to

engage with a person via Skype window in the Web page than a phone
call. This is particularly important in India and Asia, where over 50% of
customers share purchase recommendations via social media.
The focus should be on finding out the customer’s problem and
actively solving it, rather than actively pursuing a sale. This needs to be
as efficient and low cost as possible. For example, 90% of quotes need
no human involvement so should be instantly assessable once customer
details have been imputed, whether this is via an agency, or the internet,
or Facebook, or text, or phone app, or quote aggregator, or a broker. A
field agent should be able to sell an underwritten policy, or process a
claim, on a mobile device at one meeting, as long as standard parameters
are met. Most insurance types should be able to be purchased, amended,
and claims made via any channel. For example, NTUC Income, a
Singaporean insurer, allows customers do all three via a phone app.
When a potential customer calls a contact center or into a branch, all
the information which they imputed online should be available to the
agents so there is never a need for customers to provide any information
more than once. In a health insurance claim, for example, the doctor
should be able to upload and sign documents online, the client should
be able to Skype the claims adjuster and area specialist simultaneously,
and thus have any claim paid within 2 or 3 minutes of the insurer being
aware of it. This is only possible if all claims administration and check-
ing are done instantaneously via software.
Poor customer experiences are increasingly destructive as customers
have the means via social media to distribute their complaints to more
than their friends. While previously companies facing bad news sto-
ries could delay and diffuse its effect, this is very hard to do with social
media spreading news and opinions about it. Required response times
to adverse news or complaints are now in minutes or hours, not days.
Millennials and later generations, in particular, are exhibiting a
strong tendency to switch suppliers based on reputation or social media
feedback. While the proportion of customers who currently use social
networks to make product decisions is currently low, McKinsey (2012)
argues that this will rapidly increase. Yet, only a few insurers encourage
online unaltered reviews of their products and performance.
150    
M. Naylor

Social Dynamics
An increasingly important aspect of social networks is the ability of
people to form groups with like-minded individuals across the globe,
even if they have never personally met them. The social bonding in
these groups has been shown to be strong, and thus, they have a pow-
erful influence on product decisions. Any positive or negative service
experience to one member is thus likely to have a wide-spread ripple
impact. Some of these can involve members with high-level skills in par-
ticular areas who may thus become highly influential opinion makers.
Failure to actively engage with these opinion makers can be disastrous,
yet insurers are traditionally very cautious about engaging in debates
about their products or customer relationships even within traditional
avenues and almost never debate within the newer media. They need to
adopt the new approach of being extremely open with information and
actively engaging with society. They can only successfully do that if they
have quality products and robust procedures which they are proud of. It
is all about customer trust and producer reputation. Social media and
online review sites can be powerful tools but because of the tendency of
social media to attract a lunatic fringe, successful use can be difficult.
Producer reputation is thus growing in importance, yet is still not
actively managed by insurers. An issue is that insurers may have find
their social capital eroding without any way of tracking which poten-
tial customers are alienated, as currently insurer data systems can do lit-
tle more than identify which customers had a bad experience. Insurers
have little data on non-customers and why they choose to not purchase.
Verint (2016) reports that the majority of customers who have a posi-
tive experience with an insurer will not give the insurer feedback, but
will tell friends or write a positive review. The behavior of a client’s
social network is very powerful. For example, research has shown that
if a person switches mobile phone providers, their friends are 7x more
likely to also switch.
Data systems need to be able to identify exactly which areas of the
process caused the issue and determine whether this is due to customer
differentiation or generic failings. These need to be achieved in real time
(inter-daily), with the results distributed on social media, so potential
 5  What Insurance Companies Need to Do    
151

customers who are following the complaint are impressed rather than
discouraged. Modern analytics can achieve this, yet are little explored by
90% of existing insurers. The data will soon be available, and if incum-
bents do not proactively use it, then external disruptors will.
Effective client engagement software systems also have the poten-
tial to drastically reduce client contact costs. There are significant scale
economies to these systems, so there is a strong first-mover advantage,
with innovators likely to attract younger, lower-risk, higher-profit cus-
tomers, thereby trapping firms with lagging legacy systems in a down-
ward spiral of shrinking non-digital customer base with increasingly
higher risk and higher cost.
Effective insurer customer engagement systems have the potential to
revolutionize the fortunes of successful companies. The key is to turn the
currently typical unengaged insurance customers into engaged custom-
ers, who benefit from frequent positive insurer contact and who pro-
mote the benefits of the insurer within their social network. Bain (2015)
argues that promoting-insurance-customers stay longer, buy more, bring
in extra customers, and cost less to service. Their lifetime value is worth
more than seven times the value of disengaged customers. Once insur-
ers take the next step of using feedback from, and proactive engagement
with, these promoting customers to change their products and processes,
the value of these promoting customers doubles that figure.

Active Engagement
Active engagement with customers is thus vital. Yet, most insurers typi-
cally only interact with their customers at initial sales, at renewal, or at
claim time. These give customers a very shallow and negative impression
of the product, as their experience is restricted to the insurer demand-
ing money or asking tough questions. This naturally leads to most con-
sumers having a negative impression of insurers. The Edelman Trust
Barometer for 2010 found that only 41% of customers trusted insur-
ers ‘to do what is right,’ versus 71% for technology firms. The major
reason for this different was a lack of positive engagements, with 90%
of insurer engagements found to be stressful or negative. Surveys of
152    
M. Naylor

those who have made major claims find that about 70% felt that the
insurer ‘did not care.’ This is despite the industry rapidly paying 95% of
all claims. What is missing in insurer–customer contact is not efficiency
but empathy.
In contrast, digitally savvy companies monitor their customers on
a continual basis over a wide range of channels. For example, a travel
booking Web site, like hotels.com, uses its software to monitor a cus-
tomer’s travel progress and send updates by e-mail about possible travel
issues, or to send useful hints. They heavily encourage customer feed-
back about hotel experiences and have a generous refund policy if issues
arise. A single five-day trip may generate five or more customer contact
opportunities, and knowledge learnt from customer feedback will be fed
back into future contacts. The customer leaves with the impression of a
company which engages and cares. Contrast this with a typical insurer -
who may have non-positive contact once a year with a customer. These
disengaged customers are receiving frequent, positive, contacts from a
range of firms in other sectors, firms which could easily add insurance as
an additional product.
This nonexistence of a sequence of frequent, positive, customer con-
tacts is the basis of the current failure of the insurance experience. It
is the main reason why customers feel disengaged from insurance,
and why insurance is regarded as a product to ‘be sold, not brought.’
Insurers need to adopt a new mission - ‘to make insurance joyful.’
One lesson from banking is that once customers could view their
account balances online, they interacted with the bank Web site on a
far more frequent basis than expected. One of the keys to a successful
digital transformation for insurers is thus the use of social media and
big data to create a sequence of continuous, enhancing, positive cus-
tomer contacts. These contacts should be relevant and useful to the cli-
ent, and include discounts or promotions tied to life events and loyalty
programs, so that they are welcomed by the client as useful, interesting,
and educational.
Insurers need to create some activity which incentivizes custom-
ers to interact frequently. Real-time dynamic premium pricing systems
could achieve this. Research has shown that customers get an emo-
tional reward for even small price changes resulting for their beneficial
 5  What Insurance Companies Need to Do    
153

actions. These frequent contacts can be linked to news, to education,

and to pricing specials. Innovative new idea could be trialed, like offer-
ing a 40% reduction in this weekend’s driving premiums to those who
achieve a set safe driving level for the telematics. These could be linked
to specials from businesses in the wider ecosystem. This embeds the
active customer engagement. A useful part of this engagement comes
from friendly empathic staff. Therefore, management engagement inside
the company is also vital. It also needs to be remembered that distribut-
ers are digital clients as well.
This type of engaged contact will also help at claim time. Many life
and health insurers only assess client data at application time. Much
data is second hand - basically insurers know very little about each cli-
ent. The use of telematics and other data sources on a continuing basis
qualitatively transforms the insurer knowledge base, enabling person-
alized underwriting and personalized advice on reducing their risks.
Higher-risk clients can have their health condition monitored, and if
they do not follow treatment protocol, insurers can increase premiums
or withdraw cover.

Innovators
An example of an engaged innovator is AllLife Insurance which offers
affordable life and disability cover to those suffering from existing con-
ditions, who most insurers regard as uninsurable. By getting clients to
agree to intensive monitoring and strict medical protocol, AllLife has
both improved client health outcomes and increased profit, and is cur-
rently growing at 50% per annum.
Another example of an engaged innovator is the South African
insurer, Discovery, whose health focused product, Vitality, which gives
incentives for healthy living in ways which align with customer priori-
ties. Discovery has consequently experienced rapid growth, while lower-
ing customer risk profiles and cost. Vitality collects a significant volume
and variety of client data, including wearable devices, and has discov-
ered new links between client activities and health risks. It has discount
arrangements with gyms and monitors client purchases of health food.
154    
M. Naylor

Clients are sorted into groups and are given vitality points on a real-time
basis based on health-promoting activities, with higher groups offered
substantial discounts on insurance and health-related products. Clients
are given news on health-related areas. These innovations have for the
higher groups reduced lapse rates by 52% and reduced mortality rates
by 34%. Discovery has used its leading-edge software and engagement
systems to expand rapidly, with joint ventures in the UK, the USA,
China, Singapore, Australia, and recently Europe.

Competitive Advantage and Dynamic Capacities

One of the fundamental issues is how an insurer can achieve and sustain
competitive advantage within a modern fast-moving business environ-
ment where ownership of scarce or rare resources is now insufficient.
Teece et al. (1997) define ‘dynamic capacity’ as ‘the firm’s ability to
integrate, build, and reconfigure internal and external competencies to
address rapidly changing environments.’ This ability to react dynami-
cally is now more important than any existing advantage which the firm
may hold over competitors. This is especially true in a case where poten-
tial competitors may arise from outside the sector, so management may
be unaware of comparable strengths and weaknesses, thus the threat
may arise unexpectedly.
Floreddu and Cabiddu (2014) disaggregate dynamic capacity into
the capacities (i) to sense and shape opportunities, (ii) to seize oppor-
tunities, and (iii) to maintain competitiveness through enhancing,
combining, protecting, and reconfiguring the tangible and intangible
assets. This is essentially a firm’s cultural capacity to visualize a different
future of as-yet-to-occur challenges and rearrange physical and personal
resources to meet those challenges in a manner which exceed the capac-
ity of future competitors. The leading-edge use of IT and other technol-
ogies is a vital part of dynamic approaches, but they are tools, not core
capacities. The core capacity is cultural.
Floreddu and Cabiddu do, however, argue that effective IT systems
and innovative IT staff add substantially to a firm’s dynamic capacity.
To create an effective omni-channel response, they argue that firms need
 5  What Insurance Companies Need to Do    
155

to evaluate existing IT resources, data, competences, and capacities to

deploy different channels, what marketing activities implement to sup-
port the omni-channel strategy mix, and competitor capacities. IT
resources need to be understood in terms of activities rather than jobs
and modularized so that reconfiguration is easier to implement. Given
that IT-based information flow will be at the heart of surviving insurers,
required future skill sets need to be identified and capacity sourced.
Omni-channel approaches require a change from data collection and
management focused separately in each channel, to a customer-centric
approach with data from all channels integrated. This should also give
managers a richer data resource on customer behavior and start to allow
insurers to switch to a two-way rather than a one-way dialogue with
customers.

Peer-to-Peer Insurance
Peer-to-peer insurance is a new and growing area. It involves groups of
customers jointly insuring each other. There are many types of peer-to-
peer insurance. The first type is often created by a third party, like an
insurance broker who helps form customers into small groups, gener-
ally via a Web site. A part of the insurance premiums flow into a group
fund, the other part to an insurance company. Minor claims by a mem-
ber are firstly paid out of this group fund. For claims above the deduct-
ible limit, the insurer pays out. If there are less than expected claims, the
member either gets refunds or a credit toward the next policy year. The
broker is paid via a commission.
The second, more recent, type is similar to the broker model except
that as the peer-to-peer provider is the actual insurance company. If
the pool is insufficient to pay for the claims of its members, the insurer
pays the excess from its retained premiums and reinsurance. If there are
lower than expected claims, the ‘excess’ is given back to the pool or to a
cause the pool members care about. The insurer takes a flat fee for run-
ning the operations of the insurance enterprise. A third type is where a
group is set up by the members themselves, via a social or Web-based
network.
156    
M. Naylor

The only real requirement behind these groups is that all group mem-
bers must have the same type of insurance or something in common,
such as being members of the same club or believing in the same charity.
In structure, peer-to-peer insurance is very similar to traditional
mutual insurers. The main difference is the use of automated software,
and Web distribution has substantially reduced administration costs. In
that sense, while the growth of peer-to-peer insurers may reduce tradi-
tional insurer market share, the innovation does not involve any sub-
stantial alteration of the insurance market as a whole.
Some P2P insurers, however, are making use of intensive end-to-end
software systems to rapidly expand. They also use big data and behavio-
ral algorithms, speed of processing, and easy-to-use client interface, to
make themselves feel natural to digital natives.

Cyber-Insurance
One of the future growth areas is cyber-insurance. PWC (2015a) notes
that cyber security policies will be a rapid growth area, with insurers
able to gain service-based revenue from monitoring cyber-attacks and
proactively responding. The majority of corporations are only starting
to recognize their vulnerability to cyber-attack even though there are
currently 10,000 attacks a day worldwide. These attacks are starting to
transform from minor incidental damage to major deliberate attempts
to delete and destroy operating systems.
Defense against these attacks is well outside the capacity of most
firms. For example, software producers post details of problems with
their software on their Web sites, alongside the required upgrade patch.
Reading these notices gives hackers explicit instruction on the vulner-
ability and thus how to exploit it. Software producers expect companies
to download the patch asap. Currently, hackers will attack half of all
computers within 28 days of a patch being released. Despite this being
low-tech hacking, surveys repeatedly show that less than 30% of firms
regularly download these patches, so that 70% are vulnerable. Research
shows that 95% of firms do not have the capacity to proactively react to
high-tech cyber-attacks.
 5  What Insurance Companies Need to Do    
157

The damage cyber-attacks cause, their cost, and the intensive nature
of the response required are a huge opportunity to an insurer who can
offer specialist expertise in data security and can alert clients in real time
to attacks. It needs to be noted that as both industrial and service sec-
tors integrate their entire production systems into complex software sys-
tems, hackers will not only be able to access data, they will also be able
to disrupt production. Autonomous networked cars are at high risk and
could be used in terrorist attacks. Even households will be vulnerable
once house appliances are integrated, as net-linked appliances have very
basic security and most owner leave them on the default password.
This market is substantial and is growing at over 30% per annum, so
it is a useful source of new revenue. The advantages to insurers of being
able to convince clients to allow real-time integration of the clients’ tele-
matic systems with the insurer’s security alert systems are immense, as
not only can the insurer take immediate action to stop failures as they
occur, but also the access to a rich stream of real-time data on failure
will allow them to gain new insights into causes of failure, and thus pre-
vent them.
Even if insurers get a strong market share however, cyber-insurance is
very difficult to price as it is a very new area and strongly subject to con-
tagion and the possibility of catastrophes/black swans. It is like earth-
quake insurance in that there could be a long period of calm and low
claims, leading to low premiums, and then an out-of-scale loss due to
a worldwide software virus. Therefore, reinsurance will be vital in this
area.
The major issue for insurers in this area is that success in this area
demands extremely high levels of IT skills and very fast response times.
Customers are far more interested in a company which can offer pro-
active real-time protection than one which merely offers compensation
after the event. Many new, currently unimagined, issues will arise, so
insurers will have to possess a very high level of IT skill to stay up-to-
date. Surveys find that most insurance managers are scared of the dif-
ficulties involved in combating hacking, rather than excited by its
business opportunities. Actuaries regard the area as the preserve of
techies. It is thus more likely that IT firms will dominate this area than
insurance firms will.
158    
M. Naylor

Business Ecosystems
Insurers who respond aggressively and proactively to the opportunities
offered by big data, telematics, real-time dynamics, and complex system
analytics will leap ahead of their competitors who do not, as the per-
transaction costs will drop by multiples. Note that this involves linking
the insurer’s data analytical system into customers’ Webs of telematic
items in a way which is very foreign to how insurers currently operate.
A key aspect of this is using analytical predictive systems to move
away from compensating clients when things go wrong, toward predict-
ing when things will go wrong and acting to stop problems. This is the
kind of value-added services which will create the profits which will dis-
appear from existing areas. While overall per-customer profit margins
may shrink, innovators will compensate by an expanding market share,
as well as expansion into new higher margin areas.
There are two aspects to this reinvention: (i) cutting distributions
costs from over 100% of initial year premiums to closer to 5% and (ii)
integrating the products into everyday life so they are seen as relevant.
While this reinvention will probably occur initially in car insurance, it
will rapidly spread out into all product lines.
Another aspect is the recognition that the range of specialist skills
required to achieve sector leading expertise in each area will expand past
what a typical single insurer can be expected to create. Insurers will thus
have to make strategic decisions to specialize in a narrower range of skill
set and to source skills they can’t provide themselves from permanent
partners in related areas. The success of insurers at integrating them-
selves into a wider at this network will be a defining feature of survival.
Morgan Stanley/BCG (2015) thus argues that an increasingly impor-
tant competitive concept will be that of an interconnected ‘digital eco-
system’ of related companies, who interact to create combined services
and generate value, rather than as large vertically integrated stand-alone
bodies. This is not a new idea, as Iansiti and Levien (2004)’s survey
shows, and originates well before the internet. They show that the use of
ecosystems is a vital aspect of Walmart’s and Microsoft’s success. Despite
this, the majority of management education focuses on the internal
operations of a single company rather than how to integrate a network
of assets, the majority of which your company does not own.
 5  What Insurance Companies Need to Do    
159

What is new is thinking of a network in terms of data flow, rather

than physical product flow. The ecosystem is enabled by a central stand-
ardized software/technical platform, a ‘keystone,’ which connects pro-
ducer and customer devices, applications, data, products, and services.
This keystone is created by a platform owner, who allows producers and
data services to interact in a customer-centric manner, and whose IT
platform integrates the ecosystem. Note the key role of telecom provid-
ers in linking the telematic device to the cloud, so a continuous stream
of data is downloaded. The size of these ecosystems will allow a degree
of hyperscaling not possible for an individual firm.
An important competitive aspect of the future, therefore, will be that
insurers will have to be a part of an interconnected digital ecosystem
of related companies. An example would be a car platform which links
automobile manufacturers to software providers, to telecommunications
providers, to breakdown rescue services, to vehicle service providers, to
insurers, etc., so customers can be offered an inclusive, yet customized
product for autonomous cars. Google’s current push to use the Open
Automotive Alliance to establish Android as the automotive base system
is an initial example of this. Working within an ecosystem is now stand-
ard for manufacturers. Car manufacturers’ proportion of the parts they
make has dropped from about 80% in the 1980s to about 35% now.
Insurers need to undertake a similar outsourcing/insourcing analysis.
Figure 5.1 shows possible insurance ecosystems.
The role of an insurer is thus to source and work with the leading
platform innovators. The role of the platform supplier is to enable har-
monious integration, to provide a high level of low-cost customer ser-
vice, and to provide a wide range of real-time data to partners to enable
continuous product and platform improvement. There have to be com-
mon data protocols and common application programing interfaces.
The key advantages of an ecosystem for insurers are that customer
service can be provided by a firm with higher levels of customer rela-
tionship skills and social capital than insurers have traditionally dem-
onstrated, as well as data feedback from a wider range of customer
engagement areas than an insurance-only product would provide.
In terms of telematics, cooperation with related product suppliers
has obvious cost advantages as telematics can either be jointly installed
or have a dual purpose. Currently, it is hard for insurers to justify the
160    
M. Naylor

(a) Car Ecosystem (b) Health Ecosystem

Telecoms Hospitals
IT/
Data Data
Telemetrics Telemetrics

Smart Smart
Devices Devices
Platform Platform
Insurance Insurance

Govt Road Specialists

Networks
Social Network Social Network
Finance HMOs

Ratings Ratings
Reviews Reviews
Accidents Medical Emergency
Car Sales equipment Services
Maintance

Fig. 5.1  Possible insurance ecosystems

cost of installing telematics in terms of reducing claim cost by itself.

However, within an ecosystem, where data used by all parties, the costs
per ecosystem member reduces to economical levels, once the profits
from cross-selling by all members are factored in.
Ecosystem-wide integration requires that IT design is no longer
about building or buying software systems on premises and ensuring the
system links to a few external partners. It will be about understanding
and integrating the end-to-end customer experience across the whole
ecosystem regardless of the firm actually involved. Thus, an individual
firm’s IT system has to be viewed as just a specialized part of a greater
whole, based around secure but open IT protocols.
The entire ecosystem has to seem seamless and integrated from a
customer viewpoint, so they can move rapidly from product search to
reviews, to sales, to media, to financing, to insurance, to maintenance,
to claims, all on the same screen with no discernible difference between
companies. The platform provider has to ensure this customer interface
is high quality and user-friendly. The platform has to run 24/7 globally
for millions of customers.
These systems will have to be designed from the start to be mutable,
able to be adjusted as new innovations arise. The greatest error insurers
can make to allow themselves to be integrated into an increasingly ad
hoc spaghetti IT system. Strategic decisions will thus have to take prec-
edence over strictly technical decisions.
 5  What Insurance Companies Need to Do    
161

This data integration is a challenge to insurers, as they have tradition-

ally operated as closed systems with firewalls and encryption for reasons
of security. Data has been used only internally and not shared with cus-
tomers, nor have customers or other stakeholders been able to self-input
or edit data. Creating the ability to integrate with the data systems of
ecosystem partners, and myriad external data sources, while maintaining
data security, will be a substantial challenge. This is too important to be
left to CIOs.
The disadvantages of digital ecosystems to insurers who hesitate are
obvious as others will rush to fill any obvious product gaps, and insurers
who do not find a secure place in the leading digital ecosystems will find
themselves increasingly disconnected from customers and limited to low
profit-margin roles which others find unattractive. They are also likely
to be left with higher-risk customers as the high-quality data which
embedded digital ecosystem insurers obtain allows them to offer differ-
entiated premiums to lower-risk customers, as well as substantially low-
ering cost via software-based customer contact and underwriting. There
are substantial economies of both scale and scope with IT and big data
so that successful first movers may develop unstoppable momentum.
The initial leading ecosystems may grow to dominate.
The experience of Ant Financial is instructive. It is an affiliate of
Alibaba, China’s largest Web, e-commerce, and mobile device company.
Ant Financial leveraged off Alibaba’s huge client base to establish first a
mobile payments system, then credit, wealth management, and inter-
national financial services, before introducing insurance. Ant started by
offering merchandise-return insurance and has moved onto to health
policies, and peer-to-peer lending as well as structured finance. These
services are integrated and software based, with very low per-product
costs. The combined client data created by Ant’s wide product range
and Alibaba’s range of e-commerce, entertainment, healthcare services,
etc., also allows Ant to leverage off an unprecedented width and depth
of knowledge about clients to offer customized products and premiums
in an inter-active manner. It needs to be noted that Asian internet plat-
forms are moving into a range of services, especially financial, in a way
in which western platforms are not. Therefore, the main platform dis-
ruptors are likely to originate in Asia.
162    
M. Naylor

It is an open strategic question as to whether insurers should establish

their own ecosystems. It is unlikely that insurers will have sufficient cli-
ent trust or software integration expertise to either front the platform
or build it. However, insurers will have a competitive advantage in areas
like selecting clients with the high profit margins, segmenting clients
into group based on data-derived characteristics, supplying capital, and
integrating across multiple segments. The kind of client profiling and
segmentation offered by firms like Bought-by-Many is an example of a
profitable direction for insurers. The main hindrance for insurers is that
their legacy systems and instincts work against a model of deep-data
personalized prices and premiums. Insurers will also face the problems
that data across an ecosystem may be specified in quite different ways
and will require substantial work to integrate.
Most current platform providers are not inherently interested in pro-
viding insurance; they just take the opportunistic view that it is one of
many possible products they could provide via their platforms. If they
can provide it cheaper and easier than existing insurers can, then they
will invest. However, they will be happy to integrate existing insurers
into their platform if those insurers show themselves able to live within
the new digital world.
Note that the platform supplier cannot unduly exploit their power
position, or the ecosystem decays, leading to a Yahoo style collapse.
The nurturing and maintenance of partners are vital so system health is
maximized.
Insurers will have to very clearly define the advantages and skill
sets which they can bring to an ecosystem so they can sell themselves
as the preferred partner to the leading platforms. Among possible skill
sets would be experience with pricing product from intensive client
data, existing client relationships, and brand name. Platform suppli-
ers will assess applicant insurers on their ability to supply customized
product at a low price and the ability to establish and maintain a very
high level of client trust and engagement. Platform suppliers, who will
probably originate in IT-based sectors, will have a very low level of tol-
erance for hindrances caused by legacy systems or instances of poor cli-
ent service or inflexibility in coping with rapid change. They will expect
 5  What Insurance Companies Need to Do    
163

an insurance partner to already have a state-of-the-art IT system which

can easily be integrated into the ecosystem platform so the insurer can
actively use the new flow of data and can immediately provide an excit-
ing, dynamic, customer-friendly product which matches the high expec-
tations of the platform’s digital native customers.
Insurance pricing and approvals will have to occur in real time, less
than one minute, based on data sourced from the entire ecosystem. This
is easier to achieve than insurers currently assume because the premium
does not have to be static, set at the start based on the insurer obtain-
ing all possible information. Instead, because of constant telematic data
feedback, premiums can be dynamic, with an indicative premium ini-
tially offered, which is then adjusted later based on revealed client char-
acteristics from feedback.
To be able to achieve this, insurer IT systems need to operate at dif-
ferent levels; they must be intuitively accessible to customers, they
must operate in real time, core systems must evolve and improve while
rapidly absorbing innovations, new ideas must be rapidly trialed and
implemented or rejected, costs in all areas reduced, while core business
and cash flow is not unduly impacted. This will require both heavy IT
investments and substantial cultural changes in insurance management
systems.
It needs to be noted that many of the new online insurers have been
established with the long-term aim of being sold to platforms, rather
than expanding as stand-alone insurers. It may be hard for existing
pre-digital (dinosaur) insurers to convince platforms owners that they
would make a better partner in the new world than the savvy digital
insurers.
It is vital to note that within a wave theory of innovations, successful
embedding into an ecosystem can only be a short-term solution. When
next wave of innovation occurs, customers within the ecosystem will
probably not be aware of or prize the attributes of the new innovation.
The failure of both incumbents and their best customers to respond
forces innovators to seek new customers which often results in the
majority of companies within the incumbent’s ecosystem failing. The
issue for an incumbent is that it is not just a matter of one incumbent
164    
M. Naylor

switching, as they will either have to convince the new innovation eco-
system to slot them in, or convince the existing ecosystem to switch.
Failing that, an insurer needs to attempt to jump ecosystems, to a plat-
form which offers newly superior technology.
A key attribute of insurance survivors will thus be managing this net-
work of partners, ensuring that the entire network out-performs and the
insurer continues to provide the product, data, and process attributes
which makes it an attractive network partner. The inter-connectedness
of the ecosystem means that the insurer will have to assess the perfor-
mance of partners to ensure they provide leading-edge attributes and
take action if the failure of these partners starts to create issues.
Data within an ecosystem is an issue. IP, liability, privacy, profit shar-
ing, regulatory/compliance all are more complex with a seamless eco-
system, when the customer may not be aware of which company they
are interacting with. Software licensing & data ownership and CRM
issues arise. The reputation of each member may depend on the CRM
of other partners. The ecosystem needs an agreed set of standards, skills
certification, and vender performance management. Flexibility and fast
reaction need the ability to trial beta systems and accept product/system
failure as part of that. Agreed compensation protocols will be required.

Social Capital
Insurance purchase can be an emotional experience for customers and
can tie in with major life events. Insurance is an intangible product
and often complex for consumers. Thus, consumers have a tendency
to either delay purchase or rely on advisors or brand trust, rather than
do their own research and evaluation. Effective use of social media to
identify these trigger events and induce customer purchase can thus be
more useful in insurance than other products, yet insurers are laggards
in social media use. As previously explained, this leads to the idea that
‘insurance has to be sold’ as consumers, even though they recognize the
importance of the products, feel disengaged from the process. Insurance
is also different from most products as nothing of value is given at the
time of purchase, only a promise of good behavior at a later date.
 5  What Insurance Companies Need to Do    
165

Thus, a key to future insurer success is to recognize the importance

for insurers of customer trust and thus social capital. This trust is not
created by advertising but by actions and presentation style. Nahai
(2013) shows how customers judge their level of trust in Web sites
based on a range of specific subconscious indicators. Insurance Web
sites do poorly on nearly all of these indicators. Very few insurers have,
for example, used crowd-sourcing to generate ideas for policy features,
an idea common in technology firms. Danske Bank in Denmark did
try this with new product ideas and gained substantial feedback, which
improved the product and its mobile phone interface. Insurers will have
to invest heavily to keep up-to-date with the latest trends in Web and
customer relationships software, while offering a client-focused inter-
face which is sophisticated enough to appeal to the IT savvy and easy
to use so it appeals to the less IT savvy. Morgan Stanley/BCG (2014)
found that 55% of customers would switch based on a better online
experience.
Floreddu et al. (2014) point out the growing importance of social
media in determining corporate reputation and the increase in avenues
available to aggrieved customers. They use the example of Dave Carroll
who after 12 months unsuccessfully requesting that United Airlines
replace a guitar they broke wrote a song (‘United breaks Guitar’) which
he posted on YouTube. This received over 1 million views in the first
4 days and led to United’s share price dropping by almost 10%. United
then replaced his guitar and have since taken great care to handle cus-
tomer complaints with speed. When the next broken guitar complaint
occurred, they flew the complainant to China to consult with the maker -
an exercise which cost about $50,000, but gained them invaluable social
media credit.
Halliburton and Poenaru (2010) reported that only 13% of cus-
tomers trust advertising, whereas 90% trusted recommendations from
friends and 70% trusted online customer reviews. They also found that
key internet commentators were vital. Unless insurers are proactively
engaging with customers and key stakeholders they don’t have a future.
People don’t now place much importance on advertisements or state-
ments of a company, as they do on what their family, friends, and social
networks say. The main origination source for customers in the future
166    
M. Naylor

will probably be social networks. Morgan Stanley/BCG (2014) shows

that 68% of under 50-year-olds currently source a substantial propor-
tion of the information they require for product decisions from social
networks, and this is steadily rising, and 54% will use social media peer
recommendations when buying insurance in the immediate future. Of
particular interest is that 72% have, or will, researched social media
comments on customers’ claims experiences before they decide on an
insurer. This is strongest in the rapidly growing markets of Asia. Thus,
establishing high social capital is the key to future success for insurers.
Existing insurers, however, despite having established brand names,
tend to score very low in surveys of social capital. Internet retailers,
social Web networks, mobile phone operators, and even banks consist-
ently rank higher than insurers in social capital surveys. This gives new
entrants from the IT sector, firms who have high social capital, a dis-
tinct competitive advantage. Unless insurers transform this area, they
are extremely vulnerable to disruptive market entry. Companies which
act like AllState did in the 2007 Pincheira case will find that their level
of social trust plummets at an unsustainable rate and given the grow-
ing need for positive feedback, will find their client base disappearing.
Reputation is not what you say but what you do day to day - ‘do you
meet or exceed customer expectations?’
The basic element of social capital is trust, and to establish this, insur-
ers need to be seen as ‘authentic’ on an intuitive level by potential cus-
tomers. Digital natives will be more skeptical of the insurance promise
than prior generations and more prone to flight at any hint of non-
authenticness. Insurers need to generate brands based on more than
intensive advertising in disappearing media formats. Authenticness is a
fuzzy concept, and potential customers will use a range of quite subtle
clues to assess it. Outbound sales and inbound claims need to be inte-
grated, as Perdue University (2014) showed that the swift resolution of
a complaint is often the most effective time for additional sales.
A key issue for insurers here is the degree of ‘client stickiness,’ which
is the extent to which clients stick with existing suppliers and do not
periodically shop around. Traditionally, personal insurance has had a
high degree of stickiness, while general insurance had been lower, but
 5  What Insurance Companies Need to Do    
167

still moderate. Insurers have thus not had to work hard to retain exist-
ing customers and have focused their advertising on attracting new
customers. One of the key customer benefits of digital platforms is the
extent to which it drops the costs to customers of both product compar-
ison and supplier switching, in term of both cost and time. Therefore,
products which are presented on a digital platform have sharply reduced
levels of stickiness. In many markets, insurers have tried to slow this
process by refusing to cooperate with aggregator Web sites. However,
the attractiveness of these sites to the newly digital-savvy public means
that this strategy only delays the required changes and makes the mar-
ket more attractive for an outside disruptive entrant. Refusing to engage
with platforms is a dead end. The key question should be; Does a qual-
ity ecosystem platform exist?
To be successful in a world of social capital, insurers will have to col-
lect a qualitatively wider and deeper range of data so that they can gain
true insights into the lives of clients. Insurers do have an advantage over
possible entrants as customers are used to providing a wider range of
personal data than for other producers. In the world of real-time data,
however, the type of data required is qualitatively more personal, and
therefore, deep trust is required in the holder of that data. There is clear
potential for this data to be misused, and even a single example of mis-
use could devastate an insurer’s social capital.
An issue with current insurers is that they collect client data for rea-
sons which do not relate to the products provided. An example would
be requiring a client’s full name, phone number, and street address
before a sample quote can be provided, without justifying why the
street address or phone number affects the quote. Insurers also tend to
ask standardized information from each customer, even if they already
know the information, or if the information is irrelevant for that cli-
ent. In a digital era, insurers will need to provide up-front a clear reason
for each item of information, and why it is in the customer’s interest to
provide it. Any hint of information being collected for the company’s
own benefit, rather than the client’s, will destroy social capital. Getting
clients to spend time providing unnecessary information also sends the
message that the insurer is disrespecting the potential customer’s time.
168    
M. Naylor

Companies have to view social media as a positive way to communi-

cate with customers, not as a threat. However, speed is vital - it used to
be said that the first 24 hours of an adverse story were vital, but now it’s
the first 24 minutes.
Insurers will have to accept that digital native clients will review
and rate their services and Web platforms online in detail and insurers
which do not offer a very sophisticated, efficient, and friendly service,
will be rated down, and will struggle to attract new clients. Firms will
rise or fall in ratings fast, so that long-established businesses can disap-
pear in a few years. Firms which try to discourage ratings will effectively
disappear from the view of digital natives, whereas firms who encourage
ratings and respond to feedback will find demand strongly increasing.
Using social networks to advertise to related groups will multiple this
tendency. McKinsey (2012) argues that use of social network links can,
through client acquisition and fraud detection, generate insurer value of
up to 26% of the cost base in call centers.
It is vital for insurers to realize that customers are likely to be already
creating social media network discussions around the insurer’s reputa-
tion and customer service, based on existing social media sites. If an
insurer is not aware of this, then the social capital of the insurer may
already be negative. Insurers cannot hide their head and ignore it.
Neither can insurers expect to engage with social media once their repu-
tation or client trust is lost, as commentary will simply move to closed
sites. Proactive engagement is necessary from the start so enough trust is
retained to allow aggrieved customers to engage with the insurer.
Customer engagement will thus become a key insurer skill. As dis-
cussed before, this partly means creating frequent communications
with customers which are not just related to annual renewal notices or
claims. It is vital, however, that this communication is useful to the cus-
tomer and welcomed by them. This is easier in some areas than other,
but once telematics are widely used, it is possible in all areas. Health
and life insurers can offer health news as well as feedback on customer
efforts. Car insurers can offer dynamic premiums based on current driv-
ing, or traffic info. For example, insurers could link car telematics to
the central panel LCD and change premiums minute by minute based
 5  What Insurance Companies Need to Do    
169

on driving, thereby encouraging best-practice driving. House insurers

can link with household telematics, appliance suppliers, and interior
designers to offer household news about appliance innovations, grocery
specials, or design ideas. Insurers, however, do not have a wide enough
range of skills to engage with customers by themselves. The key com-
ponent is thus the selection of market-leading media partners and the
creation of a dynamic business ecosystem, where an attractive package
can be created.
Insurers will have to actively monitor their social capital and have
proactive programs to increase it. Insurers will have to offer a 24/7 ser-
vice across a range of social platforms and modes. They will have to
actively engage with clients and promote publicly available customer
feedback. This openness goes against the grain of current insurance
companies, who tend to conduct market research in secret and not
make adverse feedback public. This is no longer viable, as it signals fear
and generates mistrust.
Communication also has to be two way, instead of the traditional
insurer response to a crisis of issuing bland reassurances and not actively
listening to grievances. In a digital age, successful companies will have
to positively invite public feedback, negative as well as positive, as they
will be judged by potential clients on their response to that feedback.
Digital natives are impressed by companies who actively engage in pub-
lic debate and provide reasonable justifications for decisions. In a digi-
tal age where firms have lost control of market research data, adverse
feedback can no longer be hidden, but has to be rapidly handled.
Customers understand that not all feedback will be positive and will
judge based on overall trends. Insurers who are successful in terms of
Web site usability, security, cheapness, Web site and response speed, cus-
tomer support and feedback implementation will give customers posi-
tive reasons to move from their current insurer. Any insurer, however,
who is consistently rated low can expect to lose any new customers.
It will be particularly important that insurers are seen to respond rap-
idly and positively to claims as reviews around claims experiences are
the key factor in establishing favorable overall social capital. One key
to this is to ensure that customer engagement occurs over the full range
170    
M. Naylor

of insurer activities and not just around claims. Insurers may have to
strategically manage flexibility around claims payouts, as a tight claims
policy will immediately be spread via social media and will rapidly be
reflected in shrinking new business. Given that it is easier to lose a
reputation than gain one, a short-term tightening reaction to a spike
in claims has the potential to adversely and severely impact on market
share.
Insurers will thus have to understand that in a digital world claims
refusal will be substantially different, with each claims refusal which is
based on a ‘difficult-to-understand-interpretation-of-abstract-terms’
likely to be widely discussed and reviewed on social media, with mul-
tiple customers able to comment on each claim refusal. Even one clear
example of what is seen as ‘unreasonable behavior’ will impact very neg-
atively on insurer reputation, as the speed and breadth of adverse social
commentary multiples reputation losses. A repeated pattern of ‘unrea-
sonable’ claim refusal will have the potential to destroy an insurer’s
social capital. Given the dynamic nature of future policies, insurers with
a bad social media rating may lose customers and cash flow very fast.
Because bad press from poor claims responses will impact on the rep-
utation of the wider ecosystem, affected insurers may lose their place in
prized ecosystems as other parties will fear contagion of their reputa-
tions. While mistakes can be forgiven by social media discussions, any
systematic adverse pattern of behavior will be quickly and widely dis-
cussed. Insurers will have to understand that the creation and retention
of client trust are the basis of future success.
Aggregator sites will evolve which allow customers to mix and match
insurable activities. For example, one insurer may offer cheaper rates
when the car is in the garage, another offers the best rate during com-
muting, and a third offers the best rate for the summer vacation drive.
Companies may decide to offer ‘specials’ for Friday’s commute to those
who switch for that morning from competitors. Note that firms which
create effective aggregation software will sit at the center of the eco-
system and thus collect customer data, will hold immense power over
insurers, who will be relegated to mere product suppliers, always at risk
to disintermediators who specialize in one insurance activity.
 5  What Insurance Companies Need to Do    
171

Insurance Policies
Most current insurance policies are technically complex, with the exact
definition of terms having substantial weighting in terms of costs.
Therefore, it can be vital that customers take time to read and under-
stand policy detail, and possibly seek expert advice. This is necessary in
a world where insurance is a permanent and serious purchase, where
policy terms and premiums are static.
However, in the future world of digital real-time polices, the over-
whelming bulk of customers will have neither the time nor the incli-
nation to carefully read and understand small print - very few online
customers read the ‘terms and conditions’ in a mobile phone tick box.
Heavily reliance on policy terminology is also unrealistic from a cus-
tomer viewpoint if polices terms and premiums are dynamic, ever-
changing based on telematic feedback.
Customization of insurance to the needs of individuals is also vital.
As non-Americans understand, Starbucks coffee is low quality and high
cost. The reason why Starbucks is so popular is that customers can
specify a range of choices, so they can get exactly what they want - not
‘Starbucks coffee’ but ‘John’s Coffee.’ In this kind of customer expec-
tation world, insurers cannot continue to offer ‘life insurance,’ where
the terms and conditions are standard across multiple customers. The
future has to be an individualized policy, whereby customers can amend
terms and conditions and thus premiums, until they get ‘John’s Life
Insurance.’ Yet, it is unrealistic from an insurers’ cost viewpoint if poli-
cies are customized to individual clients.
The only solution in the more dynamic future is for policies to be
offered as modules, with terms vastly simplified and standardized.
There, however, also needs to be some flexibility within conditions or
added options, so customers can go on a Web site and play around with
adding or subtracting, and watching how the premiums change, until
they are satisfied with their customized version.
The meaning or terms also have to be very clear. It is vital that what
the average customer expects a term to mean is what the term is inter-
preted by the claims side to mean. It is vital that what the insurer seems
to an average person to be offering is what they are actually offering,
172    
M. Naylor

not what the underwriter thought the term meant. Polices will have to
be explained in clear and simple terms and be able to stand up to a sys-
tematic public comparison with those of competitors. Insurers who do
not distribute their policies online and allow them to be compared and
reviewed in detail on social media will disappear from customer view.
Digital natives will not accept insurer’s defining policy terms based
on legal rulings or long-established industry practice if these seem
unreasonable. Thus, insurers will have to establish policy terms based
on what seems reasonable to a wide range of the public. Premiums will
have to be costed on the basis of a high level of customer goodwill and
an understanding that long-term reputation and social capital are sub-
stantially more important than short-term gain by a strict claims policy.
Since claims clarity and goodwill versus competitors will be reviewed
and compared online, insurers will have to ensure their policies and
claims responses are at a level they can be proud of.
Premiums will need to be individualized and policies modularized
and made dynamic. An additional factor is that because feedback from
telematics will give customers a good idea of their risk level from each
activity, customers will no longer be prepared to be lumped into a pool
of higher-risk individuals. This feedback will be dynamic, and some cus-
tomers may find this exciting.
Policy diversity across insurers thus needs to be reduced and plain
English terminology adopted. This has to occur alongside increased cus-
tomization, flexibility, and real-time information flow. For example, a
major problem with some Web-based policy application systems is that
customers often don’t understand the choices or terminology within an
excessively lengthy questionnaire, so use advisers to guide them through
the process, defeating the purpose.
This policy customization is not possible if humans are involved in
setting up or administrating policies, as the marginal cost is too high.
The advantage of a software-intensive underwriting and administrative
insurer is that the marginal cost of allowing individualization drops to
a viable level. There is a strong first-mover advantage to this software-
intensive system as human-based competitors will find themselves una-
ble to compete and faced with falling revenues alongside rising costs of
software investments.
 5  What Insurance Companies Need to Do    
173

References
Accenture. (2013). Running Business for Growth. Insurance-blogs.
Adner, R. (2002). When are technologies disruptive? A demand-based view of
the emergence of competition. Strategic Management Journal, 23(8), 667–688.
Bain & Company (2013). For Insurance Companies, the day of digital reckoning.
Bain & Company. (2015). Global digital insurance benchmarking report 2015:
Pathways to success in a digital world.
Boston Consulting Group. (2016). How digital switchers are disrupting US auto
insurers, July 26.
Breznik, L. (2012). Can information technology be a source of competitive
advantage? Economic and Business Review, 14(3), 251–269.
Capgemini/Efma. (2016). World Insurance report.
Christensen, C. M. (1997 & 2011). The Innovator’s Dilemma (4th ed.). Harper
Business/Harvard Business School: USA.
EY. (2013). Insurance in a digital world: The time is now. Ernst-Young: New York.
EY. (2014). The future of underwriting: A transformation driven by talent and
technology. Ernst-Young: New York.
Floreddu, P. B., & Cabiddu, F. (2014). Multi-channel strategy and sustain-
able competitive advantage: A dynamic capacity approach (Working Paper).
University di Cagliari.
Floreddu, P. B., Cabiddu, F., & Evaristo, R. (2014). Inside your social media ring:
How to optimize online corporate reputation. Business Horizons, 57, 737–745.
Halliburton, C., & Poenaru, A. (2010). The role of trust in consumer relation-
ships. ESCP Europe Business School.
Iansiti, M., & Levien, R. (2004). The keystone advantage: What the new dynam-
ics of business ecosystems mean for strategy, innovation, and stability. Boston,
USA: Harvard Business School Press.
IBM. (2006). Insurance 2020: Innovating beyond old models. USA: IBM
Institute for Business Value.
Johne, A., & Davies, R. (2000). Innovation in medium-sized insurance
companies: How marketing adds value. International Journal of Bank
Management, 18(1), 6–14.
KPMG. (2014). An industry transforming: Insurers and intermediaries in a fast
changing, digital world. White paper.
KPMG. (2015a). Automobile insurance in the era of autonomous vehicles: Survey
results, June.
KPMG. (2015b). Big data: The big factor driving competitive advantage.
Frontiers in Finance.
174    
M. Naylor

KPMG. (2016). Empowered for the future: Insurance reinvented. KPMG International.
Markides, C. (2006). Disruptive innovation: In need of better theory. The
Journal of Product Innovation Management, 23, 1.
Martinez, A. G. (2016). Chaos monkeys: Inside the Silicon Valley money machine.
London, UK: Ebury Press.
McGill, R. (2008). Technology management in financial services. London:
Palgrave-Mcmillan.
McKinsey. (2012). The social economy: Unlocking value and productivity through
social technologies: Consumer financial services. McKinsey Global Institute.
McKinsey. (2016a). Transforming life insurance with design thinking, May.
McKinsey. (2016b). A digital crack in banking’s business model, April.
McKinsey. (2016c). How companies are using big data and analytics, April.
McKinsey. (2017). A future that works: Automation, employment, and productiv-
ity, January.
Morgan-Stanley/BCG. (2014). Insurance and technology: Evolution and revolu-
tion in a digital world. USA: Boston Consulting Group.
Morgan-Stanley/BCG. (2015). Insurance and technology: The emerging role of
ecosystems in insurance. USA: Boston Consulting Group.
MyCustomer/Microsoft. (2015). Customer experience in the insurance sector:
Insight Report. UK.
Nahai, N. (2013). Webs of influence: The psychology of online persuasion. FT
Press.
Neirotti, P., & Paolucci, E. (2007). Assessing the strategic value of informa-
tion technology: An analysis on the insurance sector. Information and
Management, 44, 568–582.
Peters, T. J., & Waterman, R. H. (1982). In search of excellence. New York,
USA: Harper & Row.
Progress Corticon. (2014). Flexibility and control in insurance operations. White
Paper. www.progess.com.
PWC. (2014). Insurance 2020: The digital prize—taking customer connection to
a new level.
PWC. (2015a). Insurance 2020 & beyond: Necessity is the mother of reinvention.
PWC. (2015b). Insurance banana skins.
Siegal, E. (2016). Predictive analytics: The power to predict who will click, buy,
lie, or die. USA: Wiley.
SwissRe. (2013). Automated underwriting: Threat or opportunity? New York.
Teece, D. J., Pisano, G. G., & Shuen, A. (1997). Dynamic capacities and stra-
tegic management. Strategic Management Journal, 18(3), 509–533.
Verint. (2016). Customer-Centricity in financial services: The rules of engagement.
White Paper, USA.
 6
Impact by Type of Insurance

Autonomous Vehicles
The majority of the insurance industry managers are under the impres-
sion that rise of autonomous driving will not impact on car insurance
until 2030 or later. This is not true, however, as fully (100%) automated
cars are not required before technological advances severely disrupt the
car insurance market. Semi-autonomous car technology will have a
substantial impact far sooner. One impact is that premiums will drop
substantially as there is mounting evidence that even semi-autonomous
drive cars are safer than manual-only cars.1 As of August 2016, all the
crashes advanced semi-autonomous cars have been involved in have
been caused by human drivers either running into an auto-drive, or in
one case, a human deciding to put the Google car in manual mode.2

1US National Highway Traffic Safety Administration data show that as of 2016, semi-autono-
mous cars had a 60% lower crash rate than fully manual cars.
2Note that the Aug 2016 crash of a Tesla car did not involve an autonomous car, but a car with

advanced cruise control capacity, which required constant human supervision.

© The Author(s) 2017 175

M. Naylor, Insurance Transformed, Palgrave Studies in Financial Services Technology,
DOI 10.1007/978-3-319-63835-5_6
176    
M. Naylor

Since advances in vehicle automation depend on perfection in a

range of physical technologies, like sensors, rather than just software,
overall advance may occur slower than software-only technology. The
pace of autonomous technology is, however, faster than many insurers
recognize.
One reason is that while media has been focused on fully autono-
mous Google cars, a range of other manufacturers have made sub-
stantial progress in semi-autonomous cars. Note that these are 90%
autonomous cars but have the advantage over Google cars, as there is no
need to intensively pre-map roads like Google does. In April 2015, an
Audi with Delphi technology drove from San Francisco to New York,
3400 miles of road it had never driven on before. Its human driver only
took control on 1% of the densest city roads. 2016 and 2017 saw the
start of extensive trials of cars which will incorporate 80% of the tech-
nologies required to be fully autonomous, including highway adop-
tive auto-cruise (speeding up and slowing down with traffic, passing
other vehicles when necessary, and alerting the driver when their exit
is near so they can take over), autonomous braking, the ability to han-
dle stop-start traffic jam driving, and self-parking. While a driver will
be required to be at the wheel, the car may be in auto mode 80% of the
time. Most elements of this range of technology will probably be intro-
duced into the 2018 generation top-of-the-range Mercedes, BMW, and
Audi cars.
Car manufacturers are deliberately introducing these innovations in
steps, to allow customers to get used to them. There will thus not be a
big moment when cars go from 100% human control to 100% auto-
mated; instead, cars will go from 40% to 60% to 80% to 100% over
the next decade or so. This self-drive ability will be extended to subur-
ban streets by 2020. Fully autonomous cars will follow with some com-
mentators arguing this will involve up to 50% of all cars sold by 2025.
Similar features are being installed in trucks. It is very likely that black
boxes to record crash details will be required in all cars. The EU made
basic black boxes compulsory on all new cars from 2016.
One aspect of these cars which commentators have so far not
explored is their ability to be in constant Wi-fi communication with
other cars and with a centralized data center. SwissRe (2016) argues that
 6  Impact by Type of Insurance    
177

by 2020, more than 2/3rds of cars sold will have some form of connec-
tivity, or 260 million extra cars a year. There are two kinds of connectiv-
ity: (i) tethered, which involves a 3rd part device such as a mobile phone
and (ii) embedded, which transmits from a device which is part of the
car. Most current networking works via tethered devices, but since most
new cars will feature screens with Internet, embedded networks will
soon become the norm. All Tesla cars have embedded networks. A sub-
stantial proportion of commercial trucking fleets is already networked as
operators find that data on vehicle location, on driving time, etc., offer
significant scheduling advantages. Therefore, most of the innovations in
this area will probably be trialed on trucking fleets.
Since embedded connections have two-way capacity, there will soon
be a vast stream of driving and road data available to all those compa-
nies within the car ecosystem. This will allow insurers to analyze risk in
far great detail, but will also accelerate advanced driver assistance sys-
tems and autonomous driving software. Because of this data, by 2020,
new cars should be able to handle autonomous driving in most express-
way and stop-start driving environments, as well as self-parking.
The impact of networking will be substantial. Currently, an autono-
mous car is maneuvering along a road by itself forced to recognize other
cars and road dimensions and conditions with very limited external
feedback. The current situation of one autonomous car traveling down
a dumb highway is thus like a blind man moving by trying to find the
footpath edges and bumps by echolocation. It is extremely difficult for
even a smart car to tell a rumpled newspaper from a rock, to identify
a faded pedestrian crossing, or to spot a temporary roadworks sign.
Battered road signs can be hard to distinguish.
A major hindrance in auto-drive development was that prior system,
like the Google system, requires cars to retain huge data-bases of road
layouts and finds it hard to deal with roads which have not been inten-
sively pre-mapped. Gomes (2014) thus argued that these mapping and
computer requirements meant that nationwide autonomous cars may
never be possible. Google currently needs to employ people to spot road
signs, etc., from its road videos. This is impossible for a large country.
The real breakthrough with autonomous cars will therefore occur
when they are networked so that they can communicate with each
178    
M. Naylor

other, with road side objects, and with a central IT database. This is
because networking removes existing stand-alone limitations, as real-
time feedback from sensors will mean that even when automated cars
are manually driven, they will map roads and add to the collective data
base. With each car-mile driven, more data is added, so roads get exten-
sively mapped. And, as long as one autonomous car drives along an
unmapped road, then all subsequent cars will be able to access a data
base to receive details on that road. Tesla cars are networked, so Tesla
continually collects road data from every car as it drives along, so their
road data is constantly growing and takes note of changes in road condi-
tions. Once a substantial percentage of cars have network capacities, the
cars can start to talk to each other. If we add to this ‘smart roads’ which
have embedded telematics in street lights, traffic lights, pedestrian cross-
ing, in the lane-maker bumps, or metal incorporated into road paint,
then cars will receive continual feedback on what it needs to know. The
first car through a roadworks diversion can explain to the rest of the
cars how to handle it. Road workers can put down Wi-fi-chipped cones
which will map out for cars exactly how to drive around the road works.
Rural roads or off-grid tracks can be added to the data base.
There is the possibility of issues with initial navigation systems if
roads are predominantly dumb. It has become not uncommon for
humans using software-based navigation systems to blindly follow the
navigation advice even if it is obviously faulty. The issues could be worse
with a car, which may be less aware of surroundings and no way of res-
cuing itself if it gets lost. However, once smart roads are created and cars
networked, then real-time feedback from other cars, as well as a database
of all cars which have made the journey previously, should avoid these
issues, except in very rare cases, for which protocol can be established.
Thus, the real breakthrough with autonomous cars will not be the
ability of cars to sense their surroundings, staggering as this is in engi-
neering terms, but their ability to communicate with each other, and
with telematics, to collectively gather information on roads and the
movements of other vehicles. The future situation will be like the man
gaining sight, as telematics will light up the road way, roadside objects,
traffic lights, and the position of all other cars.
Networked cars can continually pass on info to each other in real
time on car position, road conditions, exchanging data on current
 6  Impact by Type of Insurance    
179

position and speed, where they are going, mutually arrange directions
so they can travel faster and closer, and receive information from cars
miles ahead as to road or traffic conditions or roadworks, over as many
future miles as is required so that each car has a far better idea of the
road and other cars than any human driver. Cars can then even coordi-
nate with each other over traffic information to enable the group of cars
to mutually choose traffic routes which minimize total traffic. The cars
will talk to traffic lights and traffic management systems, sending flocks
of cars through lights, or re-routing down less busy streets. The highway
system will be able to move cars as coordinated groups, like flocks of
birds. Traffic police will be able to issue general instructions to all cars
in an area, if a diversion is needed, and to stop cars they have suspicions
about. The collective wisdom of cars will vastly exceed what each indi-
vidual car could achieve.
These cars will be, thanks to radar and networked roadside mark-
ers, able to operate just as safely at night, or in fog, or storms, as they
will not be solely dependent on visible light as human drivers are. They
will talk to parking space sensors before they arrive, will be allocated
an empty space, will drive to it autonomously, and then charged for
it. Each car will not need extensive onboard maps or large computing
power, as these can be stored in the cloud.
It is useful to note the huge size of data which new model cars pro-
duce. A car can have 1000 sensors, the processing capacity of 20 PCs,
and produce about 25 GB of data per hour. This will increase at an
exponential rate, especially once cars start to network. The complexity
of car software is rapidly exceeding the complexity of the mechanical
engineering, so that an auto-drive electric car is essentially a computer
on wheels. This complexity has left some automobile manufacturers
struggling, e.g., the components in Ford cars are supplied by a range of
contractors, each with their own software systems. This means that even
though Fords are heavily computerized, Ford is struggling to leverage
this due to difficulties in integrating the parts into an in-car network.
Over time, this huge data base of real-time information on the move-
ments of cars will allow algorithms to be created to make autonomous
cars far safer than any human driver. The system will treat cars driven by
human drivers, or non-networked cars, as dangerous and warn all other
networked cars to establish wide leeways around them. Note that fast
180    
M. Naylor

computers, visual sensors, big data, Wi-fi, telematics, cloud-computing,

and advanced software will combine in a way, whereby the sum of the
parts vastly exceeds each individual contribution. There are significant
scale and scope economies, with possibilities exponentially expanding as
the percentage of people and devices connected rise. Exponential rates
of improvement in auto-drive software imply that by 2020, human
drivers may be 10x as unsafe as 90% self-drive cars and over 100x less
safe by 2025.
It is vital to note that autonomous driving is not the only car soft-
ware innovation made possible by networking. An example is the Tesla
S model which is externally linked to the Internet. Since all parts of the
Tesla S are connected to an internal computer, this allows Tesla engi-
neers to spot a problem and solve it by uploading a software fix. Tesla
engineers can download data from all the cars, discover common issues,
and invent improvements, without leaving their office. Sensors inside
the car will continually monitor performance, for example, using an
exhaust sensor to monitor fuel burn and then adjusting engine param-
eters to improve performance. Wear rates of different parts can be
tracked and changes made to manufacturing. Software can alert owners
when parts are near to failure and book maintenance, which a self-drive
car could drive itself to.
Once a car is networked, then required software upgrades can take
place by push-downloads without the owner being aware of it. This means
that Tesla owners discover that their cars receive continuous upgrades in
capacity without them having to do anything. For owners, their experi-
ence of owning car becomes like a software product - with continual
improvement expected. This will be particularly useful once electric cars
are common, as these are easier to modify remotely. For example, Tesla has
been able to substantially increase the capacity of its car battery cells by
software updates sent remotely. Insurers need to be able to link into this
car software network, or they will be left with no market.
In another direction, Mercedes-Benz already allows their car to
inform home-software that the owner is coming home, so that heating
and other home appliances like the coffee maker or oven can be turned
on. This link could be used in other ways, for example, the fridge or
food cupboard could send a shopping list to the mobile phone, and the
 6  Impact by Type of Insurance    
181

car can then communicate with local shops to find the best specials and
deliver the driver to the optimal combination of shops.
The elderly, disabled, or children could be driven rather than have to
rely on the courtesy of others. Those who drink would have no issues
getting home. Some of the autonomous vehicles may look nothing like
today’s cars. For example, if a vehicle is delivering pizza or groceries,
why would it need seats or height? - it could be built like a smallish box.
A self-drive taxi may only need space for one or two people.

Automobile Insurance
There are many issues for insurance companies around autonomous
cars. The biggest is that once networked autonomous cars are freely
available, KPMG (2015) estimates that there will be at least an 80%
drop in crash rates. Car crashes kill more than 30,000 people in the
USA annually. Think of all the factors that affect a human driver: physi-
cal (tired), emotional (angry), psychological (confused), or intellectual
(distracted) factors all come into play when a person gets behind the
wheel. Based on current experience, the US NTSB estimates that since
93% of all car crashes are caused by human error, so it is possible that
crash rates for auto-drive cars could drop by more than 90%, even if
autonomous cars are not 100% crash-free. Semi-autonomous Tesla’s
have so far had a crash rate 5% of other cars. Volvo has an aim of elimi-
nating car crashes caused by the Volvo by 2020. Theft will disappear,
as these cars will also be nearly impossible to steal, as they will both
require voice and face recognition to start and can be programed to lock
the doors and deliver the thief to the nearest police. This leaves only
damage caused by external factors like things bumping into the car or
weather.
The result of this is that car insurance premiums will fall drastically.
All State’s 2015 annual report admitted that automated cars could
destroy their auto-insurance business. Those want to drive manually will
pay a hefty surcharge, probably 10 or 20x the automated premium, so
drivers will choose to forego self-driving if possible. Even before autono-
mous cars are used, telematics will enable insurers to identify the bad
182    
M. Naylor

driving habits of each driver and reeducate them, as well as identifying

the worst drivers, who create an abnormally high number of crashes, and
use high premiums to discourage them from manually driving. Override
controls could be used to disable cars when drivers exhibit dangerous
patterns of behavior, or enforce time restrictions on learner drivers. Car
insurance in an era of very few accidents is a terminal industry.
International Transport Forum/OECD (2015) predicts that use
of auto-taxis will cut the number of cars used in major cities by up to
80%, and that private vehicle ownership will only exist in less popu-
lated rural areas. While this decline in per unit premiums may be off-set
somewhat by a rise in vehicle value, in general, it is likely that a severe
drop in premiums will impact heavily on insurer cash flow. Activities
like supporting sales branches could become unviable.
As noted earlier, commentary has been mainly focused on 100%
autonomous cars, and since these may be a decade or more away, insur-
ers have felt no sense of urgency. However, the key factor for insurers
is the arrival of 80% autonomous cars, and these started to arrive for
high-end models in 2017. Most manufacturers have already conducted
extensive trials. Since these 80% cars will have far lower crash rates,
declines in cash flow will impact on insurers from 2017 onward at the
premium end of the market.
Even using conservative assumptions, KPMG (2015) estimates that
by 2040, car insurance premium cash flow will be no more than 40%
of the current size. This is probably too conservative as SwissRe (2016)
found that current-tech adoptive-driving systems will, by 2020, reduce
crash rates by 45%. Note that this includes both driving and parking
incidents. More advanced systems are expected to reduce crash rates
by up to 90%, with most remaining crashes involving crashes caused
by 3rd party low-tech cars. Therefore, I would argue that using reason-
able assumptions, the impact on premium cash flow will be probably
more intense and a lot faster, with premiums dropping by over 80%
by 2030. Before this, the turning point, where intense and probably
destructive completion starts, is likely to occur before 2025, conceiv-
ably from around 2022. Changes to car insurance will be enhanced
once cars are networked, and a separation can be made between insur-
ance when the car is parked and insurance when it is driven, as well
 6  Impact by Type of Insurance    
183

as the introduction of real-time dynamic insurance. Note that since

the weak link will be human drivers, car insurers can be expected to
offer substantial discounts (up to 90%) to drivers who keep the car in
autonomous mode. The dynamics of this is discussed in more detail in a
subsequent chapter.
Three probable outcomes arise from the above changes: (i) Accident
frequency will decline to where the difference among driving behaviors
becomes negligible, and it is difficult to charge a meaningful premium
for insurance. (ii) Insurance will take the form of commercial product lia-
bility instead of personal driver liability as we let software do the driv-
ing. (iii) Vehicle utilization will rise and cars on the road will decline as
one car can serve the driving needs of multiple travelers per day.
Given that crashes caused by the autonomous cars will likely be the
fault of software or the telematics, product providers may well have to
cover costs under normal product guarantee laws, or will provide very
cheap life-of-the-car insurance at purchase time. Volvo and Tesla have
already announced plans to offer buyers automatic car insurance with
purchase, with Audi and Mercedes indicating similar plans. Issues with
product-based insurance are discussed in a subsequent chapter.
It needs to be remembered that the automobile industry itself is on
the cusp of its own disruption tsunami. At the center of this is Tesla and
Elon Musk. While Tesla is currently very small, Musk is promising to
produce 1M cars a year by 2020 and predicts they’ll be larger than Ford
by 2025. By 2020, the Tesla model 3 will be the same purchase price as
a gasoline car, yet energy costs will be 5% (or less if paired with home
solar charging). Musk expects, that by 2020 or soon after, to increase
recharge distance to 620 miles/1000 km, exceeding gasoline cars, along
with Level 4 autonomy. All Tesla cars come with the hardware required,
so level 5 timing becomes an issue of how fast AI can be evolved.
Within a business ecosystem, the power will reside with the firm
which receives the continuous steam of driving data, which will be the
key to underwriting risk. Unless insurers take charge of this stream, they
will be reduced to suppliers of risk algorithms for dynamic insurance.
Samsung is investing heavily in the autonomous car parts suppliers and
can be expected to be a major platform center, sub-contacting insurers
to provide product/system-based insurance for the entire group of cars.
184    
M. Naylor

Insurers want access to a database of telematics data to help them set

personalized premiums for individual drivers, but arrangements gov-
erning how that information is gathered, managed, and accessed could
be subject to scrutiny by competition regulators. Telematics data can
constitute personal data, and therefore fall subject to data protection
laws, on the basis that it records the activities of individual drivers, or
a number of individuals. Insurers will need to be able to make sense of
this data via a model with predictive capabilities based on frequency of
driving, hard braking, sharp turns, time of the day, and a handful of
other factors to determine the personalized premium for each customer
based on risk profile. UBI programs offer many advantages to insurers,
consumers, and society. Underwriters also need to be aware of specifics
of ADAS systems and insist that repairs are qualified and sensor align-
ments are checked regularly.
Linking insurance premiums more closely to actual individual vehi-
cle or fleet performance allows insurers to more accurately price premi-
ums. This increases affordability for lower-risk drivers, many of whom
are also lower-income drivers. It also gives higher-risk drivers the ability
to control their premium costs by incentivizing them to reduce miles
driven and adopt safer driving habits. Fewer miles and safer driving
also aid in reducing accidents, congestion, and vehicle emissions, which
benefits society. Note that there are two differing concepts; pay-how-
you-drive (PHD), where rates change based on risk, and pay-when-you-
drive (PWD), where rates change based on usage. Both of these would
be used to set dynamic insurance rates.
Car insurance is thus the test area of the new IT revolution, the
‘canary in the coal mine’. The key variable is the speed of the spread
of the 80% or 90% technology. BCG (2015b) estimates that the per
unit cost of 90% technology, predominantly sensors, will initially be
about $6000 extra per car, while car premiums may drop by $400 per
year. While manufacturers may initially restrict the technology to high-
end cars as a sales incentive, the desire to spread the upfront cost over
as many units as possible guarantees its spread to middle-level cars by
2020. Costs will drop rapidly as volume builds so that all but the low-
est end cars will have the technology built-in by 2024. Young drivers
in the UK, who are charged an average of £2300 per annum premium,
 6  Impact by Type of Insurance    
185

are already switching to telematic-based policies. Note, that in contrast

to the high cost of the current method of installing black boxes in cars,
future cars will come with sensors pre-installed.
Some commentators have argued that the high cost of sensor systems
will limit their appeal. However, since networked cars will be able to
form conveys (closely packed cars moving at high speed) on express-
ways, they are likely to get priority lanes and get passengers to work far
faster. Alternatively, these auto-cars can network with each other and
work out mutually advantageous commuting routes. These substantially
faster commuting times for auto-use will be attractive to customers and
make them willing to pay a premium. Note that all cars produced since
2006 include an OBD chip which records miles driven speeds, speed
of turns, braking, and gas consumption. Linking to this is sufficient to
allow insurers to offer driving based discounts, so black box installation
is not necessary. Progressive already offers insurance based on this to its
customers.
The scope for stand-alone car insurers will rapidly shrink, with their
main future business backing product guarantees by manufacturers. At
this point, stand-alone car insurance is a niche product, as for most drivers
it will either be a software-linked product or a lifestyle choice for the few
remaining drivers who choose to manual drive despite the high premiums.
BCG (2014) argues that the major mistake insurers are currently
making in the telematics area is that of inadequate levels of investment.
A small product portfolio in this area heavily restricts insurer flexibility
as there is not much they can do without large enough scale to gain a
competitive advantage. BCG (2013) states that correct use of car tele-
matics changes key elements in the insurer value chain and is positive
for insurers as (i) it enables insurers to select and price individuals more
accurate, offering innovative solutions to low-risk groups, (ii) it enables
accurate and rapid assessment of claims, (iii) it enables insurers to pre-
dict loss incidents, and (iv) it enables insurers to differentiate their value
proposition and protect their customer base from direct offerings. It is
urgent that insurers gain experience with telematics in the auto area as
this is vital learning for insurers to cope with telematic expansion into
other areas of insurance. A very good starting area for gaining experi-
ence with big data derived from traffic telematics is integration with
186    
M. Naylor

commercial vehicle operators who are already using real-time location

and usage data. Some insurers have already created this ecosystem link.
The future of cash flow for car insurers will be from the provision of
web/telematic-based car services, rather than from risk-based premiums.
It is useful to understand that within the wider car ecosystem, there
will be growing sources of cash flow, though these will tend to arise
from offering dynamic services, rather than a static risk-based insur-
ance product. Insurance survivors will those companies who can transi-
tion toward an engaged customer dynamic relationship, whereby they
became an active part in customers’ lifes via data-based services. The
problem for insurers is that, in general, they do not currently have ade-
quate skills or social capital to be able to compete against IT firms who
already have active engagement with customers.
Imagine that you step into a self-driving car of the future. The car’s
heads-up monitor issues an alert: ‘I sense that you’ve been drinking. I
probably don’t need your driving assistance, but there’s always a slim
chance I may, in the event of an unexpected emergency. If you want to
operate this vehicle under the influence of alcohol, your insurance rate for
the next 12 hours will increase to $2 per hour. Press here to accept this.’
Insurers will have to switch away from linking their insurance to one
product to a general client activity. Thus, instead of insuring a client’s
car, they insure the client whenever they use a share car, even if that car
is always different. This switch to a mainly share fleet will require the
use of a ‘mobile client risk profile’ which is a software-based tool actively
monitors the telematic feedback from whatever car the client is driving.
Thus, premiums are based on a driver, and their habits, and not linked
to any particular car. Note that this implies that monitoring client use
needs to be automated, as clients can’t be expected to inform the insur-
ers whenever they use a share car and exactly what they do with it. This
is especially true for autonomous cars, as users may send these on a trip
without physically being in them - like picking up children. Insurers
whose costs are not low enough to provide this type of micro-cover will
not last long.
McKinsey (2016) argues that insurers will need to complete a
range of tasks before they can move into offering successful dynamic
auto-insurance products. They need to (i) identify and engage with
 6  Impact by Type of Insurance    
187

an ecosystem to create the appropriate system of telematics, (ii) create

appropriate and sophisticated analytical tools to integrate data from dis-
parate sources, analyze it, and provide useful feedback, (iii) create user-
friendly customer interfaces, (iv) enlarge the customer data pool, (v)
digitalize interactions, and (vi) rapidly expand the depth and scope of
internal capacity. The vast increase in sensor feedback data, and the abil-
ity to adjust driving parameters based on feedback from other cars, will
mean that insurers will have a far more in-depth understanding of con-
ditions which led to crashes. They will then be able to set dynamically
adjusting premiums, based on those conditions.
There are major risk-pricing issues if the majority of driving decisions
are taken by software rather than a human driver. For example, do you
set premiums based on the risk rating of the software or the ability of
drivers to oversee it and take control if things are going wrong?
There does not seem to be an issue of customer reluctance to pur-
chase these semi-auto-drive cars. It is interesting to note that on social
media discussions of current 80% autonomous highway mode cars,
the predominant feedback was ‘why only on highways?’ BCG (2015b)
found that only 22% of those they surveyed were averse to buy a self-
drive car, and it would only take relatively minor evidence of lower
crash rates and therefore lower premiums, to convince most car owners
to switch. Capgemini/Efma (2016) shows that the younger generations
anticipate that they will self-drive cars in far higher numbers than insur-
ers expect. Evidence from auto-taxi trials shows that female passengers
strongly prefer taxis without a human driver and would pay more.
This indicates that the general public is generally accepting of auton-
omous drive vehicles, supporting the auto-manufacturers’ strategy of
introducing the technology in stages. After all, adaptive cruise control
and self-parking are now accepted as normal so the next step to autono-
mous highway cruising is small, and the next step to autonomous city
driving is small, and so on. It is insurance companies which are lag-
ging, and they will either have to catch up or lose their market to new
entrants. If we add to this the possibility of differing policy options like
pay-as-you-go premiums as well as rapid claim handling due to telem-
atic feedback, then insurers who are not market leaders or fast followers
will be facing bankruptcy.
188    
M. Naylor

An additional issue is reduced private car ownership. ITM/OECD

(2015) postulates that there is likely to be a move in larger cities away
from car ownership toward rental or toward shared cars. SwissRe
(2016) points out that car ownership rates in large western cities are
already dropping. Many urban drivers would car share if they could
find certified safe fellow-passengers going from around their home
to around their work place. This is an easily solved software informa-
tion issue, for which apps are already being developed. Auto-taxies are
currently being trialed and will be introduced widely by 2020. The
removal of the driver ensures that they will have low marginal costs,
especially if shared, well below the cost of driving yourself. BCG
(2015a) estimated that auto-taxis with that least two occupants would
be 35% less costly than current taxis and comparable with mass tran-
sit. Uber estimates costs will be more than 50% below current taxi
rates. BCG thus argues that within cities, auto-taxis or car-sharing will
become common, resulting in declining rates of private ownership -
why own an expensive vehicle which just sits in an expensive car park
95% of the time?
Current estimates by industry analysts are that by 2025, 25% or
more of all cars could be owned by fleet owners such as Uber or Lyft,
predominately autonomous drive. These fleet owners will have the
capacity to largely self-insurance, especially if the predominate cause
of accidents are software related. For insurers, this ownership switch to
fleets run by a firm like Uber is a concern as it will lead to sharply lower
margin bulk deals. Elon Musk argues that by 2030, the low margin cost
of driver-less drones means that individual ownership of a cars for urban
dwellers will seem like owning a horse3; something done for leisure
enjoyment. John Zimmer of Lyft argues that this transition could occur
by 2025.4

3https://www.businessinsider.com.au/elon-musk-owning-a-car-in-20-years-like-owning-a-horse-

2015-11?r=US&IR=T.
4 https://medium.com/@johnzimmer/the-third-transportation-revolution-27860f05fa91#.

kxmuehhxd.
 6  Impact by Type of Insurance    
189

House/Contents Insurance
The use of big data and analytical techniques means that house insur-
ance will be able to be priced to individual houses, using data like house
building materials, distance to neighbors, proximity to schools or trans-
port routes, soil conditions, flooding risk, etc. House telematics can be
linked to insurance company computers to report intruders, or fires,
etc., to accurate track risks in real time. Houses will have finger print
or voice-activated security linked to doors or windows. Sensors linked
to electrical wiring can shut down electricity flow if overheating and a
potential fire is detected. This will lead to less claims and thus lower pre-
miums. Aviva has mapped ground conditions in every UK city, allowing
them to risk-rank house by house.
House contents insurance also faces a substantial drop in premium
income, as nearly all products will come with networked chips, thus
enabling the owner or police to track the location of stolen goods in
real time, and remotely disable them. This will make burglary unprof-
itable. The end result is that burglary rates will plummet, fires will be
controlled faster, and customers will be discouraged from building in
higher-risk location. This will substantially reduce insurance premium
cash flow.
Estimates in general predict at least a 60% drop in risk for a con-
nected house. Of course, house upgrades occur at a far slower rate than
car renewal, but as insurers fight over an increasingly smaller premium
flow, profit margins will fall faster than risk levels. The overall impact on
premiums cash flow will be less than that of car insurance, as the risks
like flood or earthquakes will remain mostly unchanged.
However, given that general insurers tend to obtain 40–60% of their
income from car insurance rather than house/contents, then as car
insurance becomes a shrinking overall income pool, competition can
be expected to increase for house customers, thus reducing profit mar-
gins in all products to near zero. Given that the expense ratio for P&C
insurers in the USA (costs and claim over premiums) has increased from
96.2% in 2013 to 100.3% by 2016, any reduction in premiums due to
InsurTech would leave many P&C insurers financially non-viable. On
the upside, effective real-time customer relationship systems combined
190    
M. Naylor

with house level data will allow innovative new customer contacts and
potential income from sale of data-based services. For example, clients
can be alerted about bush fires, flooding, or tornados likely to affect
their house. Alerts about a recent spike in local burglaries can be pack-
aged with education from a related home security firm.
The increasing use of household robots will add an expensive new
item to be insured, and thus, a valuable new area of premium flow in
the short run. Insurers may be able to compensate for lower-risk pre-
miums by working with their business ecosystem to provide additional
services like linking with security monitoring firms to capture telematic
data for all kinds of risks houses face and respond to those in real time.
This can include health telematics so an appropriate response by an
ambulance etc., can be arranged. The insurer will only gain from this if
they are proactive and use their skills in data interpretation to set them-
selves up as the co-coordinating center of this response network, rather
than just a supplier of a decreasing value product.
Disruption in this sector can be expected to occur by 2020. Neos
Insurance in the UK is already offering house and contents insurance
alongside a suite of smart-house telematic-based services. Their profit
comes from service fees as their focus is on preventing adverse events
and reducing insurance need.

Life Insurance
PWC (2014) argues that life insurance products have tended to be
overly complex and expensive to distribute, and this has restricted sales
to well-off market segments or older customers, keep client contact to
an occasional payment, and has ensured low conversion rates and build-
ing client resentment.
Yoder and Rao (2015) argue that the main reasons for the percent-
age of western populations holding life cover dropping over time has
been (i) decreasing percent of the population aged 25–40, which has
decreased the market, (ii) a sharp decrease in death rates below age
70, which led younger couples to devalue the need for premature
death cover, (iii) increasingly product complexity and, (iv) a shift from
 6  Impact by Type of Insurance    
191

products being sold via workplaces toward selling to individuals. These

have led to an increase in the difficulty of selling products which has
necessitated the creation of a network of brokers and agents to explain
the products to clients. The commissions which this network requires
have led to unfavorable distribution economics, both reducing insurer
profits and reducing demand from clients.
Yoder and Rao then argue that Generations X and Y expect simple
products, transparent pricing, quick delivery, and being able to deal
with product suppliers where and when they want. Insurers have been
far slower than other product suppliers in recognizing this change in
consumer expectations and argue that this has led to three intercon-
nected downward vicious cycles. The first is that the need for advisers
leads to slower service, to decrease in product transparency, to increased
product complexity, and therefore a need for advisers. The second is
that increased distribution costs lead to decreased insurer profitabil-
ity which leads to decreased life sales which lead to slower service and
which finally leads to increased distribution costs. The third is that
increased distribution costs lead to customer reluctance which leads to
a need for insurance ‘to be sold not bought’ which leads to the need for
advisers which lead to increased costs. They contrast these cycles with
the mutual fund sector which has grown strongly and has experienced
a virtuous cycle of simple products, decreasing distribution costs, sales
without promotion, commissions dropping below 1%, increased sales
volumes, and rising profits.
Yoder and Rao thus argue that life insurance, and the associated
income replacement and trauma products, needs to be reimagined.
They need to be made simpler, easier, and faster to obtain, with cheap
and transparent pricing and distribution costs. They need to be made
more relevant to the lives of young couples who see premature death
as extremely unlikely. Life insurers need to reduce administration costs
to near zero and offer simpler higher quality products. This can be
done by allowing customers to self-create policies by combining blocks
of products into a suite. The focus has to be on engaging more deeply
with customers so purchases are made with greater understanding, so
a lifecycle services approach can be created. Data analysis has already
proved to be highly predictive for death rates, especially during the last
192    
M. Naylor

two years of life. One US insurer is using its predictive end-of-life data
model to provide estate/tax/trust/counseling/palliative care services.
Similarly, telematics could potentially have a large impact on life and
health insurance, engaging clients with insurance services, adding useful
new sources of insurer income, and providing a stream of valuable data.
One suggestion Yoder and Rao make is to shift emphasis from death
toward promoting health and quality of life. Wearable bionics can then
track blood contents and fitness levels, scan for signs of sickness, and
alert clients to potential problems. Discounts can be given for health-
related activities, warnings given if clients engage in unhealthy activi-
ties. They argue that this shift will change client perceptions about the
insurer, integrate them as a part of weekly life, and encourage younger
people to take out policies. Even though health telematics are still
crude, several companies have tried this approach with significant suc-
cess in terms of client reaction.
Yoder and Rao (2015) argue these advances, combined with auto-
mated distribution and underwriting, will decrease costs and induce
younger, more affluent, clients to see cover as relevant and therefore
proactively seek out cover further lowering distribution costs. It will
allow clients to be individually underwritten and grouped into small
market segments. It will allow detailed data analysis so that deep under-
standing can be gained around the relationship between client activities
and well-being and sickness.
Feeding the insights back to clients so they can understand how their
life choices impact on their health will both improve client life choices
and create the sense of connectedness which the insurance sector is cur-
rently lacking, and provides a clear value-proposition to clients. Note,
however, that insurers will find it difficult to get clients to agree to share
a real-time stream of personal information unless they can establish
both inherent trust and a clear benefit. From the subsequent reaction of
clients to this increased information, insurers can then build behavioral
models of how to incentive clients to reduce their risk profile, possibly
by trialing differing methods with separate groups. The increasing rich-
ness of the data will build a virtuous cycle, whereby insurance is sought,
not sold.
 6  Impact by Type of Insurance    
193

McKinsey (2015) argues that competition will ensure that most of

the benefits of the changes will flow to clients in terms of lower pre-
miums or increased access to health information. Insurers who do par-
ticipate will face large investments in data collection tools and software.
Insurers who do not participate will rapidly lose market share among
the non-chronically sick, and will face eventual bankruptcy.

Health Insurance
Health insurance has also been heavily criticized for offering an overly
complex and unnecessarily diverse product range. This results in a high
lapse rate and subsequent high product distribution costs. Energesse
(2015) argues that the major reason for policy churn in Australia is the
customer confusion caused by customers being offered a choice between
more than 25,000 different policies. High lapse rates (over 20% of cus-
tomers per annum with some insurers) mean high lost revenue and a
significant financial impact on insurers due to the relatively tight net
profit margins. In addition, insurers and customers waste a significant
amount of time negotiating and resolving issues related to poor pur-
chasing and claims experiences. Foster (2012) found that while 80% of
companies claim that they deliver superior customer service, only 8% of
customers agree. Given that 70% of customers decide to buy based on
how they feel they are treated, this is a major industry issue.
Energesse (2015) found that the highest lapse rates for health insur-
ance occur within 6 months of purchase, and that the number one
reason for purchase was price. They argue that this is caused by two
factors: (i) customers find health insurance products confusing and
do not understand the value of the product. It is thus a grudge pur-
chase dominated by fear. In the medium-term customers try to ration-
alize their purchase and suffer regret, as they realize the product does
not meet their actual needs. (ii) Insurers are thus forced to compete
on price so have developed products with complex exclusions and
restrictions. Customers do not initially understand these, but in the
medium term, they either realize the product’s poor quality or feel
uneasy at the non-understandable complexity. This creates a vicious
194    
M. Naylor

cycle in customer retention and experience, whereby customers struggle

to understand benefits, so they switch to cheaper products or options,
which then do not provide the required outcomes at claim time, which
then sour them on the concept of health insurance.
Energesse argues that the real reason for poor customer experiences
can be summarized into these four major themes: (i) customer percep-
tions of confusion and lack of value, (ii) regulatory and competitive
forces, (iii) sub-optimal systems, processes, and data management and,
(iv) health system dynamics.
There are some key factors inhibiting client perception of value. One
factor is that health insurance has unusual delayed gratification, so that
many customers do not understand their product choice until claim
time, when it is too late to switch providers. The second factor is that
since only 5–10% of customers ever make a claim, most customers per-
ceive very limited value. The third factor is there are few opportunities
for the insurer to engage deeply with the customer. The fourth factor is
that payments tend to be made direct to health providers so that cus-
tomers may not be aware of the costs of their treatment. These factors
mean that today’s customers perceive lower levels of customer service
from health insurers than they do from other companies which they
deal with. There is thus a substantial likelihood of customer switch if
those companies offer insurance. The current structure of the insurer–
client relationship means that there are very opportunities created for
the insurer to impress the client with superior engagement. This has to
change, as was discussed in an earlier chapter.
These issues can only get worse given the drive toward individualiza-
tion of underwriting and increased customization. The only solution to
these issues is for health insurers to reinvent their administration sys-
tems, and offer simpler‚ higher quality products. As with life cover, the
way to do this is to use software as the base administration system and
create policies by combining blocks of products into a suite which cus-
tomers can pick and choice from. The focus has to be on engaging more
deeply with customers so purchases are made with greater understand-
ing. There also needs to be greater integration between the software
systems of insurers and those of health providers, so that the customer
experience is seamless.
 6  Impact by Type of Insurance    
195

Given the complexity of health insurance products, involving

humans in giving advice is vital. These staff need to have excellent, com-
plete‚ information on customers, as well as the ability to transfer client
advice to dedicated area specialists. The software then needs to be capa-
ble of generating policies, handling claims, as well as interacting with
hospital software systems and extracting required information from doc-
uments created by a wide range of medical staff and arranging health
care. This needs to be in real time, so that transactions can be completed
while the customer is being advised. Digital natives expect responses in
minutes rather than days, let alone weeks.
An example of health innovation is the agreement in the UK between
DeepMind and the Royal Free NHS Trust to use algorithms to assess
and monitor kidney function tests and send alerts when blood results
are abnormal. In the UK, 10,000 people a year die from preventable
acute kidney problems. The App can also be used to monitor sepsis,
liver dysfunction, and general organ failure.
Health insurance will be as affected by telematics as other insurance
areas, as the growing use of wearable biosensors will allow hospitals and
insurers to collect real-time data on the biochemical changes which
occur before a person gets sick, and thereby start to predict illness and
alert that person to seek appropriate medical help. The positive media
news an insurer would receive from such a lifesaving response would be
invaluable.
The problems with current measurement methods of patient health
is that: (i) they provide only episodic readings, (ii) they can require
high-cost physical collection, (iii) patients often fail to adhere to treat-
ments, and (iv) they do not detect problems which occur before a
patient exhibits chronic symptoms. Currently, if a doctor thinks there
is an issue with your health, (i) they have to see you, (ii) request a
blood test, (iii) you have to physically attend a blood test clinic, (iv) the
blood sample has to go to a specialist, (v) the results have to be sent
to the doctor, and (vi) you have to go to the doctor to discuss treat-
ment options. The test also only shows blood contents at the time of the
test. Continuous monitoring via biosensors overcomes all these issues,
as they provide dynamic real-time data and sends real-time blood analy-
sis to the doctor’s AI program, alerts them if an issue occurs, and tracks
196    
M. Naylor

blood contents vs activity on a 24/7 basis. This can be linked to sensors

in your bathroom scales and what your food cupboard thinks you ate.
McKinsey (2015) argues that the use of biosensors will cut the cost
of treating chronic disease by 50%. These sensors include (i) wearables,
which collect information via the skin, (ii) implantables or injectables,
which collect more intimately, and (iii) non-wearables, which collect
information periodically such as Wi-fi-linked scales. These can be unob-
trusive - MC-10 and Sano Intelligence already produce patches, ‘bios-
tamps,’ which stick on your skin and connect to a mobile phone App.
These stickers can detect a broad range of metrics, e.g., blood pressure,
hydration, glucose, oxygen saturation, temperature, heartbeat, stress,
sunburn, sleep levels, wound healing, drug levels, and movement pat-
terns. They are powered by mobile phone transmissions. MC-10 is
working on stickers which can be fitted to internal organs. They could
also be used to control pill boxs or medicine injectables.
While there are serious regulation issues with implantables or
injectables, beta-versions of wearable devices are available now and are
expected to exceed 1.3B by 2025. Injectables are in clinical trials and
should start to be available by 2020. In the meantime, there has also
been work on including sensors in mobile phones which can analyze
blood components by scans. Pills bottles will track when medicine is
taken, and blood sensors can then track the biochemical response to the
medication. Hats can track brain patterns. Mobile phones can record
food purchases. More intrusive telematics will face medical trial delays
so maybe only be used outside of higher-risk patients before 2028.
Nearly, all data, however, can be collected externally; for example, bowel
microfauna can be analyzed from a toilet-based sensor. Some sensors are
now the size of a full stop.
A market segment to start trials with is patients with critical health
needs, who should agree to wear cumbersome sensors. Health special-
ists can decide which metrics need collecting, and data analysts can then
start to analyze the flow of data to both feedback alerts to doctors and
to ascertain if trends can be found which predict adverse changes before
they occur. Another market segment will be those keen on fitness or
professional athletes, who will agree to wear sensors so they can learn
from the data analysis. From these experiences, engineers will discover
 6  Impact by Type of Insurance    
197

how to make sensors smaller, and data analysts will discover enough
useful features to induce the average user to wear them, as well as induc-
ing insurers to pay for them or offer discounts to wearers. Note that this
is not the far future as these devices are rapidly approaching a useful
state, probably within the next five years.
If we take the case of an elderly woman with congestive heart failure
and diabetes: beside her bed there could be a sensor to record her weight
as she stands up; bed sensors monitor pressure points to detect early
signs of ulcers; Sensors in the floor monitor her walking gait to assess
her risk of falling; A patch on her arm monitors her heart rate, blood-
oxygen level, blood pressure and glucose levels; Drinking cups and toilet
sensors monitor her hydration levels; Sensors in her clothing can moni-
tor if she has collapsed as well as track heart rhythms; Her medication
pill has a tiny sensor which tells her arm patch that she has taken it.
All these sensors automatically transmit data via her phone App in an
encrypted format to the public health system. The data is analyzed for
risk factors, and medial help is sent if required. Why should an elderly
person be provided with a medical alert which they have to activate after
they collapse when sensors can automatically request help before they
collapse because health metrics drop below preset levels?
Recent technological conferences have featured experimental but
already existing items like (i) bionics which monitor health and are barely
visible plasters, which can send out data via the wearer’s mobile phone,
(ii) exoskeletons, for use in dangerous of heavy lifting situations, which
are predicted to cut injury rates by 70%, and (iii) body suits with neuro-
interfaces which can exercise limbs of paraplegics in a way which either
encourages nerve regrowth or allows limb control directly from the brain.
Horvath (2013) argues that biomarkers involving the epigenome and
methylation systems are good predictors of how rapidly a human body
is aging in response to environmental and lifestyle factors.5 Accessing
these results will provide accurate mortality profiles, helping under-
write life insurance on a more individualized basis. Insurers could also
set rates based on promised change in lifestyle and use periodic tests of
these biomarkers as feedback to check if these promises are kept. This
provides data to underwriters and is an incentive for customers.
5For a summary, see Gibbs (2014).
198    
M. Naylor

The range of data that can be collected is widespread: heart rate and
patterns, temperature, glucose, blood serum, enzymes, hormones, sleep
patterns, body fat, activity levels, mood, blood chemistry, oxygena-
tion, EKG brain functions, and bowel microfauna. Wound dressings
can track healing progress. Employers can use watches or wrist straps
to detect workers who have used illegal substances or who are drowsy in
unsafe environments. All can be tracked in real time, 24/7. The analysis
of data should discover trends which predetermine sickness and enable
doctors to call in and treat patients before they get sick. Patients who do
not take medicine can be alerted and warned.
The analysis of this river of health big data from millions of patients
from diverse sources like physical activities, eating patterns, sleep,
health metrics, and food purchases, will provide many links to be
made between social habits and disease or between the specific genes
of a person and the lifestyle they should lead. This will not only allow
health care providers to set up alerts for intervention with the client,
which maybe immediate or asking the patient to come in, but it will
also provide a rich and deep source of data linking everything the cli-
ent has done prior to the adverse health event, including physical activ-
ity or changes in blood chemistry due to food etc., to health outcomes.
This will give health researchers the capacity to understand physiological
reactions and segment clients by their reaction type. For example, UK
insurer, Aviva‚ has combined data from a range of non-health sources,
like shopping or online behavior and has found that these can predict
future health outcomes nearly as accurately as blood or urine tests.
Patterns of Facebook ‘likes’ have been found to be a good predictor of a
number of health indictors. This illustrates the importance of an insurer
exploiting the advantages of data sourced from the wider ecosystem.
The largest collectors of health data are likely to be healthy, rather
than sick individuals. The availability of cheap and easy to use sensors
embedded in clothing will allow people to quantify their health over
their life span. Linking telematics in clothing, household appliances,
mobile phones, toilets, computers, etc., will allow people to track their
health and their productivity to find patterns: Are they most produc-
tive with e-mails on a Wednesday afternoon? What level of coffee intake
maximizes decision making while minimizing heart stress? How does
 6  Impact by Type of Insurance    
199

that third beer on a Saturday night impact on belly fat? What pattern
of weekly eating produces the best gut microfauna? Those who seek fit-
ness can have sensors in their watches or shoes to give feedback from
blood or muscles as to improvements and the effect of differing exercise
regimes. Each house will own a pill maker, which will use daily data
on blood chemistry, heart patterns, gut microfauna patterns, and gene
details to produce the optimal daily vitamin. A medical program will
provide a monthly health and productivity report.
Medicine can then move away from an annual visit to a doctor who
on very limited data provides a ‘this-tends-to-cure-people-your-age-
&-gender’ toward a personalized prescription based on detailed per-
sonalized data and a huge data bank of outcomes from similar people
internationally. Medical research is based on trials involving as few as
500 heterogeneous people for a very short time frame. Big data would
provide data on millions of people on a continuous basis. This would
enable data on the ‘people-99%-like-you’ to be extracted and a person-
alized cure created. Insurers can offer inducements to clients to follow
a personalized optimal health/activity/eating plan and provide regular
health information highly applicable to you.
Before this data can be useable, however, it needed to be standardized
and its collection automated. In many countries, medical records are still
based on paper, so systems will need to be changed so all medical personal
data is inputted digitally probably via tablets. Another issue is that cur-
rently, most medical records are kept in proprietary formats/data stand-
ards, and providers view holding the data as a competitive advantage to
be denied to competitors. Yet, the economics of data scale dictate that the
network of companies with the largest and most diverse data collection
will gain the most insights and be able to price most accurately. Thus,
companies who cooperate by forming ecosystems to increase data size will
progressively pull ahead of companies who do not collaborate. Linking to
public health or hospital data is essential‚ preferably internationally.
Social capital will be vital so that customer trust can be gained so
vital and confidential health data can be gathered. Industry wide meth-
ods of secure data storage may need to be created so there is an infra-
structure of trusted, neutral, and secure storage. On a positive note,
many hospital software systems are currently used to compare the
200    
M. Naylor

outcomes of treatments across a wide range of patients, so that add-

ing in telematic data will be easier than in other insurance areas. Many
hospital risk prediction systems already accurately predict the risk of
death or complications from different types of surgery based on per-
sonal factors including genetics, as well as other areas like length of time
required to recover. Riskprediction.org.uk provides this to the public.
The access to a huge database of health, medical, and lifestyle met-
rics will enable insurers to analyze treads and predict health problems‚
or relate health metrics to exercise in a detailed way. Vital here will be
access to as much data on client lives as can be obtained, so that data
analysts have a complete picture of what aspects of life cause health and
ill-health. The successful insurers will have to convince clients to grant
them access to a constant steam of real-time lifestyle data. For exam-
ple; men who skip breakfast have a 27% higher risk of heart attacks,
yoga participants have less mental diseases, while those who buy alcohol
in the evenings have shorter lives than those who buy alcohol in the
afternoons. Note that these types of data outputs are often not directly
­casual, as a third factor may be causing both to change - those men who
skip breakfast have more stressful lives.
The combination of genomics, IT, nanotechnology, biotechnology,
and cognitive science will create ‘precision medicine,’ whereby doctors
will be able to tailor treatments to a patient’s individual biochemistry,
and remotely track in real time how the treatment is evolving and make
changes if required.
General Electric (2012) argues that upgraded health systems software
connected to smart devices could cut health system costs by over 25%.
They add that information inefficiencies in health care exceed 10% of
current costs, and as such the sector is ripe for technological disrup-
tion. Given that health care accounts for about 10% of global GDP, or
US$7.1Trillion in 2011, the impact of a 25% cut is substantial. These
substantial possibilities to reduce health care costs and the resultant
commercial opportunities is generating a flood of telematic InsurTech
investments, estimated to be US$120B p/a by 2020. This will progress
fastest in countries with existing quality hospital software systems. This
incentive is likely to ensure that health insurance a fast follower behind
auto-insurance in the timing of transformation. Insurers in the USA,
 6  Impact by Type of Insurance    
201

who spends about 18% of GDP on health, and where hospitals already
tend to have quality data systems, are likely to be leaders.
When combined with a multitude of other health innovations in
development, per patient spending on health will drop substantially,
leading to a drop in health insurance premiums and therefore cash flow.
The aging of the population may hold up aggregate spending, though
a large share of innovations is focused on diseases of old age. The need
for heavy investments in telematics and associated analysis with a stag-
nant cash flow will be difficult for insurers. They may choose to segment
their client base into those prepared to contribute more toward inclu-
sion in the full program and those left within the old system.
Those insurers who do invest will find costs dropping as data is col-
lected remotely, is analyzed by software, adjustments are made to medi-
cine machines by software, emergency teams are sent by software, doctors
are consulted remotely, and humans are only involved if the software
finds trends exceeding preset parameters. Costs related to precautionary
tests will substantially drop. For example; AI image recognition is already
having an impact on the diagnosis of CT or X-ray scans for detection of
abnormalities, as leading AI systems have proved to be more than 50%
better than human specialists at reviewing scans with radically lower cost.
One of the largest impacts of telematics and big data will be on
claims. Because insurers will be collecting real-time data on a wide
range of factors, they should be aware that an adverse event has
occurred in real time and should have automatically collected nearly all
the elements required for a claim without the client informing them.
If a client falls sick, the health bracelet will tell the insurer that a medi-
cal emergency has occurred, and send a record of recent health metrics‚
while the insurer’s computer can arrange emergency assistance, analyze
scans, and make payment of health bills without any need for patient
involvement. This proactive approach will increase insurer engagement
to a qualitatively different level, increase the social capital of the insurer,
as well as allow cross-selling of more services.
The key benefit, however, will occur from a switch in insurer focus
from ‘sickness’ to ‘health’ as they move to predictive pricing. Insurers
can offer relevant, real-time, health tips to clients by providing real-
time feedback on the metrics gathered from telematics. This will give
202    
M. Naylor

them the opportunity to interact with healthy customers on a regular

basis and be seen as a ‘positive’ company. There will also be substantial
opportunities for sales of data-based value-added services.
Successful insurers will thus have the opportunity to become more
involved in clients’ lives because of their expertise in real-time informa-
tion management and analysis. Insurers will have to switch from the
current focus on cost and control of benefit payouts to heavy focus on
customer responsiveness and creating reason to engage. For example,
AXA Singapore’s health product arranges and pays for: a person to fetch
children to and from school, a nurse to attend to follow up treatments,
rehabilitation services, for doctor’s home visits, modification to the
home to accommodate disability, transportation to the hospital, house
cleaning and meals, and psychologist consultations. This establishes the
insurer as a positive, helpful, partner in life and grows social capital.
Another growth area will be the use of domestic home-health robots
to provide care. These can be used to decrease the need for long hospi-
tal stays, or to free other family members, or provide for elderly. Given
that these domestic robots will initially be expensive, about the price of
a luxury car, insurers can start to add provision or purchase of these as
an attractive policy option if specified health conditions occur. These
robots will also need intensive servicing and linkage to a health care
telematic and monitoring system. Any health insurer who manages to
position themselves as the coordinating hub of a health ecosystem will
not only receive a significant level of cash flow, but they will also have
the chance to cross-sell, to gain intensive data about client physiology,
and to create proactive communication avenues with clients who will
welcome relevant information.
Successful insurers will also have to be heavily involved in social
media and be proactively aware of ratings/review systems. Insurers who
manage to switch their focus to being data intensive will find that this
will create an access to the rapidly expanding markets in other areas of
information management as our world becomes deeply interconnected.
Given that society is aging and there is a tendency to increase the pro-
portion of income on health as people get richer, the volume of cash
flow from health-related services should rise, which will give health
insurers the incentive to invest heavily, as well as give outsiders the
incentive to enter.
 6  Impact by Type of Insurance    
203

Note the degree of sensitive personal health information provided

by biosensors and other telematics mean that there is an extraordinary
high level of client trust required for insurers to successfully create this
transformation. This will be a major barrier for existing insurers who are
not currently creating a dynamic proactive client communication strat-
egy. Many insurers with established low trust levels due to focus on cost
control will find it impossible to establish the required high level of cli-
ent trust to access the data.
An aspect of this trust issue is that major issues arise if health biosen-
sors are vulnerable to hacking, especially if they are linked to medica-
tion systems. Therefore, proving a high level of system security will be
vital. Another aspect is that clients will be relying on health telematics
to provide appropriate emergency health responses, so that clients are
literally placing their lives in health ecosystem response software. Given
the low social capital of existing insurers, clients may find it hard to
trust their lives to the integrity of the insurer’s software and systems, and
incumbent insurers may find it hard to build a high enough level of cli-
ent trust. The sector is thus open to disruptive entrants with technical
knowledge and existing high social capital.
Overall, the extremely large size of the health sector and health insur-
ance, combined with what is perceived to be high levels of waste and
low-quality IT systems, makes it a very attractive market for disruptors.
Regulation, experience, and the scale to negotiate costs would seem to
make incumbents attractive partners for disruptors, but their legacy sys-
tems and conservative culture discourages disruptors from working with
them. There will be a longer delay in market disruption in health, but
the likelihood of disruptive external entry is high. First-moving insur-
ers who do survive without adverse incidents will find that the positive
media from providing lifesaving real-time services will give them a near
unbeatable competitive advantage.
The increasing accuracy of health risk predictive systems raises
the issue that some clients with high-risk factors will only be offered
restricted insurance or at high rates. This raises important regulatory
issues, which are discussed in a subsequent chapter. However, feedback
between premiums and lifestyle indicators has the potential to improve
overall health outcomes. AllLife, which links premiums to following
doctor’s advice, has shown that nearly all clients improve their lifestyles.
204    
M. Naylor

Commercial Insurance
The use of telematics in commercial transport fleet is already routine,
as the real-time data about vehicle usage, vehicle wear, and goods track-
ing, is useful to transport businesses. Linking these sensors to insurer
networks is therefore easier than linking private automobile fleets. Real-
time container tracking via telematics is also routine. These existing
telematic sensors lead naturally into an automated claims process and
flexible underwriting. Dynamic insurance solutions are therefore already
being trialed for large commercial transport operators, and will rapidly
be expanded into all areas of commercial transport and marine insur-
ance. Morgan-Stanley (2016b) argues that telematics are already dis-
rupting small business insurance, and the impact will be affect 25% of
the market by 2020. This can be expected to rise to 60–70% by 2025.
Some commentators argue that because commercial insurance con-
tracts and processes tend to be customized, this is hindrance to the use
of process automation. I would argue that this is not true; software can
easily be designed for customized systems if AI systems have enough
programable rules. In fact, the large size of commercial policies and the
existing widespread use of telematics make commercial insurance an
ideal test-bed for the creation of modularized customer-selected pol-
icy construction and automated processing. The size of these contacts
means that small insurance savings add up to substantial amounts so
pressure will be on insurers to offer the most modern deals. Insurers
who offer proactive monitoring and claims services will be welcomed.
Even more welcome will be insurers who use the collected data and AI
analytical skill to provide accident/loss/fraud avoidance advice or acci-
dent/loss response services.
Note, however, that since firms like Octo-Telematics have been pro-
viding this type of data and associated analysis since 2002, covering fleet
use, location and use, as well as crash-causation, insurers who seek to
enter this market in competition have a fifteen-year lag to overcome.
Commercial transport fleets and associated goods can therefore be
seen as the area were procedures and protocols around data, commu-
nication, and policies will be explored and defined. Insurers who aim
to survive the looming disruptive waves will thus have to proactively
 6  Impact by Type of Insurance    
205

seek opportunities to offer dynamic insurance so they can gain expe-

rience with real-time data collection, use, and analysis. The ability to
respond to adverse situations as they occur, to handle the huge flows of
often incompatible data, and to integrate these into a platform will be
essential skills to be developed. To do this, they will have to integrate
themselves into the transport ecosystems. For example, they will have to
work with telematic suppliers to ensure that the correct range of data is
collected.

Cyber Security, Drones, and Hacking

The proposals outlined in this chapter assume that insurers will be col-
lecting an immense amount of personal and confidential data about
their clients including real-time location and links to the control of
items like cars or vital medical equipment. Clients will only agree to this
if they have a high level of trust in the insurer’s honesty and data secu-
rity ability.
These databases and links will be very attractive targets for hackers,
either to commit mayhem or blackmail. Data security needs to cover
both the insurer’s IT system and any customer telematics which the
insurer could be seen as liable for. Already, multiple monthly cyber-
attacks on insurers occur, and this can only exponentially rise as the
value of the data rises. Given the high level of client trust required for
the data collection, any successful hacking event could immediately
destroy the reputation of the insurer, and given the reduced brand stick-
iness, the insurer may find that their business rapidly disappears. Thus,
advanced level data security will have to be the top priority.6
The downside to telematics is their vulnerability to hacking. This
is especially true for smart household appliances, where research has
shown that 90% of owners do not change the default password. This
hacking could be malicious to cause damage, to ascertain if the owner is
home, or to spy on your activities. While smart appliances are worthless

6For a detailed cyber security summary, read Morgan-Stanley (2016a).

206    
M. Naylor

to steal (as they can be tracked), blackmail unlocking scams could be

common. Insurers may have to start insisting on professional installa-
tion. The hacking of health telematics in high-risk patients is extremely
concerning as changes to these could cause death. Insurance relating to
smart device hacking could thus be a useful growth area, as would be
the sale and service of IT firewall systems.
Insurance relating to drones is another growth area. SwissRe (2014)
extensively describes the potential uses for drones and argues that they
will have a reasonably high rate of insurable risk and therefore attract
a useful premium flow. They argue that spending on drones will reach
$US11B per annum by 2023.
SwissRe (2014) also argue that drones will become a standard tool
for claim assessors as they can survey locations easier and faster than
humans can, when a physical presence is required. For example, if insur-
ers pre-arrange a national network of drone bases, then assessors can
assess a claims location remotely from thousands of miles away. With an
array of built-in sensors and an ability to crawl, drones could cover 90%
of assessment tasks. This will be particularly useful for the agricultural
sector, where a farmer can show an assessor crop or other damage even
in remote locations which have no roads. There will be no need for an
assessor to climb ladders or enter dangerous areas. A recent California
start-up, DropIn, assesses automobile claims by using claimant mobile
phones to conduct interviews and drones to assess the crash site. They
use a nationwide network of over 1000 drone operators contracted
to respond immediately. Often no operator is needed; for example,
AI linked to a drone can assess roof tile damage after a storm with no
human involvement by using visual recognition tools.

References
Boston Consulting Group. (2013). Telematics: The test for insurers. BCG
Perspectives.
Boston Consulting Group. (2014). Bringing big data to life. BCG Perspectives.
Boston Consulting Group. (2015a). Robo-taxis and the new mobility. BCG
Perspectives, April 21.
 6  Impact by Type of Insurance    
207

Boston Consulting Group. (2015b). Revolution in the driver’s seat: The road
to autonomous vehicles. BCG Perspectives, April 21.
Capgemini/Efma. (2016). World insurance report.
Energesse. (2015). ‘Future Solutions in Customer Experience for Health
Insurers White Paper’.
Foster. (2012). The customer experience index.
General Electric. (2012). Industrial Internet: Pushing the Boundaries of Minds
and Machines. P. Evans & M. Annunziata (Eds.). USA.
Gibbs, W. W. (2014). The clock-watcher. Nature. 508.7495 (Apr 10, 2014):
168–170.
Gomes. (2014). Driving in circles. Technology. October 21.
Horvath, S. (2013). DNA methylation age of human tissues and cell types.
Genome Biology, 14(10), 3156.
International Transport Forum/OECD. (2015). Urban Mobility System
Upgrade: How shared self-driving cars could change city traffic.
KPMG. (2015). Automobile Insurance in the Era of Autonomous Vehicles: Survey
results, June.
McKinsey. (2015). The Internet of Things: Mapping the Value Beyond the Hype.
McKinsey Global Institute.
McKinsey. (2016). Shifting Gears: Insurers adjust for connected-car ecosystems,
May.
Morgan-Stanley. (2016a). Cybersecurity: Rethinking security. Blue Paper. USA.
Morgan-Stanley. (2016b). Insurance—property & casualty—digital disruption
in small business insurance. North America Insight, June 29.
PWC. (2014). Reinventing life insurance.
SwissRe. (2014). Insurance and the rise of drones. New York: SwissRe.
SwissRe. (2016). The future of motor insurance: How car connectivity and ADAS
are impacting the market. Berlin: SwissRe/Here.
Yoder, J., & Rao A. (2015). Insurance at a tipping point. Insurance Thought
leadership, PWC, June 20.
 7
The Dynamics of Decline

Introduction
Most industry commentators agree that the insurance industry is facing
unprecedented innovations. The issues in dispute are: (i) Are the innova-
tions incremental or disruptive, (ii) How long the waves of innovations
will take to disrupt current market conditions, and (iii) Can incumbents
react in time to survive?
The answers to these questions can only be speculative and will dif-
fer for each sector of the insurance market. Analysts need to approach
these questions using a range of innovation disruption concepts. When
an analyst discusses the future, it is well understood by the industry that
any predictions involving exact dates are likely to be wrong. The useful-
ness of analyst reports is instead found within the conceptual reasoning.
This chapter uses auto-insurance as an example to explore the con-
cepts which underlie the dynamics of decline. These decline dynamic
concepts can be applied to other sectors. Cognizant (2017) found
that few insurance managers regarded changes in the automobile mar-
ket as worthy of concern before 2030. I will show that this lack of
concern could be based on misunderstandings about dynamics of
­

© The Author(s) 2017 209

M. Naylor, Insurance Transformed, Palgrave Studies in Financial Services Technology,
DOI 10.1007/978-3-319-63835-5_7
210    
M. Naylor

sector decline as disruptive change often occurs far faster than incum-
bents expect. Remember that modern mobile phones were only intro-
duced in 2007, yet now dominate our lives. As noted in the Chap. 1,
history shows that 3/4rds of social change occurs in the last 1/4 of the
disruption time period. Therefore, ascertaining where on the disruption
time-line each insurance sectors sits is vital.
The auto-insurance industry is significant. In the USA, it generates
$220 billion in annual revenue and supports 277,000 jobs, about the
same number of mechanical engineering jobs. For many P&C insurers,
auto-insurance is over 50% of their total cash flow. Therefore, under-
standing its demise is critical.

Profitability vs Premiums
A key concept is that there is a difference between the rate of spread of
a technology and changes in profitability, in this case, spread of autono-
mous cars and the decline of auto-insurance profitability. If high-end
cars embed 80% technology by 2018, then this will probably be embed-
ded in middle-end cars by 2021 and low-end cars by 2025. If there are
large differences in crash rates or telematic costs drops due to scale, then
this may happen faster. This implies that new low-end manual cars (less
than 80% auto-drive) will still be on sale in, say 2023, and cars sold
then will not be off the road before at least 2030. It is thus likely that
the majority of cars on the road in 2025 will still be manual.1
Note that there is a ‘valley of death’ for crash rates between 80% cars
and 100% automatic. This is because more than 80%+ cars require so
little of the driver, that despite being at the wheel, they are likely to
not pay active attention to the road, and instead do other activities.
Yet these software systems are not perfect so accidents will still occur.
Therefore, there a strong argument that level 4 autonomous cars are

1Germany has recently passed a law-banning new cars with internal combustion engines from
2030. This means that all new cars will likely be auto-drive. Given low crash rates, similar laws
can be expected to ban manual drive cars.
 7  The Dynamics of Decline    
211

more dangerous than either level 3 or level 5 cars, and therefore car
manufacturers will halt at level 3 and then jump to level 5.
What needs to be clearly understood, however, is that the demise of
the car insurance industry is not about percentages of cars with X or Y
type of technology, nor is it really about declines in premiums flows.
Demise dynamics depends instead on profitability. The general rule is
that at the point when premium cash flow starts to decline by 5 or 10%
per year, the natural reaction from insurers will be to cut premium
profit margin to retain market share. This tends to lead to ‘destructive
competition,’ where in order to increase market share in the hope of
holding premium income constant in declining markets‚ companies
compete until profit margins turn negative. This means that car num-
bers and premium flows could still be at 70–80% of previous levels,
but insurance profits may no longer be positive. Profitability falls faster
than premium flow, which falls faster than the technological change.
This drop in profitability will also impact heavily on insurer stock
value. If incumbents cannot convince the market that they have a via-
ble path to the future then stock markets will sell, which will severely
handicap insurers’ ability to raise the required funds to transform them-
selves. An example of this is a comparison of the 2017 stock value of
Tesla versus Ford.
If we apply this analysis to the suggested rate of technological spread
in the first paragraph, this implies sharply decreasing auto-insurance
premiums from 2020 on. In addition, since the richest customers will
be the owners of the 80% cars, whose level of safety implies low premi-
ums, an increasing proportion of premiums will come from insuring the
lower income owners of the remaining manual cars. Insurers who want
to remain will thus have to compete for a shrinking pool of decreasing
quality clients, or those few who chose to manually drive as a hobby.
The winner in this kind of market is the supplier who is prepared to
supply at or below cost. The only firms with an incentive to stay in this
type market are car producers, or those niche insurers used to providing
specialist cover for higher risk clients.
Thus, it is vital to understand that the turning point for car insur-
ers does not occur when 100% autonomous cars arrive on our roads in
substantial numbers, which is a decade away, though selected brands
212    
M. Naylor

will introduce these by 2020. The insurers’ turning point occurs when
a sufficient number of 80% autonomous cars exist to substantially cut
crash rates, and thus cut premiums, and thus cut insurer cash flow, and
thus cut profit margins. The demise of auto-insurance profitability is not
about the rate that 100% auto-cars are introduced but is instead about
the full package of current technology, 80% auto, which is being intro-
duced from 2017. Data shows that 80% technology cuts crash rates by
about 40%. If we remove the crashes predominately caused by the other
driver, this increases to a 70% drop in crash rates. 90% auto-tech, due
by 2020, will cut crash rates by over 80%. Premiums for top-of-range
cars may thus fall by corresponding levels. The speed of diffusion of this
technology down to the average car is a debatable point, but it’s likely to
underway by 2023, with an 80% transition by 2035.
At that turning point, those insurers who get more than 40% of their
cash flow from car insurance will be forced to either back out of the
market and prepare for sale to another owner‚ or start to compete inten-
sively in an arena of a declining market. Some commentators will argue,
correctly, that ‘car insurance will still exist for decades.’ I’m sure it will;
people will still drive manual cars for fun for decades. My response is
that horse insurance still exists, but I’m not sure that current insurers
visualize their future in that way. In leisure driving market, niche auto-
insurers are more likely to survive than the major ones.

Tipping Point
The crash rate of an individual car, however, does not just depend on its
own level of technology; it also depends on what percentage of other cars
have crash avoidance technology. Therefore, if some cars behave so as
to avoid crashes, avoiding cars being driven erratically, then it becomes
harder for manual cars to crash. There, overall crash rates do not just
depend on the rate of spread of auto-drive technology; it also depends on
the overall proportion of cars with crash avoidance technology in a par-
ticular area. The existence of an increasing number of 80% cars will mean
that crash rates will start to fall faster than the spread of the technology,
as these 80% cars increasingly detect and avoid straying manual cars.
 7  The Dynamics of Decline    
213

This is not a linear process but is a downward S curve. At a certain

point called a ‘tipping point,’ the environment changes rapidly from
one where one type of behavior dominates to one where another type of
behavior dominates. In a networked environment, manually driven cars
can be identified well before the auto-car gets to them and they will be
given a wide margin. At this stage, it becomes difficult for manual driv-
ers to find a car to crash into.
Studies show that for this to occur the percentage of 80% cars does
not have to be close to 100%, even 50% is enough. Using models
derived from diseases and financial crises, it can be argued that there will
be a tipping point when about 40–60% of current cars have 80% tech-
nology, though a reasonable case can be made for a lower percentage.
While the details of this can be debated, it is clear that crash rates
in general will start to drastically drop, well, before fully automatic cars
become the predominant vehicle on the roads. There will thus be a ‘tip-
ping point’ in crash rates in advanced economies about 2023–2026,
when there will be a sufficient proportion of 80% cars to ensure that
even badly driven manual cars will find it hard to create crashes. Once
50% or more of cars are autonomous and networked, then car crashes
will only occur due to software bugs or hacking or poor maintenance or
roadside objects or manual cars driven erratically.
This concept can be applied to other insurance types, though there
are less networking effects with personal insurance.

Telematics and Turnover Rates

The next issue to be examined is the rate that telematic devices and other
technologies can be introduced into the market. In this case, the rate that
80% crash avoidance technology will be introduced into the automobile
fleet. The faster the introduction of pre-installed telematics and the higher
the automobile turnover rate, the faster is the impact on insurer cash flow.
Some commentators argue that telematics with not be widely used
in automobiles because, based on current black-box telematics, they
are too expensive for all but the highest risk clients to install. This
ignores the fact that increasingly cars are being manufactured with quite
214    
M. Naylor

detailed telematics pre-installed. The increasing use of these devices

is leading to their cost to start to drop sharply. When Google first
built their autonomous car the LIDAR system cost more than
$70,000. Today, that same hardware can be purchased for $250 and is
expected to fall to $90 soon. Even low-end cars will soon have these
devices installed. In 2017, 83 million cars with smart functions will be
sold. Top-end cars are also increasingly being data networked, though
at this stage only within the manufacturer’s network. Note that, that all
cars produced since 2006 include an OBD chip, so for basic dynamic
insurance there is no need for expensive sensors.
Given that it will be low cost for insurers to access these networks
and gather data from pre-installed telematics, why wouldn’t they? This
linking to telematics is happening now in the commercial fleet, which is
where these techniques will be trialed.
The turnover rate (replacement of old cars by new ones) of automo-
biles between countries varies greatly. In the USA, it is approximately
11 years. In general, if we use existing turnover rates, then cars with
80% avoidance technology and telematics will not constitute over
50% of the vehicle fleet until after 2028. This has led to commen-
tators arguing that change will not be substantial until about 2035.
However, within an exponential change, a forecast based on a linear
projection of past trends is dangerous, so insurers are likely to be sur-
prised and thus wrong-footed. And, as noted above, the tipping point
for a rapid decrease in auto-insurer cash flow occurs at about 50%, or
around 2028, with sharply increased competition five years earlier.
It needs to be noted that the inclusion of telematics has increased
the cost of cars, especially exposed components like wing mirrors or
windscreens.2 The installation of these has become more complex and
expensive. This has led to a temporary rise in crash payouts. However,
by about 2020, the rate of crashes will decline faster than component
cost will rise.

2Windscreens increasingly carry many sensors and even small errors can results in substantial mis-

alignment in ADAS systems.

 7  The Dynamics of Decline    
215

For personal risk insurance, external telematic devices have a low

cost, and a group of customers have been voluntarily buying them.
Hospitals will soon start issuing these to high-risk patients, so they can
be monitored at home. There will be some delay for internal devices as
medical testing is required. This use by high-risk patients will enable
producers and insurers to gain experience and thus to lower costs and
create stable systems. Widespread introduction can thus be expected to
start after 2025.
For house and contents insurance, there is a first mover issue as cur-
rently homeowners have little incentive to pay more for smart connected
appliances, thus installation rates are low, so there are few incentives of
the insurer to invest in telematic-based service systems. Once one firm
moves in with an attractive offer, then homeowners can be expected to
switch rapidly, especially at the top end of the income scale. Insurers
who do not respond will be left with lower income clients.

Electric and Automatic-Drive Vehicles

A related issue linked to turnover is the rapidly decreasing drop in
the cost of electric vehicles (EV) against the cost of internal com-
bustion engine vehicles (ICE). Tesla has been an EV innovator,
and while still a niche player is raising genuine concern among the
incumbents. Realizing that the main cost was the battery pack, Tesla
has built an extremely large battery factory. Given that batteries are
about $12–14k of the cost of a $35k car, a halving of the battery cost
makes the car competitive. By the end of 2017, Tesla’s Gigafactory
will be approaching or below $100/kwh as production ramps up.
Continuing cost reduction will make EV cars cheaper than ICE cars
after 2020. EV range is also rising rapidly, with Tesla predicting a
driving range equal to ICE cars by 2020 and double the ICE range by
2028. Tesla’s success has motivated competitors to step up their game.
BYD, Foxconn, LG Chem, Nissan, Dyson, Apple, and Samsung have
all made announcements with plans to build large scale battery fac-
tories. BYD’s plans are the most aggressive - a goal to add 6GWh
216    
M. Naylor

of production per year, hitting 34GWh by 2020 (matching Tesla’s

35GWh capability). This manufacturing attention could accelerate
the cost curve downward even faster.
Car purchase costs are not the only factor, as most ICE cars cost
owners over $10k in annual running costs. Given that EVs have a
about 90% lower fuel and maintenance costs of ICE cars, with superior
performance, it can be assumed that once they can be purchased at a
similar price to ICE cars EV cars will probably be purchased as fast as
they can be produced. Current trends indicate that somewhere between
2020 and 2025 is the tipping point - the point at which it would make
no financial sense to purchase a new ICE vehicle for the average vehicle
buyer. These predictions do not lie in the success or otherwise of Tesla
to meet production targets. And, they also do not assume that indi-
vidual ownership will decrease. The prediction is based on the immatu-
rity of current EV manufacturing and the expected rapid drop in price
as technology improves. Fundamentally, EV are substantially easier to
manufacture than ICE cars. FoxConn, a contract IT manufacturer, esti-
mates that by 2025, it will be able to make family saloon EVs at a price
of $US15k.
This replacement of ICE cars by EV cars is a one-off disruption,
which will jump the turnover rate substantially within the next ten years.
The main limiting factor will be the speed at which incumbent manu-
facturers can respond by switching production to auto-drive EVs, at that
price range. Assuming all these cars will come equipped with an inbuilt
advanced sensor and auto-drive systems then the tipping point to auto-
insurance viability in cash flow terms could occur surprisingly soon.
Progress to date with 100% automatic-drive systems has been slow
due to physical issues with car telematics. These have been largely
solved, so delays from this year will be primarily related to AI experi-
ence. AI progress depends on accumulating sufficient training data.
Given that Tesla is collecting driving data from all its cars, Elon Musk
predicts that software issues should be primarily solved by 2020, for
areas without extreme winter conditions. Note that networked cars
are not stuck with the auto-driving ability possible at the date of their
manufacture, as software downloads can regularly improve that perfor-
mance. Within an EV future, cars become software devices on wheels.
 7  The Dynamics of Decline    
217

Given the exponential rates of software improvement, it would be a

brave insurance manager who would argue that auto-drive cars will not
be introduced rapidly. If the auto-drive software capacity doubles and
telematic cost halves every two years (say), then why would crash rates
and associated premiums not start to drop by 10–15% per year?
It needs to be remembered that any technology offering a substan-
tial reduction in the costs of car accidents is likely to be proactively
embraced by both the public and regulators. Europe is likely to ban the
use of non-auto and non-electric cars by 2030. Turnover rates are thus
likely to sharply rise.
It also needs to be remembered the change to safer telematic linked
EV cars will occur first for rich and upper middle-class clients. This
implies that insurers who do not provide a dynamic networked product
will be left competing for a shrinking pool of lower income clients driv-
ing old manual ICE models.
An increasing trend in large cities will be the use of car or ride shar-
ing Apps. Already car ownership rates have peaked in key demographic
urban categories. In particular, these groups will be less likely to buy
a 2nd or 3rd car. The net impact of this is still unknown, but since
profit margins on fleet insurance deals are low, the impact on premium
flow could be noticeable.

Rapid Innovation Switches

The time it takes for an external disruptor to enter is shrinking; it took
Uber only 4 years from its foundation in the USA to start to disrupt
the Australian market. Examples exist of unexpectedly rapid technol-
ogy switches. For example, urban transport in cities went from 99%
horses in 1900 to 95% cars by 1912. The key is the pace at which tech-
nology advances - cars were a new technology that was faster, cheaper,
longer living, and always followed directions. These benefits were so
overwhelming that a technique of travel that humans had been using
for millennia disappeared in a little over a decade. The lesson here is
that once a disruptive technology reaches a particular tipping point, it
doesn’t just take market share incrementally from the incumbent indus-
try but rather completely rapidly replaces it.
218    
M. Naylor

Another example is the German electricity sector. In 2010, the major

companies regarded their market as stable and secure. The threat of dis-
tributed solar power was regarded as a minor issue. By 2015, both EON
and RWE were in internal crisis and announced that they were spinning
off the fossil- and nuclear-dependent parts of their businesses so they
could focus on renewables and producing consumers, ‘prosumers.’
As an important enabler of innovation, insurers could evolve with
technology and even thrive. Some cars today can tell when you’re
drowsy, or if you don’t have a hand on the wheel, or if you’re speeding,
and so on. These are enabled by listening devices, biometrics, and tele-
matics in your car - monitoring you and your driving habits.
Given this flood of real-time data that cars can generate, the current
model of charging annual premiums seems so last century. Dynamic
insurance is the future. These are complex issues and all should be
debated by insurance strategists. However, I would argue that it is bet-
ter for insurers to be prepared too early than to be rushing to catch up.
Assuming no change is the risky choice. There is a lot of money cur-
rently going into InsurTech. Are auto-insurers willing to bet their firms
on the unlikeliness of disruption not occurring over the next 5–10 years?

Barriers to Entry
The insurance industry has traditionally had higher barriers to entry by
new firms that other industries do. These include financial, data, expe-
rience, reliability, switching costs, and regulatory barriers. The auto-
insurance market, because it is a temporary and simple product where
switching is easy, has substantially lower barriers than other insurance
sectors. This reinforces its role as the leading edge of reform.
Personal insurance is, by nature, a long-term contract. Once a cus-
tomer has taken out a life or health policy, then any health conditions
which subsequently occur, would be covered by their existing policy
but excluded if they switch. They are thus locked into their existing
supplier for life. Because of this, customers must trust their insurer to
continue to exist. Switching is also not straight forward as there may
be medical questions or tests required. Regulators thus require insurers
 7  The Dynamics of Decline    
219

to hold long-term capital and guarantee stability. Incumbents also

have a data and underwriting experience advantage. Because of these
hindrances emerging InsurTech firms are unlikely to poise a major
threat in the personal insurance market until they have grown to be
large, trustable companies in other sectors. Competition is more likely
from large firms in other sectors who add an InsurTech option to their
­existing product line.
These hindrances are not as important in the P&C insurance market.
Customers face few barriers to switching their car or house insurance,
and telematic users like Octo or Tesla have the data and experience
advantage. Bankruptcy of an insurer is, thus, not a serious problem for
customers, nor are regulators as strict. Dynamic insurance also offers
more in the P&C sector. For these reasons, auto-insurance will be
impacted by external disruptors before other sectors are. I would argue
that while InsurTech start-ups have entered via niche markets, there are
only limited barriers to their expansion outside those sectors, especially
if they have a substantially lower cost model.
The major barrier to entry to the new world of insurance will be
(i) an automated end-to-end processing system, (ii) expertise with AI
­analysis of real-time big data, (iii) a large relevant data set, (iv) an abil-
ity to transform as successive waves of innovation roll in, and (v) expe-
rience with engaging and delighting customers. Incumbent insurers do
not rank well in terms of these attributes. Large Internet platforms do.

Regulation
There has been commentary that the introduction of autonomous cars
will be held up by conservatism on the part of legislators in passing the
necessary laws. However, the relative safety of autonomous cars versus
the dangerousness of the current human drivers will provide enough
impulse to push through the required legal and cultural changes.
Worldwide more than 2 million people are involved in car crashes, with
a total cost exceeding US$212B. Each year in the USA, 30,000 people
die. If auto-drive cars reduced this by 90%, it would be brave legislators
who would oppose restrictions on the use of manual drive cars.
220    
M. Naylor

In fact, cities could well ban self-drive cars from the CBD, and coun-
tries could ban them from expressways, probably by 2025. There will
be a strong incentive for smaller countries or states to promote them-
selves as ‘autodrive’ friendly via enabling legislation, in an effort to
attract investment and workers into their jurisdiction. This is why the
change may occur first somewhere like Singapore, where regulation
can be trialed easier. Within fifteen to twenty years self-driving a car
could become a quaint sport like horse riding - something rich people
do on weekends on an enclosed track. After all, people still own and
run antique or classic cars for fun. They are just not used for normal
­commuting and do not constitute a large insurance sector.

Reference
Cognizant. (2017). The work ahead: Seven key trends shaping the future of work
in the insurance industry.
 8
The Response of Incumbents

Incumbents
As explained a perfect storm of technological disruption is about to hit
the insurance industry. A key question is thus - ‘are existing incumbents
able to respond fast enough to avoid being swept away?’ The main issue
for incumbents is that they are set up to run their existing activities as
efficiently as possible, as well as respond to any immediate, real, crisis.
They thus do not have the existing spare resources, the spare money, or
personal, or the intellectual flexibility, to respond to medium term pos-
sible crises and to remake their business while also running the existing
business. It’s like trying to rebuild a car while it’s driving at speed.
A company subject to disruption is hard to manage as the core
business is suffering decline yet still produces the bulk of revenue and
employs the bulk of staff, whereas the expanding business is a cash drain
and requires skills/culture which may be foreign to existing staff. This is
worse if the disruption sector is exotic to the company culture and dis-
ruptor is still minor so the threat is potential rather than actual.
Added to this is the fact that most possible threats don’t eventuate, so
that executives who focus on looming possibilities and divert scarce cash

© The Author(s) 2017 221

M. Naylor, Insurance Transformed, Palgrave Studies in Financial Services Technology,
DOI 10.1007/978-3-319-63835-5_8
222    
M. Naylor

and personnel to combat something which doesn’t eventuate, risk being

seen as over-reactive. Incumbents thus have to determine if a threat
on the horizon actually exists or is noise, and if it will involve a small
change, a major change, or a complete reimagining of the industry. Yet it
is impossible to determine if a looming threat is disruptive or hype until
it impacts on existing cash flow in a major way. Who could have imag-
ined in 1990 that the phone industry would destroy the film industry?
As explained earlier, an issue is that management response to disrup-
tive change cannot be the same as the response to incremental change.
Disruptive change will require transformation of processes and culture.
The strategy that got you to the top is probably not the strategy you’ll
need for the road ahead. This change has to be substantial, as there is a
big difference between ‘an insurance company which does digital’ and ‘a
digital company which does insurance.’ It is hard to switch from one to
the other.
The problem is that because of the nonlinear nature of disruptive
change, by the time cash flow is affected, it is probably too late for a
company to react. Late adoption is thus often not viable. Therefore,
businesses can either (i) adopt early and become part of the disruptors,
or (ii) downsize and try to survive with the old model, or (iii) shut up
shop, maybe by sale to the disruptor. There will often be instances when
no viable path of transformation exists for the incumbent.
It needs to be understood that an incumbent following best practice
for their current sector is not going to save the incumbent if a disrup-
tor uses a different business model. Incumbents may have to imagine
an alternative approach to their current business model and disrupt the
market themselves.

Stages of Disruption
Capgemini (2015) argues that there are three stages to technologi-
cal disruption by an external entrant. These are (i) onset - typically
within the first year and is marked by the entry of the external disruptor
(ii) spread - typically takes place with the first two or three years, when
the disruptor starts to grow in popularity, which leads to multiple me-too
 8  The Response of Incumbents    
223

entrants, and (iii) mainstream adoption - the disruption technology

achieves widespread acceptance, typically within four years from entry.
Capgemini’s analysis shows that only 26% of incumbents respond
to the disruptive entrant during the onset stage, with 35% responding
during the spread stage, and a worrying 38% responding during the
mainstream adoption stage. Needless to say, the substantial majority of
incumbents who went bankrupt were those who responded late.
Capgemini argues that there are five reasons why incumbents tend to
delay responding to disruptive entrants.

1. Slow Decision-making Cycle: Traditionally, companies have used an

annual planning and strategy cycle. Satell (2012) shows that the
technological cycles are becoming, on average, shorter than annual
and are becoming increasingly shorter as the IT revolution pro-
gresses. Companies are thus finding that their markets are becoming
disrupted before they have had a chance to analyze the change. An
example is Uber, which spread worldwide within 18 months of estab-
lishing itself, well before taxi companies could think about and react
to its entry.
2. Complacency about Existing-Business Model: Management typically
fails to sense the need for drastic change due to inertia. This occurs
because management tends to focus on improving the efficiency of
the existing successful business model. Their position in their market
sector during the onset stage is often unchanged and initially shows
no sign of strain. The disruptor is small and it is easy to ignore or
pretend it will come to nothing, especially as most new technologies
do fail. The disruptor often has an entirely new business model which
incumbents cannot visualize. Often the disruptor has not quite
worked out how to make a profit, so financially switching seems idi-
otic. Thus, very few companies ever question their business model
until it fails. Even technically savvy companies can fail to foresee
disruptive changes. For example, RIM/Blackberry ignored the tech-
nological challenge posed by the all screen iPhone as they remained
the market leader in their business market sector, failing to see that
Apple’s new approach was destroying that market sector. Because the
224    
M. Naylor

change is exponential, it is easy to initially assume the threat is over

hyped, and then be surprised when a tidal wave of change hits.
3. Fear of Cannibalizing Existing Business: Most companies’ initial con-
cern is that new business models will destroy existing cash flows. As
this is true, a choice has to be made between current profits which
may disappear and uncertain future profits. An example here is
Kodak who delayed investing in digital despite their initial tech-
nological lead, vs Fuji, who invested heavily in digital processing
machinery.
4. Lower Margins in the Transition: Digital businesses often have lower
margins than traditional businesses, so that moving into digital is
normally perceived as a risky bet on future revenues. An example is
newspapers which have been reluctant to move into digital since the
advertising revenue is a lot less. The heavy investment required in
new areas often means they run at a loss for a lengthy period.
5. Key Resources unaligned to Opportunities: The siloed nature of most
companies, where unit size equates to manager salary and prestige,
tends to mean that no manager wants to contribute cash or mate-
rial to a new unit. Managers thus tend to try to retrofit new units
into the existing structures, failing to realize the essential cultural
and business model changes required to make the new technology
successful.

The ideal response to disruption is to treat the new-area business with

a venture-capital style management, so that growth rather than cash
flow is focused on, and flexibility is stressed. The existing-business area
should have private-equity style management, with a focus on cash flow
and efficiency. An incumbent can then disrupt itself, using the new area
to destroy the existing area.
Unfortunately, what tends to happen in practice is that the new area
is treated with suspicion and endures a high level of scrutiny, particu-
larly if cash flow estimates are not meet, while the existing business is
treated leniently. Existing-business managers have been right in the past
and events will show them to be right until they are wrong, whereas
new-area managers have no history of success to build on. The skill set
which made existing-business managers successful is unlikely to be the
 8  The Response of Incumbents    
225

skill set which works within the new area, so existing-business managers
may not recognize or value the new-area manager skills. This tends to
delay a successful response to external entry until the threat to cash flow
realizes, and then the transition has to occur within an environment of
rapidly declining existing-business cash flow.
It needs to be remembered that in the initial stages of the revolution,
the external disruptor will only impact on the poorer quality firms, so
that the better firms can tell themselves that they will survive intact. As
long as you’re not the weakest firm, you may not be under imminent
threat. An analogy here would be a herd of wildebeest being chased by
lions. An individual wildebeest can tell itself that as long as it runs fast
enough to stay in front of the slowest runners, it won’t be eaten. What
it fails to recognize is that as the back wildebeest are eaten, the lions
are getting larger and faster and its own turn will come sooner than it
thinks. The best survival is to go early, learn to run faster, and try to
transform into a lion.
This relationship between the old and the new business areas is made
even more difficult because the main target the new area wants to dis-
rupt will be the old area. Basically, customers have to be migrated to the
new area as fast as possible, even though profit margins in the new area
may be initially far lower than profit margins in the old. The impact of
this on the state of the accounts can cause huge internal management
disruption and needs strong leadership to ride out.
Capgemini (2015) argues that there are four ways of successfully
responding to disruptive entrants:

a. Acquiring digital talent: hire appropriately skilled staff. This was used
by 48% of successful responders.
b. Mimicking: deploying significant resources to create similar products
and business models. Used by 32% of successful responders.
c. Acquisitions: purchasing new entrants and then using them. Used by
36% of successful responders. Note that the incumbent has to use
the acquired firm to change existing core business culture or the
acquired firm will ossify and employees will leave. New competitors
will arise.
226    
M. Naylor

d. Judicial/regulatory: new entrants often have new business models

which are not covered by regulation, so incumbents raise legal or
regulatory concerns. Note that this can only delay disruption and
can slow down incumbent adoption until change occurs with a tidal
rush. Used by 32% of successful responders to give them time to
react.

Note that these methods can be used together and are the only effective
long term if they aid the incumbent to transform its business model.
Ultimately, successful responders need to embed a culture of open dig-
ital innovation and customer centricity within a new business model.
Management needs to be constantly looking for possible new disruptive
elements at an early stage and prepare for them. The pace of technologi-
cal disruptions will increase in all sectors and will increase at an increas-
ing rate.
Business as usual is not an option. Existing business units will have
to be evaluated and ones which are not useful for the future need to
be sold-off while they still have value. McKinsey (2016a) argues that
the senior management and board need to step back and ask funda-
mental questions—‘do the businesses we own make sense in the new
world which is unfolding, and are they going to create value, and are
we the best owners of those businesses to maximize value, and what
else could we be investing in, what’s the bar on the capacities which we
have, and how much has it gone up, and are we up to scratch or are we
going to end up being out-competed by someone else who’s just recon-
ceived what’s possible?’ Given that during a disruptive transformation,
the future will be unusually hazy; this strategic thinking has to be done
via loose scenarios, rather than precise spreadsheets. The board needs to
focus on lost opportunities in the new area, rather than threat to reve-
nue in the old area. The culture needs to change from a ‘Fear of Making
Mistakes ’ mind-set to a ‘Fear of Missing Out ’ mind-set.
The role of the CIO is vital and probably too important to be left to
existing CIOs. This is because traditionally CIOs have been routine and
reactive, focused on efficiency and trouble shooting. What is needed for
IT transformation is a person who can combine IT knowledge with a
strategic overview and focus on innovative and proactive initiatives. In
 8  The Response of Incumbents    
227

particular, the job of establishing IT as the core element of an insurer

will need close integration of IT development with all sectors of the
firm and with customer groups, with a special focus on usability and
visualization appeal.
Cultural issues at board level are a problem. This is because the exist-
ing board was chosen for their skills at evaluating and running the exist-
ing business with its well-understood model and stable cash flow. The
new-area business will probably require a different set of skills from
board members, more attune to those of a venture capitalist board. It
takes a courageous board to approve management investing heavily in
terms of personal and money in ambiguous projects in areas where the
outcome is uncertain, cash flow is unstable, and the pathway to success
subjects to unexpected shifts. For this reason, survivors are more likely
to be incumbents who were in trouble and therefore had a board pre-
pared to gamble, than the best-run and most-profitable incumbent.
Increasing investment in the new area and declining cash flow from
the old area mean that profits will probably drastically shrink and the
fall in dividends may trash the stock price. Stock investors who liked
the old cash cow model will become disgruntled. The board may be
tempted to underemphasis the financial pain necessary during the tran-
sition phase, which only leads to a larger crisis later when the scale of
problems is realized. Survival may thus require a different class of
investors.
An additional problem is that boards often have a ‘group-think’
and thus may fail to see problems which would be obvious to a per-
son from an external sector. Thus, a shake-up in board membership may
be an essential part of survival. The introduction of experts in disrup-
tive transformation and in the core skills of the new area is required,
as well as a deliberate increase in the diversity of board member world-
views. Disruptive thought leaders should be prized, yet they are usually
not selected because their presence makes other board members uneasy.
Strong and decisive leadership at the top is essential. New board mem-
bers should be chosen because they make current members uneasy.
228    
M. Naylor

Oligopolies and Disruptive Entry

A lot of the description in this book, about what will be required in
the immediate future from insurers in terms of new patterns of manage-
ment behavior, will strike many currently in the insurance industry as
fantastic speculation - they will respond that the described changes seen
as essential will never occur because they are so far away from current
modes of behavior. Some participants will, however, not find it unrealis-
tic as they will either already be starting to engage in these behaviors or
will have already contemplated these issues. Readers in disrupting sec-
tors may regard my speculation as trivial and old-dated. This contrast
in reactions is not unexpected as such dichotomy is the norm during
revolutionary upheavals.
The unique convergence of the technical innovations outlined in an
earlier chapter will create an unprecedented period of change in service
administration across all industries, a period when there will not only
be a major change but change will occur at an increasing rate; exponen-
tial change. As outlined in figure 1.1, the human mind cannot generally
cope with this kind of change, initially being disappointed that change
is occurring at a slower than expected change, before being shocked by
the unexpected rapid change. The natural reaction is to initially dismiss
exponential change as ridiculous, then to fight it, before reluctantly and
belatedly rushing to catch up, and lastly accepting the change as nor-
mal, and finally wondering what the fuss was about.
The pace of change in the insurance sector will probably be faster
than most participants expect. Internet giants and platforms suppliers
are already starting to move into the insurance sector, and their data
analytical skills should make them strong competitors. Insurers are
starting to be forced to learn big data and social network skills. The base
concept is that these two sectors, Internet technology and insurance,
are starting to converge in terms of the fundamental skills required.
Insurance has two basic skill sets: (i) data analytics and (ii) customer
relationships. These are also the core skill sets of many IT sector firms,
and in general, Internet technology firms are generally more efficient at
these two skills. Insurers are definitely playing catch up.
 8  The Response of Incumbents    
229

Customer reluctance to use out-of-sector firms is also dropping.

Capgemini/Efma (2016) shows that one-quarter of Generation Y cus-
tomers and nearly one-half of Latin American and Asian-Pacific cus-
tomers say they are very likely to purchase insurance from a technology
company.
It also cannot be understated that the scale and speed of the com-
bined changes amount to a revolution in the way service administra-
tion is carried out. It also needs to be repeated that there is no visible
end to the process of change, in that it will not be a series of discrete
changes with a defined end date, a one-off change. It instead seems
to be a permanent increase in the rate of change, a new norm which
businesses will have to adjust to. Insurers cannot thus just visualize
the change as one of moving from their old arrangement to a new sta-
ble arrangement - they will have to reconfigure themselves to accept
­permanent change.
All service sectors will be impacted by these changes but insurance
will be impacted more than most service sectors because (i) other sec-
tors, like banking, have made more adjustments already, whereas insur-
ance has stayed fairly traditional; (ii) insurance depends more deeply
on skills in data analysis than most sectors (iii) insurance is inherently
relationship and trust based; and (iv) insurance in most markets tends
to be traditionally less competitive. The finance sector, for example, has
moved away humans trading; it has extensively disintermediated and
has coped with extensive changes in company ownership and structure.
Insurance has done none of these.
How quickly the insurance sector in each market reacts to the grow-
ing pressures to evolve will depend on local market structure. Many
markets are oligopolies with a small number of firms dominating. These
markets are not subject to intense competitive pressures, with market
participants able to reach de facto understandings to only engage in cer-
tain types of competition. Innovation thus has tended to be limited.
In many markets, this has limited the introduction into the insur-
ance sector of a range of technological changes which have impacted on
other service sectors. It is thus easy for these insurers to conclude that
they will remain safe from the onslaught of technological innovations
discussed. The reverse is true, however, as escape is impossible from a
230    
M. Naylor

Potential efficiency gap Market entry

Cost Possible Cost Efficiency

Cost of disruptor Gap B
Gap B entry

Actual
Entry point Gap A

Time Time

Fig. 8.1  Market Entry Dynamics

wave of innovations which are of a type which effect a broad range of

administrative processes, as well as a range of customer behavior pat-
terns, especially when these innovations both cut internal costs by a
substantial magnitude and intensify customer acquisition and retention
processes.
What the differing types of market structure will affect is the type
of disruption experienced. Within competitive markets, firms are used
to responding to changes in the business environment and thus will
be prepared to ditch existing frameworks faster. This higher efficiency
reduces the scope for an external disruptor to enter. Within less com-
petitive markets the cozier an oligopoly is, then the slower the rate of
internal sector change will be, thus the more the sector will fall behind
the general service sector norm, and therefore the more attractive the
sector becomes to an external disruptor and the less that existing firms
will be able to compete with that external entrant. Thus, in competitive
markets, change will tend to come by competition forcing existing firms
to adapt to the required changes, whereas in less competitive markets,
change will tend to come by the entry of external firms who are strong
in the new skill sets and the subsequent disappearance of existing firms.
It is up to insurers to decide which fate is theirs.
The basic concept behind disruptive vs evolutionary change is illus-
trated in Fig. 8.1. There is a cost to entry imposed on a non-sector
outsider, PV amount A. The dynamic changes discussed above mean
that this cost of entry will probably decrease over time. There is a gap
 8  The Response of Incumbents    
231

between efficiency/profit generated by best practice and efficiency/profit

generated by existing firms, PV amount B. This increases over time if
existing firms do not innovate. As soon as amount A is less than amount
B, it is profitable for an external firm to enter and disrupt the market.
This is the ‘Uber moment.’ As time passes, the gap gets larger so the
chance of change by disruptive entry rises.
It is useful to note that the rate of disruptive entry is increasing in all
industries and the rate of established company bankruptcy increasing, as
shown by the five-year mortality rate for US public companies now run-
ning as high as 33% overall. Even for the largest established firms, the
rate is around 20%. In 1958, S & P500 companies, on a rolling seven-
year basis, remained in the index for 61 years, but by 1980, it was down
to 25 years and, by 2011, was down to 18 years. From 2000 to 2015
52% of Fortune 500 either went bankrupt, were acquired, or ceased to
exist. Realistically, it can be expected that over half of all currently listed
insurance companies will not be listed by 2027, due to bankruptcy or
takeover.
It is vital to realize that disruptive entry does not have to involve one
major external disruptor. It can equally involve many small firms attack-
ing the more profitable parts of a value chain and forcing incumbents
to disintermediate. This is particularly true in smaller markets. The dis-
ruptor also does not have to be the major world leader, as once a busi-
ness model is established, third-party innovators can establish in smaller
markets before the international giant has a chance to expand there.
The insurance sector is also highly prone to disruptive entry as their
return on equity has been persistently below the average with the US
property and casualty insurance sector falling below the average more
than 85% of periods during 1955–2010. They will thus struggle
to obtain the extra equity they require to make the large investments
needed to support the large increase in innovation needed.
232    
M. Naylor

The Economics of Digital Disruption

McKinsey (2016b) argues that the task of trying to prepare for future
innovations is very complex as companies not only have to examine
current trends, current industry competitors, and possible external dis-
ruptors, companies also have to predict trends which have yet to occur.
This involves understanding ‘what-we-do-not-yet-know’. This is vital as
the ultimate aim of insurance survivors should not just be to survive
against waves of disruptors, it should instead be to get on top of yet-to-
occur trends so the incumbent insurers can become the disruptors. This
requires an understanding of the deeper forces, that is, not which com-
panies or technologies will disrupt, but why the disruption will occur.
For an external disruption to be successful, there needs to be a ‘dis-
ruption point’; some aspect of the value chain which the disruptor can
carry out substantially better than incumbents can. Some disruption
points are

1. Unused inventory: Amazon, Uber, and Airbnb have thrived because

they are better able to organize inventory, which existing suppliers
left unused.
2. Information aggregation: Disruptors can organize production or client
information better than incumbents
3. Reputation: Incumbents have low reputation and trust levels.
4. Cost-plus pricing: Incumbents are using cost-plus pricing. An example
is the way Space-X is transforming its industry by reducing cost to
less than 10% of the hide-bound incumbents
5. Low marginal cost: Disruptors can automate processes so each activ-
ity has an extremely low marginal cost, thus enabling them to both
undercut the incumbent and to offer customization.

Incumbents need to decide in which area of their value chain they are
vulnerable and start to disrupt their own model.
McKinsey (2016b) argues that there are two primary sources of digi-
tal disruption, each with three sub-elements:
 8  The Response of Incumbents    
233

1. Modest changes: These tend to cause smaller changes in market

structure.
a. Unmet demand caused by changes in customer expectations.
Changes in technology reveal that customer expectations are not
what incumbent companies thought they were. For example,
music was delivered as albums as this was the most cost-effective
supply mechanism. Digital downloads revealed that customers
actually prefer to buy single tracks. Do customers want to own
cars or do they merely want to use private transport services? Do
customers want insurance on a permanent annual basis, or only
when they engage in particular activities? Companies are vulner-
able to disruption if some customers subsidize others, or if cus-
tomers have to buy a whole product to get the part they want, or
if customers can’t get what they want when they want it, or if the
customer experience doesn’t match the global best standard.
b. Exposed new supply because digitization allows new sources to enter
product or labor markets. For example, Airbnb unlocks the sup-
ply of previously unrentable lodging. Companies are vulnerable
to disruption if customers only use a product partly, production
is inelastic to price, supply exists which is currently unusable, sup-
ply can be utilized in an unpredictable new way, or fixed costs
are high. An example here would be Web-based peer-to-peer
insurance.
c. Market makers enter to match unused supply to latent demand.
For example, Wikipedia allows people to learn about a topic with-
out buying a large encyclopedia set. Companies are vulnerable if
there are; high information asymmetry costs, or high search costs,
intermediaries who impose fees, or long lead times to complete
transactions.
2. Extreme changes: These tend to cause substantial market disruption
and normally involve external disruptors using a different business
model.

d. A new value proposition is created due to customers’ expecta-

tions evolving past what present suppliers can deliver. This often
234    
M. Naylor

involves new suppliers providing what customers weren’t aware

they needed. For example, the ability to search the Web from a
pocket phone. Customers and thus producers were not aware that
they wanted phones to do anything apart from talking to other
people. Companies are vulnerable if increased information would
greatly enhance your product, your physical product is not con-
nected, there is a significant lag between when customers purchase
your product and when they receive it or the customer has to go
and get the product.
e. A reimagined business system is offered by new suppliers.
Incumbents often focus on perfecting their supply chains, so are
stunned when an external supplier introduces a completely new
way to make money. For example, Hard-drive manufacturers
focused on efficiency before Amazon came and transformed data
storage from a product to a service. These new approaches often
transform the scalability of cost structures, driving marginal cost
toward zero, thus flattening the supply curve. Companies are
vulnerable if there are redundant value chain activities, or well-
entrenched physical distribution networks, or higher than aver-
age distribution costs. Do customers want compensation after an
adverse event or do they want help to ensure that event does not
occur?
f. New suppliers transform the industry by scaling the quantity and
breadth of production up by previously unimagined multiples and
thereby blurring industry boundaries. For example, Amazon intro-
duced the Kindle mainly to sell digital books rather than make
money from physical sales. This allowed them to move onto the
sale of services. This disadvantaged Sony, who had to make money
from sales of Kobe. These disruptions normally involve the crea-
tion of a digital platform, which binds chosen suppliers into an
ecosystem dominated by the platform supplier. Within the car
industry, the company which is best placed to offer car insurance
is the player with the best data, which is likely to be the company
which supplies and operates the telematics platform.
 8  The Response of Incumbents    
235

Detectable Clear Inevitable New Normal

Disruption is .. Faint signals Emergence Critical mass At scale
with lots of noise of a validated of adoption and mature
model achieved
New Business
Model
Profit

Incumbent
Business
Model

© McKinsey 2016

Time
Incumbent’s move Acuity Action Acceleration Adaption
.
Common Barrier Myopia Avoidance Inertia Fit
of pain

Fig. 8.2  Stages of Digital Disruption. Source McKinsey (2016a)

Stages of Disruption
McKinsey (2016a) argues that there are four stages to digital disruption.
These can be graphed by recognizing that all businesses involve technol-
ogy which grows matures and then ages over time. This can be visual-
ized as an S-shaped curve. The only question is - Are you on the old
or new technology curve? McKinsey argues that we can visualize dis-
ruption by overlapping the old and new technology curves, shown in
Fig. 8.2.

• Stage One: Signals amidst the Noise - In the first stage, the new tech-
nology is just evolving. A clear-sighted incumbent could spot the
possible change and react to it early. However, because the environ-
ment is so noisily, which hundreds of possible threats, it’s nearly
impossible for an incumbent at the time to decipher which tech-
nology will thrive from the hundreds which will fail. Incumbents,
236    
M. Naylor

however, need to be aware of the likelihood of some technologies

being successful and being prepared to react swiftly. Scenarios based
on possible disruptive challengers will often highlight vulnerabilities.
Those incumbents who do spot possible threats and prepare to react
are well placed to survive.
• Stage Two: Change takes Hold - In the second stage, the trend is clear.
The incumbent needs to commit to nurturing multiple new incen-
tives. They need to invest in the most likely new directions and be
prepared to cannibalize existing businesses. This is difficult as the
threat is still distant and may not seem important to stockholders.
For example, when Netflix moved from DVDs to Streaming, its
stock price fell by 80% because of a sharp drop in current cash flow.
Most incumbents thus dabble, making only small investments, and
focus on protecting the existing business.
• Stage Three: The Inevitable Transformation - In the third stage, the
future has arrived and new technology is rapidly gaining market
share. Incumbents have to massively shift resources to the new area.
This can be difficult if they do not already have experience in man-
aging the new area. It is made even worse because old-business cash
flow will be dropping, so firms focus on cutting back rather than on
making the large investments required. Top management and boards,
who are normally sourced from the old business, often view the crisis
as temporary; one they can live through if they focus on old-business
efficiency.
• Stage Four: Adapting to the New Norm - In the fourth stage, the new
technology is dominant, so incumbents have become disruptors or
have to focus on the best way to exit. Often all that is left of value is
the brand name and legacy.

Digital Strategy’s Double Game

BCG (2015) argues that because of the exponential possibilities and
economics of new technologies, companies need to create defined
‘digital strategies’ to enable them to continually adapt to and seize new
opportunities. This means playing a ‘double game’: Making the most
 8  The Response of Incumbents    
237

out of today’s contests, while positioning themselves to win in tomor-

row’s contest. They have to rebuild their new business while running
their old one.
BCG argues that companies have to ask three strategy questions:

i. What can I forecast? Companies need to analyze what is possible in

the future; explicitly stating the most likely scenarios and conclu-
sions, but also running scenarios about any possible outcomes, even
if very unlikely. This reduces surprises about future and prepares
companies to act on any outcomes with decisive speed.
For example, automotive companies have been contemplating a
future with autonomous vehicles for decades. But because their busi-
ness focus was on efficiency, global expansion, and electric cars, they
were surprised by the rapid emergence of an autonomous-vehicle
prototype by outsiders, Google and Tesla, and the broad embrace of
the concept that they generated. The industry is thus urgently play-
ing catch-up.
ii. Where can I be disrupted? Companies need to examine all areas of
the supply chain as well as customer requirement. It is because an
external disruptor has a very different perspective on customer needs
that they are successful. Who knew that phone customers wanted to
play games? For example, buyers of cars do not want cars, they want
mobility, which is not the same thing. Car insurance customers do
not want 24/7 insurance, they want insurance only when they drive.
Would customers prefer to buy PAYG insurance from their mobile
data provider? Thus, companies need to explore the attacker’s per-
spective. They need to examine adjacent industries for new moves
which could endanger their current business. They need to be aware
of the activities of disruptors in all sectors. They need to actively
imagine multiple possible futures.
iii. What can I shape and where do I need to adapt? Existing supplier
advantages, brands, distribution networks, supplier relationships,
strong capabilities, and superior cost positions, etc., do not disap-
pear in a digital world. Companies need to examine how they can
draw on existing or latent strengths to acquire new capabilities or
adapt to ambient forces. BCG cautions, however, that CEOs often
238    
M. Naylor

Value chain Ecosystem

Reengineer the Value Chain

Transformation Transformation

Exploration

Enhancement
Exploration
Transformation

Offering
Enhancement Transformation

@ BCG Analysis Reimagine the Offering

Fig. 8.3  The Digital matrix. © Boston Consulting Group, used with permission.
Source BCG (2015)

overestimate their ability to predict and shape their environment,

underestimate the power of disruptive trends, and consequently lead
their companies astray.

BCG (2015) argues that charting the current and future playing fields
is necessary but insufficient. It’s also essential to frame, explore, and pri-
oritize strategic choices. An intuitive tool is BCG’s digital opportunity
matrix, shown in Fig. 8.3, which comprises two axes: (i) Reengineer the
value chain - state-of-the-art IT has become flexible, intuitive, power-
ful, and accessible. Data analytics that improves sales effectiveness across
physical, mobile, and online channels can promote significant value cre-
ation. (ii) Reimagine the offering - whereas reengineering is largely a lin-
ear process in search of efficiency and effectiveness, reimagining is more
open ended, requiring creativity and vision. Digital also creates ample
opportunities for novel products and services. These innovations typi-
cally exploit new data and powerful analytics.
The BCG diagram has three stages that are represented along the
diagonal in the digital opportunity matrix.
 8  The Response of Incumbents    
239

1. Enhancement: Strategically, enhancement is about extrapolating from

your current position. Start with where you can create immediate
value. While this is the least radical stage of digital opportunities, it
can improve the organization’s digital skills and provide tremendous
and immediate value creation that can fund the broader digital jour-
ney. Examples include predictive maintenance, streamlined digital
links to suppliers and customers, and recommendation engines.
2. Exploration: Exploration requires investigating offerings adja-
cent to the current business or pursuing larger adjustments of the
value chain. Exploratory digital strategies become executive topics.
Companies need to invest significant resources in digital businesses
and closely track their performance. In many cases, companies can
develop a portfolio of venture investments to establish beachheads in
promising areas.
3. Transformation: Transformation is an all-encompassing strategic move
that has the greatest potential to generate competitive advantage,
often over several years, but also the greatest risk. From a strategic
standpoint, executives need to reimagine a target business model
5–10 years out and then ‘retropolate’ from that vision back to the
present. Transformative change by its nature becomes the top CEO
priority. Transformation requires major investments and often the
development of new partner ecosystems.

BCG (2015) argues that three conclusions arise from this: (i) Ultimately,
strategy is about choice. Companies cannot do everything, and they can-
not even do all the things they should do simultaneously. Strategic
choices are required to prioritize and stage initiatives. (ii) Conventional
wisdom can be generally right but specifically wrong. In a fast-moving
environment, even winners stay vulnerable. Data in particular is a
problem because it is new and growing exponentially. The most valu-
able data doesn’t rest in dusty databases but has yet to be created. (iii)
Practice playing the double game. Sustained success requires actively man-
aging a portfolio of initiatives across time, running the old and new
side-by-side.
Companies find implementing these strategies hard. Managers are
increasingly nervous about the lack of progress in their digital initiatives
240    
M. Naylor

Double
Strategy Extrapolate Game Retropolate

Customer Customer Customer Customer Customer

pain points unmet needs journey engagement
Experience
Digitally
Offerings and enhanced
Data and Data-driven Software
information
Business Model products products
services products

Business Sales, channels R&D and After-sales HR, financial

and functional
Processes and marketing operations service
support

Agile Digital partners Systems and Analytics

Capabilities ecosystem technology and data
methodologies platforms integration
Transformation Lighthouses
Digital Start-up incubation Digital
talent
Accelerators and prototypes acquisition
venture funding, change
and M&A management

© BCG anaylsis

Fig. 8.4  The Action Space for Digital Transformation. © Boston Consulting Group,
used with permission. Source BCG (2015)

and consequently merely add ‘digital pixie dust’ to traditional processes

or engage in a frenzy of digital experiments and ventures, designed to
look good rather than being effective. Rather than drive competitive
advantage, these efforts leave companies more vulnerable.
Insurance customers do not want an insurance product, they want
an insurance service. Insurers need to reimagine their entire business
model.
BCG (2015) argues that companies need to operate in an action
space consisting of six interdependent layers as shown in Fig. 8.4. A
complete digital transformation requires playing in all six layers.
These layers are:

• Defining the strategy focuses on building competitive advantage in the

double game. It provides insight into short-term ‘no regrets’ moves as
well as more-transformative plays. These decisions establish the path
for the other five layers.
• Reshaping the customer experience explores how to eliminate pain
points and how to surprise customers with new levels and forms of
service today, while achieving quantum improvements in customer
experience tomorrow.
 8  The Response of Incumbents    
241

• Reimagining offerings and business models prepares companies to create

novel products and services, often by exploiting new data and power-
ful analytics.
• Reengineering business processes entails adopting flexible and intuitive
digital technologies to simplify processes and increase efficiency.
• Building capabilities, often by working with outside partners and
creating new platforms, enables companies to develop new ways of
working, new business models, and other building blocks of digital
transformation.
• Accelerating the transformation involves devising new approaches to
speed up learning, ramp-ups, and transformational plays.

BCG (2015) argues that delivering a superior customer experience is a

natural initial goal of many digital programs. New offerings designed
to retain or increase the share of wallet of current customers and to win
new customers are at the heart of many digital programs. Allowing cus-
tomer to interact with quality Web systems can enhance their experi-
ence and transform neutral customers to enthusiastic advocates. While
quality is vital, however, speed is more important. A common mistake
is letting perfection stand in the way of progress. It is more important
to launch minimally viable offerings, even if faulty, and then to upgrade
them quickly in response to customer feedback than to let time and
opportunity slip away. Customers have proven to be highly collabora-
tive - accepting initial imperfection and suggesting practical product
improvements - when their feedback is acted upon.
Speed and flexibility are more important than perfection. In today’s
world, it is more important to be wrong fast than right too slowly. It
can also be useful to set impossible targets and deadlines, even if these
are not met, as this focuses executive minds on how to get there, rather
than on the precise details of the journey. Strategies do not have to be
precisely mapped out, designations can be fuzzy, methods can be flex-
ible, and targets can be missed - what is important is momentum.
Business processes need to be reengineered. Rather than thinking
creatively about how digital technology can support novel approaches
to work or production, many companies simply digitize the legacy
processes. Companies also need to quickly find or train candidates
242    
M. Naylor

to fill new job profiles. Without active management of workforce

issues, fundamental reengineering will almost certainly stall or even
fail. Experienced digital transformation managers are particularly hard
to find. Very few executives, for example, have industry experience and
deep knowledge in prototyping and agile methodologies.
Playing the double game in this area is challenging because it requires
simultaneous action to maintain legacy technologies, embed new digital
tools, and introduce a flexible and scalable new digital architecture that
will pave the way for an even more radically redefined future. IT depart-
ments and external partners can be critical resources in this undertak-
ing, or they can be a hindrance if their specific skill profiles, business
practices, and interests do not sync with what is needed - so IT pro-
cesses will probably need transformation.
Companies are adopting various practices to accelerate their digital
transformation, including rapid prototyping, incubators, M&A, and
the acquisition of digital talent, as well as traditional change manage-
ment levers. Many of the most effective accelerators call for new ways
of working and new forms of behavior, such as cross-functional, self-
directed, teams and collaboration built around agile software develop-
ment methodologies.

Innovation Traps
Kanter (2006) argues that faced with disruptive changes, firms fall
repeatedly into the same innovation traps. She identifies four waves of
competitive changes in the last 25 years: (i) the dawn of IT in the late
1970s when PCs started to take over from mainframes, Japanese style
total quality management became the fad, and new industries were
established. Firms responded by creating product innovation ‘garages’
modeled on firms like Apple. (ii) The takeover scare of the late 1980s,
when financial innovation gave corporate predators more power and
state-owned enterprises were privatized. (iii) The Digital Mania of the
1990s, based on the threat of firms associated with the newly created
Internet. (iv) The mid-2000 consolidation following the dot.com crash.
 8  The Response of Incumbents    
243

Kanter (2006) argues that each wave has started with enthusiasm
but has been derailed by similar problems. The main tension is between
protecting revenue streams from existing businesses and supporting new
concepts which could potentially dominate future revenue streams. This
tension creates internal resistance to the new innovations which tends
to destroy most renewal programs. Kanter notes that this means major
innovations tend to occur due to external disruptors.
Kanter observes a repeated pattern of mistakes in four areas:

Mistakes in Strategy
1. Executives declare that they want innovation but are refuse to sup-
port new programs unless other incumbent firms are doing some-
thing similar.
2. Executives tend to focus on opportunities for large gains, even if
risky, rather than opportunities for smaller, more certain gains.
3. Executives screen out opportunities which do not promise a rev-
enue gain within two years, even if it has the potential to disrupt the
industry.
4. Executives tend to focus on new products instead of realizing that
innovations can be as important in areas like production or adminis-
tration or transport or CRM.
5. Executives impose too tight controls.

In the 1980s, Procter and Gamble failed to develop a new type of toi-
let bowl cleaner as analysis showed that returns were uncertain and
didn’t fit traditional market research. They thus lost substantial market
share to a rival. In the 1990s, Quaker Oats focused too much on minor
tweaks to existing products and thus missed nearly all the more health-
focused innovations.

Mistakes in Process
1. Executives strangle new innovations by applying traditional planning
and budgeting controls to new products.
244    
M. Naylor

2. Executives apply existing metrics and performance reviews to innova-

tions. They want innovation managers to stick to plans, even if inno-
vation has gone in a different direction.
3. Employees tend to be rewarded for doing what they are committed
to do, rather than acting as changed circumstances would suggest.

In the 1990s, Ocean Spray grabbed the radical paper-bottle market

because its larger competitors were not prepared to fund uncertain
innovations outside their budget cycle.

Mistakes in Structure
1. Executives are unsure of how closely to link the innovator unit to the
mainstream, making it either too loose, thereby generating misunder-
standing and resentment in the mainstream, or too tight, thereby sti-
fling the innovation team with bureaucracy.
2. Executives can create two classes of conflicting corporate citizens:
those who have fun innovating vs those who make the money. This
conflict is worsened if the innovation unit has a markedly different
culture and type of employee.

In the mid-1990s, AT&T worldwide refused in invest sufficiently in

an Internet service provider unit because existing managers were reluc-
tant to provide funds or personal, as they were convinced it would
never generate significant revenue. Once it was a going concern, it was
brought back into the mainstream budgeting process and starved of
funds.

Mistakes in Skills
1. Executives tend to undervalue and underinvest in skills relevant to
innovation. They tend to put the best technical person in charge of
innovation units and not the best leader.
2. Technical executives tend to undervalue the importance of commu-
nicating results to non-innovation units, tend to empathize tasks over
relationships, and do not build team interpersonal skills. Innovators
tend to like to work in isolation and undervalue the need to get
 8  The Response of Incumbents    
245

corporate-wide buy-in, and the need to present the radical in familiar

terms.
3. Executives tend to rotate staff too quickly. Research shows that
innovation staff are more productive after two years, yet tend to be
rotated every 18 months.

In the early 2000s, Timberland developed an innovative new line of

modular travel shoes within a separate innovation unit. However, exist-
ing units had not been included and the sales force refused to promote
the product.
Kanter argues that there are thus four remedies to create successful
innovation:

1. Strategy Remedy - Firms need to widen the search for new ideas and
broaden the scope. They should create a pyramid of innovation,
topped by a couple of long-shot ground changing ideas, but cascad-
ing down to multiple smaller ideas. Company-wide calls should be
made for new ideas in all areas, not just products. While dedicated
groups pursue big projects, temporary groups can pursue smaller
ones. This allows a lot more employees to be involved and embeds a
culture of innovation and dynamic change.
2. Process Remedy - Flexibility is required in planning and budget cycles.
Special funds should be set aside for use if ideas arise mid-cycle.
Budgets and plans need to be changeable if innovations take new
directions. A cross-department team of flexible thought leaders needs
to allow exemptions from normal processes.
3. Structure Remedy - Links between innovators and mainstream units
need to be tightened, especially human links. Productive conversa-
tions have to be created between innovators and mainstream units,
with senior leaders actively encouraging mutual respect. Innovators
need to understand that they are there to serve the mainstream busi-
ness, and the mainstream units need to understand that innovators
are securing their future. Tesla insists its research engineers’ work in
the middle of the production floor.
4. Skills Remedy - Firms need to emphasize that the most important
company asset is the level of skills within its workforce. Continual
246    
M. Naylor

retraining will be required as activities are lost to software. Employees

need to feel part of an ongoing team.

New Entrants/Innovators
Current indications are that the majority of existing insurers will be
unable to make the required transition successfully against competi-
tion from outside firms, due to their inability to break from traditional
modes of behavior.
An example of this inability to break traditional bounds is Google’s
failure to find an insurer prepared, or even able, to underwrite road tests
for its self-drive cars, despite this being an ideal opportunity to gather
innovative data. The essence of accurate risk pricing is data, and the
increase in real-time driving data collection being generated by sensors
currently embedded in modern cars means that within a few years, car
manufacturers will have a better source of driving data than insurers and
will thus be able to price risk more accurately. Manufacturers will also
be able to establish real-time proactive customer contacts. Why would
insurers not want that data experience even if it is money losing?
Once real-time feedback is perfected, insurers will be in the position
of needing to work with car manufacturers to survive, yet car manufac-
turers will by then have the capacity to establish themselves as the plat-
form at the center of business ecosystem, with its profitable value-adds.
This relegates insurers to the margins, as suppliers of low-value prod-
ucts with little direct customer contacts. Insurer cash flow will drop and
car manufacturers will be able to play insurers off against each other.
Insurers need to understand their potential positions within business
ecosystems and what they can offer partners. They need to create skills
which are useful to potential partners.
It is vital to recognize that insurance is a key end-market for data
from telematic producers and as such, insurers need to insistent that
telematic sensors are incorporated into consumer, health, and home
products as fast as possible. Given that integrated use of telematics is
the core aspect of intensifying data collection and use, and therefore
vital for dynamic premiums and substantially lower cost adminstration,
 8  The Response of Incumbents    
247

insurers who do not proactively innovate and cooperate will be at a

severe competitive disadvantage. If current insurers do not proactively
engage, then there will be a strong impulse for these telematic suppliers
to approach newer, more innovative, insurance competitors.
Disintermediation is also a major threat. A growing trend has been
for IT developers to develop stand-alone modules for various compo-
nents of the insurance supply chain. For example, Guidewire offers a
core processing suite for P&C insurers. New entrants with existing cus-
tomers will therefore be able to buy in the technical capacity which they
may be lacking, adding modules from a network of suppliers, effectively
creating an ‘insurer-in-a-box’ to add to their suite of products. SAP is
an example of a firm which already offers a comprehensive IT insurance
package. They can add to this by contracting with firms who offer actu-
arial, or CRM, or claims, or other sector-specific expertise. Insurers who
aim to survive the coming perfect storm therefore need to ensure that
they leverage off their expertise and deeper understanding and therefore
are able to offer a superior value proposition than an ‘in-a-box’ solu-
tion provider can. Given that the cost of processing via IT software is
expected to fall by about 20% per year, this will be a sizable challenge.
The shorter product duration cycle in P&C means that these changes
are both likely to impact there first, and that it will be easier for exist-
ing insurers to migrate clients to a new IT system. While life and health
insurers may be less exposed, they face more issues with migrating long-
term policies, probably already located on multiple legacy IT systems,
to a new system.
Canas (2015) argues that the likely failure of insurers to respond will
mean that change will probably come from an outside disruptive tech-
nology innovator, similar to the way Uber is disrupting the taxi indus-
try, a process now called ‘uberization.’ These new entrants will probably
be existing technology firms with high social capital. The entry point
will probably be first via comparison quote engines, then purchase of
e-aggregators, then purchase of product creators. A growing alterna-
tive new entrant is Web-based group schemes, ‘peer-insurance,’ which
is similar to old-fashioned mutual insurers. These trends will mean that
existing multiple-task insurers will have to fight strong disaggregation
trends, as new entrants cherry-pick profitable tasks.
248    
M. Naylor

Equinix (2014) argues that given that the core competencies of insur-
ers are abilities to aggregate, predict, and manage financial risk using
data analytics, then outside organizations which possess these skills
are potential rivals. The core business of companies like Google and
Amazon is data analytics, their customer relationship skills are supe-
rior, and they are technological leaders in core areas like big data, AI,
and telematics. It is easier for these companies to add insurance to their
product mix than it is for current insurers to remedy their limitations in
AI data analytics and customer relationships.

Possible Future Pathways

Morgan Stanley/BCG (2014) foresees seven possible scenarios for insur-
ers subject to technological disruption:

1. Status Quo - insurers continue to offer the whole value chain. This is
seen as unlikely.
2. Insurer-based ecosystems - insurers create the platform which is the
core of an ecosystem and thus continue to dominate the value chain,
as well as expand into related areas. This is the best outcome for
insurers but seen as unlikely due to lack of platform expertise.
3. Traditional Partnering - insurers work with an adjacent entrant, pos-
sibly a bank, to offer insurance via the partner. The problem with this
is that traditional partners for insurers may also lack platform skills.
4. Complementary Platform - insurers seek a suitable ecosystem plat-
form to offer complementary services. The insurer loses control but
is a part of a dynamic ecosystem which has a higher chance of suc-
cess that a go-alone policy offers. This is the most likely outcome for
survivors.
5. Value chain disruption - insurers work with partners along the value
chain but face erosion of its value chain as more suitable adjacent
entrants expand into the insurer value chain.
6. Disruptive ecosystem entry - entrants from adjacent ecosystem sectors
expand into parts of the insurance value chain. This is already hap-
pening in some sectors.
 8  The Response of Incumbents    
249

7. Disruptive entry - entrants from outside sectors offer a new model of

insurance cover and squeeze the insurer out of significant parts of its
market, destroying scale, or attracting prime clients. This is the most
likely result unless insurers are proactive in removing the opportunity.

One of the key points which Morgan Stanley/BCG (2014) makes is

that any pathway which separates insurers from contact with their cus-
tomers and the resultant feedback exposes insurers to a significant risk
of gradual erosion of its role in an ecosystem to being that of a margin-
squeezed processing component supplier, so that client contact is not
maintained.
IAIS (2017) discusses three contrasting response scenarios. The supe-
rior route for incumbents is that of intensifying both the spread and
depth of client contact and data and then leveraging of this to create
deep client engagement, directly useful products, and value-added ser-
vices. This kind of deep authentic branding is, however, still largely
absent in the insurance sector so they lack experience.
The current response of many incumbents is to buy or partner with
a start-up or disruptor. This can be successful, though the track record
of incumbents in other industries gaining from these is patchy. The
reasons for this, as laid-out above, are that (i) the incumbent needs to
transform its IT system to reduce its marginal costs enough to make
use of the start-ups technology (ii) the incumbent’s culture needs to
be transformed, or the innovativeness which created the start-up with-
ers. The same arguments apply to start-ups the incumbent creates itself.
Responding to disruptive innovation is difficult.

Disruptive Innovators
Canas (2015) highlights examples of seven innovative firms disrupting
existing insurance markets. These firms can’t be regarded as perfect in all
areas, but they are useful examples.

1. Zenefits - founded in 2013, this cloud-based company is transform-

ing the HR business, by giving away its software and making profits
250    
M. Naylor

instead from commissions on health insurance sold via its HR soft-

ware. Because it is software based, it has low costs and can under-
cut all its insurance competitors, as well as the advantage of being
embedded within the client’s core software.
2. BizInsure - founded by an online business insurer and a leading US
broker, BizInsure enables clients to explore policies, to obtain quote,
buy instantly, and get them all declarations and policy details in their
email inbox within 10 minutes. They also offer phone or email access
to a licensed agent able to answer any queries, at any time.
3. MetroMile - connects clients to insurers who offer vehicle cover by
the mile driven, not on a time basis. It is a software-based download-
able app which can be installed on most current vehicle models. The
company has very low costs per client and has been advertising heav-
ily at those who drive less than 10,000 miles per year. It also offers
companies the ability to track their fleets and create online vehicle
usage and cost reports, as well as discounts on maintenance, and
information on the best parking deals. It has teamed up with Uber to
be the default insurance supplier to all Uber drivers.
4. Evosure - provides a Web-based platform for commercial brokers to
outline the types of client risk they want to be covered and to receive
specific quotes from underwriters.
5. Friendsurance - arranges peer-to-peer personal risk insurance cover
by creating groups of people with similar risks, normally via existing
social connections, and then offering this to underwriters. Pools with
less than expected losses are given a refund, while pools with higher
losses face no increase. The business concept is that using groups
of friends sharply reduces fraud and improves risk selection, as the
group jointly wants a refund.
6. Social Intel - underwrites risk by using software to analyze people’s
social media posts rather than traditional metrics like driving his-
tory or credit record. This cuts data collection costs sharply without
reducing underwriting quality by much. By being software driven,
they can provide underwriters with real-time updates. This leads to
interesting concepts like the insurer knowing that a client is at a ‘rave
party’ and can therefore cancel the car insurance for 24 hours, while
arranging an Uber driver.
 8  The Response of Incumbents    
251

7. Policy Genius - founded by two ex-McKinsey consultants who consid-

ered current insurers very backward and disagreed with the concept
that insurance ‘had to be sold.’ Instead, they thought that provision of
the right information would inspire clients to purchase online, as they
did for other products. An example of this is providing information on
what ‘people like you’ need cover for and linking this to a client friendly
checklist which takes 5 minutes and leads to a summary and a link to
quotes and online purchase. It builds trust by pointing out what kinds
of insurance you don’t need, as well as how need for cover changes over
the life cycle. It has strong appeal to the younger digital natives.

Other firms of interest are:

8. Aspiration - offers financial advice linked to investment funds. It is

predominantly software based and is aimed at middle-class inves-
tors with between $500 to $100,000 to invest. It charges no fees and
instead asks that investors to ‘pay what you think is fair.’ This surpris-
ing concept has led to a rapid growth in investment funds.
9. Lemonade - a P2P insurer. They take a flat 20% of premiums as a fee,
use 40% for reinsurance (internal and external), and 40% as a small
claims pool. Any profit above that is given to nominated charities.
These arrangements disincentive fraud. Adminstration is based on an
intensive software-based system, with 80% of most client activities
requiring no human involvement. Claims are nearly all automated,
with one claim paid in 3 seconds. Feedback from customers indicates
that the ease of application, lack of paperwork, rapid query feedback
(usually chatbot based), and speed of claims payment have generated
very high social media rankings and ‘makes incumbents seem like
they are from a past century.’
Lemonade involves three aspects which are different from the incum-
bent model: (i) P2P is simply a return to the old mutual insurer
model, so nothing new. The issue with the mutual insurer model was
a lack of capital, which Lemonade solves by their financing set-up.
(ii) Reinsurers play an increased role, which may not be as cost effec-
tive as an internal capital model. (iii) An intensive software admin
system is the real innovation. Most incumbents will struggle to adjust
legacy systems to Lemonade’s low admin costs and fast processing.
252    
M. Naylor

10. Insure-the-box/Drive-like-a-girl - these sister companies offer cheap

car insurance by requiring all clients to install a telematic box device
which monitors driving. Clients who do not fulfill cautious driv-
ing parameters are dropped or required to pay hefty surpluses.
The better the driver behaves, the lower the premium drops. The
device sends real-time reports, instantly alerts a staff member to any
crashes, and sends a crash report which includes data on aspects like
location, speed, force of impact, direction of impact, and engine
condition. A staff member can then alert emergency crews with
the car location and likely extent of injuries. The increased speed of
response has already saved lives. The company is especially popular
with young drivers who can prove that they shouldn’t have to pay
the extremely high rates generally imposed on them. Since driving
frequency and location are recorded, infrequent drivers or those in
rural areas can be offered substantial discounts. If the car is stolen,
the location can be tracked in real time. Note that the company
only attracts low-risk drivers, leaving high-risk drivers to traditional
companies.

11. Bought-by-many - this is an aggregator firm who finds and groups

people who have unusual insurance needs, for example, diabetics
who want travel insurance. It finds these groups mainly by using
algorithms which are designed to respond to Web searches by peo-
ple looking for insurance. These people are offered customized poli-
cies at no extra cost, and insurers are offered a block of clients with
a specific risk profile.
12. Tokio-Marine - has partnered with mobile phone operator NTT
DoCoMo to offer limited time insurance for specific events
arranged and paid for via a mobile phone-based app. This can nor-
mally be obtained on the spot, just before a purchase is made or
an activity takes place. An example is a non-car owner who bor-
rows a friend’s car and needs cover for ½ a day. The target is unin-
sured younger participants in active events and costs are heavily
reduced by piggy-backing on existing NTT software with invoicing
via NTT’s payment system. NTT gets extra business at minimal
 8  The Response of Incumbents    
253

cost, and Tokio-Marine has access to NTT’s 60M customers. This

approach has seen sales rise by at least 22% per annum across the
range.
13. Oscar - this is a rapidly growing New York health insurer, based on
an online only platform. Insurance can be obtained rapidly, which
a quote offered in 5 clicks. Oscar offers free phone calls to doctors
and a Google-map-based doctor-finder, as well as health price-com-
parison tool.
1 4. Youi - is a South African-based insurer which uses an intensively
underwritten model to provide customized quotes and thus skim
off good clients. It has a high social media presence including allow-
ing customers to post unedited reviews on its Web site in real time.
Currently, user satisfaction on its Web wall stands at 85%, and Youi
is rapidly expanding internationally.11Note that Youi has suffered from regula-
tory action in Australia and NZ over dubious sales practices.
15. Allianz - this German insurance giant is trying to transform into
a digital business model. It is spending €500M on IT per annum
and has 5 key initiatives (a) digitization of all business processes; (b)
enabling online research, quotation, purchasing, and servicing; (c)
standardization of data and policies so data for a customer can be
viewed in an integrated manner; (d) standardization of Web plat-
forms worldwide; and (e) establishment of an experimental center
for new ideas. Key decisions were to re-create products as mod-
ules which can be assembled in different packages and to change
culture by employing a substantial number of staff from social
networking companies. Allianz is a good example of a company
which is overcoming the inherence of running a legacy model while
transforming.
16. Wells-Fargo - has used the Stagecoach Island social media game both
to teach 16–24-year olds financial literacy and to cheaply acquire
tens of thousands as new clients. The accumulated emails enable
them to run a successful educational social media service.
254    
M. Naylor

17. Progressive Insurance - pioneered dynamic car policies based on tele-

matics, via Snapshot which offers drivers discounts depending on
when and how well they drive. They have signed up 3 million cus-
tomers since 2008, collecting a massive archive of driving data from
onboard black boxes. This data has enabled them to identify poten-
tial customers by risk type a lot more finely than competitors can.
18. Brolly - a UK InsurTech start-up which offers a digital agent focused
on helping customers, understand, manage, and buy insurance. The
App helps clients gather all necessary financial and lifestyle informa-
tion and then compares these to existing insurance policies, exam-
ining cover levels and conditions in order to spot coverage gaps or
negotiate better policy terms. Customers are offered detailed feed-
back and explanations and can then OK a search for better poli-
cies. While personal details are not provided to insurers, customers
as a group can be underwritten more accurately due to qualitatively
higher quality profiles.
19. Octo Telematics - is the leading global supplier of telematic devices
for the auto-insurers. They have 4.8 million connected users and
380k crashes analyzed. From this, they have developed a range of
data algorithms and provided in-depth accident-causal data to
clients.
20. DBS bank - this Singapore bank is not an insurer, but it is an exam-
ple of how a large incumbent can embrace the challenge of a digi-
tal future. Keys to this are ensured the central software system was
recreated end-to-end, focused on the customer experience, and on
the required cultural change. They found that setting up a separate
R&D unit failed, so they embedded innovation into every employ-
ee’s job. Every employee was taught how to be technology creative
and required to run experiments. A separate online only bank in
China has been established, which aims to cut admin cost by 90%.
Bancassurance market share has doubled.
21. L&T General Insurance - this Indian insurer was faced with the
problem of reaching customers spread across a subcontinent with
poor roads and few bank accounts, when skilled staff were hard to
find. Its solution was an end-to-end IT system which handles all
customer contact and payments via a mobile phone App. This has
 8  The Response of Incumbents    
255

ensured low costs and enabled low-value micro-policies to be issued.

It is scalable and handles sizable regional fluctuations in demand for
services. All customers get proposals and automatic translation ena-
bles policies to be offered in the preferred language. New products
can be launched very fast. Growth in India has been fast, and the
model has the potential for use across all emerging markets.

Keys to Insurer Success

When current company mortality rates are low, existing companies can
fall into a ‘success trap,’ that is, focusing on exploiting existing opportu-
nities and being as efficient as possible, rather than understanding that
existing opportunities will soon disappear. Management focus should
instead be on exploring and creating new opportunities. The creative
destruction of existing management structures and replacement by flex-
ibility and innovativeness have to be the key to success.
PWC (2014) argues that there are ten attributes which will mark out
the future survivors in the insurance sector from those who will fail.
1. Contextualized - matching the customer experience with their indi-
vidual preferences, by using digital interaction to know customer
preferences and modifying the buying experience to suit.
2. Optimized - continually testing and enhancing the interaction with
the customer, by individually asking for feedback, and modifying
interaction to increase the effectiveness of capturing leads, closing
sales, and managing contracts. This has to be open and two-way, so
the customer needs to feel in charge of the engagement style.
3. Engaged - creating a compelling reason to buy by increasing cus-
tomer understanding and engagement. This should be bench-
marked against best IT sector practice.
4. Targeted - turning the increased insights into customized products
by recognizing what features customers value and then expanding
services. Expertise in predicting customer behavior is critical, as is
dynamic pricing.
256    
M. Naylor

5. Guided - helping customers decide what the correct choices are, by

using a dynamic structured decision-making process.
6. Synchronized - creating a seamless, customer-lead, multi-channel
experience, including online, social media, and human. The cus-
tomer has to choose the mix of channels. Personal advisers should
be available via multi-channels, phone, Skype, live chats, texts,
Facebook messaging, social media, as well as face-to-face.
7. All-embracing - reshaping the entirety of company processes toward
being customer focused. Every aspect of a company’s operations
needs to be re-examined so that the new approach is embedded in
every area, and all managers need to be comfortable with the big-
data decision-making style.
8. Nimble - leading rather than following the changes, taking control
of the market place rather than reacting. Perfection is not as impor-
tant as speed, and ability to change. A learning culture is vital and
staff need to be allowed to fail.
9. Efficient - delivering more for less, by pricing risk more closely to
customers, and using efficient internal processes. This does not
mean the cheapest product, if customers understand choices.
10. Sensitive - clarity about what the company is doing with data and
why. Data should only be collected if there is a reason clear and
beneficial to the customer, and privacy is critical.

To this could be added 11. Diversity - an inherent ingredient of flex-

ibility and innovativeness is employees who can approach an issue
from multiple directions. It is thus vital to avoid creating a monocul-
tural workforce with a dominant shared worldview. Instead, compa-
nies have to deliberately diversify their workforce in the right way. This
then creates an internal richness of approaches and closer understand-
ing of a diverse client base. It is vital to understand that this is not a
simple matter of employing and promoting across ethnicities and both
genders, as there is no guarantee that employees from different ethnic
groups actually have differing worldviews. It is more about deliberate
recruiting people who think differently. Important external stakeholders
and paradigm thinkers need to be involved. It is also vital that existing
 8  The Response of Incumbents    
257

management culture is supportive of this diversity and encouraging of

more complexity in internal decision making.
A simple example of the new approach would be the integration
between the Web site and the call center, so when a customer phones
the employee knows which aspects of the product the caller has been
searching, or if quotes have been received, so that they don’t have to
force the customer to repeat the procedure or re-input information.
Then prospective customers can get information and initial quotes on
the Web site, with the always available option of a virtual or human
adviser on standby.

Value-Added vs Commodity
The key thread throughout this book is that a large part of the pro-
cess of administration for insurance products is replaceable by software
which has a very low marginal cost and few barriers to entry. Thus, the
inevitable result is that insurance administration will become com-
moditized and therefore low profit. The profit, as always, will be where
insurers can add value to those products by creating a deep brand which
is in-tune with client intuitions. McKinsey (2012) argues that correct
use of social media within a deep brand can increase margin by up to
60%. McKinsey, however, then argues that while insurance as a product
has a high potential for capturing value, current insurers are among the
worst sectors at achieving that.
It is vital to note that in a world of social capital based on busi-
ness ecosystems, insurers cannot take heart in the idea that they can
just reach agreements to supply product to other ecosystem members to
distribute to clients. This is because once they lose client contact they (i)
become invisible and therefore highly likely to end up trapped in a low-
value commoditized market, (ii) do not get the data they need to grow
into the data analytics area, (iii) will therefore find it hard to expand
into the growing value-added areas of the ecosystem, and (iv) will dras-
tically shrink in terms of employee size as core functions disappear into
software.
258    
M. Naylor

In terms of pricing, three differing price models come to mind:

(i) The core insurance products and the added services are all provided
at a profit-margin about average total cost; (ii) the core insurance ser-
vices are provided at a profit-margin about average total cost, with the
added services provided at a very low marginal cost, or free, as a prod-
uct purchase inducement, or more revolutionary; (iii) the added services
become the core product priced at a profit-margin above average total
cost, with insurance product provided at a substantial discount, as a ser-
vice purchase inducement. Given that the services are likely to be more
intimately integrated into customers’ lives, and therefore valued more,
option (iii) has more to recommend it than traditional insurers may
consider.

A Vision of Dynamic Insurance

As John leaves his house, the sensor at the garage door notes that he is
leaving and instructs the house to make sure that everything is secure
until his wife, Helen, gets back at around 5 pm. The house bids him a
cheery farewell in the sultry voice which John prefers.
His car welcomes him in a deep masculine voice as he gets in and
adjusts settings to his preferences. He tells the car that for a change he’d
like to manually drive. He glances at the dashboard indicator which shows
that his car insurance has switched from his house insurer, which handles
his insurance while his car is parked at home, to his driving insurer.
As he accelerates onto the expressway, he notices that the dashboard
indicator shows that his current insurance risk rating has jumped. Several
voice commands later he is scanning his recent driving record and con-
cludes that he has recently been heavy footed on the accelerator, and his
insurer has concluded that he is higher risk. He pushes the black button to
put the car into auto-drive and sighs with relief as the risk indicator plum-
mets to 10% of the manual drive rating. He watches as his car slides over
into the fast, auto-drive-only lane, slots itself into a coordinated convoy of
cars, and accelerates to 150 m/hr. He idly wonders whether he should only
manually drive on weekends, since the car is safer and faster than he is.
He switches his attention to the car’s built-in screen and sees that his
life insurer has sent him congratulations for his recent avid gym attend-
ance and has cut his weekly premiums by 15%. He scans his personalized
news and reads an article sent by his health insurer outlining research
showing the benefits of a new kind of berry and clicks on the button to
receive his free sample pack. He thinks about the coming family vacation,
considers the recommended insurance add-ons, and asks the system a few
questions, before going to the insurer store to make his choices.
 8  The Response of Incumbents    
259

He notes that his food cupboard and fridge have been talking to his
health insurer’s computer and have agreed that if his family lifts their
consumption of greens by another 10% then the health premium will be
dropped by 8%. He notices a linked offer from his life insurer.
He opens a pop-up from Google Insurance offering driving insurance
for the next week at a 30% discount if he agrees to not drive manually. He
accepts the offer with a click and happily watches the dashboard rating
indicator drop even further.
He surfs his favorite Web sites until the car informs John that he has
reached work. Just before he hops out, he tells the car to go park itself
at the Uber car park and to accept any fares. He watches the dashboard
indicator switch to Uber insurance. He knows that auto-Uber use doesn’t
always go smoothly but he likes the extra income and remembers an acci-
dent the prior month, caused by a third-party manual driver, and how the
Uber insurance computer informed the police of exactly how the crash
had occurred, arranged his car to be repaired, and provided him with a
backup car for two days, all set up according to his specifications. All this
before he had finished work.
John briefly stops to talk to his friend Eric. They discuss how Eric’s
employer, who is a major player in the mobile messaging industry, has
been recently gaining substantial market share in the health insurance
sector, due to their superior social media software. Eric expresses his
relief that he left his old insurance company when he did and describes
with disbelieve how its dinosaur approach to technology led to its recent
bankruptcy.
As they entered the office foyer and chat, they passed Eric’s favorite
café. The barista handed him his morning double-decaf soy latte, which
his car had arranged via the café app. Both men agree that times had cer-
tainly changed for the better.

Conclusion
The CEO of Lloyds, Inga Beale, announced in a speech on July 2015,
that the disruption moment in insurance was 2015, when irreversible
forces were being set into motion. She said that non-traditional holders
of customer information like supermarkets, or social networks, had an
in-depth understanding of and contact with clients, which insurers did
not. She also said that insurance was rapidly globalizing so that the dis-
ruption will probably come from foreign providers.
260    
M. Naylor

The future of traditional insurance is looking as bright as the future

of Kodak was in 1995. Note that Kodak was the inventor of the digi-
tal camera, but executives choose to delay its development because they
feared its impact on their existing business model of cheap cameras and
expensive film. Insurance CEOs are similarly worrying too much about
destroying their current cash flows to see the larger threat from not
investing in the management changes required to meet disruptive tech-
nology. In general, there are currently two types of insurance executives:
those who believe that their sector is facing disruptive transformation
and those who are not paying attention.
One of the key reasons for the inability of existing insurance compa-
nies to react successfully is that the people in charge have achieved their
success by acquiring and displaying the attributes best suited to the past
environment. Within a new environment, they may be no better able
to acquire and display the new skills required than subordinates are and
probably less than outside entrants. This is a major problem given that
the key to insurer survival is a complete cultural transformation. This
need for managers to transform themselves is difficult for existing com-
panies and the major reason why they are normally out-innovated.
This is particularly true when shape of the future is still undefined
and therefore insurers are unsure which new skills they have to acquire
and can be sure that whatever choices they make now will probably be
faulty. In this new fluidity, the ability to experiment and restructure
rapidly is more important than perfection. Insurers need to create a
Google-X style experimental unit which requests ‘long-shot’ ideas from
staff and trials them. They need to adopt Google-X’s mantra of ‘fail fast,
fail often, fail forward.’ The acceptance of a culture of creative failure is,
however, a huge cultural change for insurers to achieve.
There is a positive feedback loop to database size, as more informa-
tion on clients enables an insurer to price more finely, and therefore
sell more, and therefore collect more data. This gives a huge first-mover
advantage, and therefore stragglers will disappear. Insurers need to
actively collaborate to create across company and across industry data-
bases, they need to be constantly developing new analytics ideas and
they need to manage data as trustees.
 8  The Response of Incumbents    
261

McKinsey (2016b) sees four key areas competitive advantages in the

new digital era:

1. Distinctive technology - obtaining proprietary low-cost or base

technology
2. Distinctive data - obtaining propriety data and using this to provide
innovative customer service.
3. Platform providers - creating broad-based software enabled platforms
upon which third parties can build applications.
4. End-to-end solutions - customizing products to provide customers
with unique complete solutions.

Successful survivors will have to understand that (i) competition is

global, (ii) social engagement is vital, and (iii) creativity is essential.
Insurers have to be constantly inventing and trialing new products, new
services, new customer connections, and expressing ideas clearly and
imaginatively. They have to reward staff for being creative and experi-
mental. Customers have to be actively engaged and connected in ways
which make clear insurers’ value proposition.
What insurers need to do to react to the looming technological per-
fect storm is reasonably clear. It is also clear that few insurers are even
aware of the issues, let alone are transforming with sufficient speed.
PWC (2014) argues that the key issues holding insurers back are: (i)
a preference for incremental over radical innovation, (ii) legacy sys-
tems - heavy investment in upgrading existing systems which leaves
little room for new systems, (iii) constantly moving performance tar-
gets, (iv) inexperience of actuaries and underwriters with big data, and
(v) success being gauged on product development and sales rather than
smart solutions and enduring engagement. Most insurers also feel lost
and inadequate in possible growth areas like cyber-insurance. They are
also unlikely to be able to excite enough innovative young IT graduates
to join them and thus will not be able to create the required skills base.
In the past established insurers have been protected from new
entrants by the high fixed costs of developing systems and distribution
capacities. In the new digital era, however, these barriers have disap-
peared, as many IT-based companies have better existing systems than
262    
M. Naylor

insurers do, making it easier for new entrants to develop the required
systems than it is for existing insurers to transform their systems. Given
the abysmal level of social capital of most existing insurers, new entrants
may have a distinct competitive advantage.
The insurance companies who survive the coming perfect storm will
be those who recognize that they are no longer primarily insurance com-
panies, but that in the future they will be primarily information com-
panies specializing in analytical risk-based client services. Insurers who
aggressively build the richness of their data and its analysis will find new
sources of profit mushrooming. Insurers need to reimagine their industry.

References
Boston Consulting Group. (2015). The Double Game of Digital Strateg  ’,
P. Gerbert, C. Gauger, & S. Steinhäuser, BCG Perspectives, Oct 16.
Canas. T. (2015). What will be the Uber of Insurance? Insurance Thought
Leadership, PWC, July 22.
Capgemini Consulting. (2015). Strategies for the Age of Digital Disruption,
Digital Transformation Review, No 7: Feb.
Capgemini/Efma. (2016). World Insurance report.
Equinix. (2014). Challenge to change: three-part series, Acord/Global Futures &
Foresight.
IAIS (2017). FinTech Developments in the Insurance Industry, International
Association of Insurance Supervisors, Basel, Switzerland.
Kanter, R. (2006). Innovation: The Classic Traps, Harvard Business Review, 84 (1).
McKinsey. (2012). The Social Economy: Unlocking value and productivity through
social technologies: consumer financial services, McKinsey Global Institute.
McKinsey. (2016a). ‘An Incumbent’s Guide to Digital Disruption’, McKinsey
Quarterly, May.
McKinsey. (2016b). ‘The Economic Essentials of Digital Strategy’, McKinsey
Quarterly, March.
Morgan-Stanley/ BCG. (2014). Insurance and Technology: Evolution and revolu-
tion in a digital world, Boston Consulting Group, USA.
PWC. (2014). Insurance 2020: The digital prize—taking customer connection to
a new level.
Satell, G. (2012). Business Models and the Singularity, May.
 9
Regulatory and Legal Issues

Pricing Regulation
In general, governments for reasons of equity have placed restrictions
on the ability of insurers to convert information about risk differences
between people into differences in rates and premiums. There are three
levels of insurer price freedom: (i) Community rating - The insurer is not
allowed to vary the price between customers based on risk. Often used
for health insurance. (ii) Restricted rating - The insurer is allowed to vary
within set bands, possibly with a small number of groups. Often used
for workers’ compensation. (iii) Unrestricted - Insurers are free to set
rates as they wish, either by groups or individuals.
For dynamic insurance to work, insurers do not have to have com-
plete price freedom, they merely need the ability to vary rates or offer
conditional discounts by an amount large enough to induce the desired
behavioral change. Research clearly shows that a substantial proportion
of customers respond in behavioral ways to small price differences.
This will probably take the form of rate reductions for positive
changes, rather than rate increases for adverse changes. Governments
can be expected to be cautious about allowing rate freedom for risk

© The Author(s) 2017 263

M. Naylor, Insurance Transformed, Palgrave Studies in Financial Services Technology,
DOI 10.1007/978-3-319-63835-5_9
264    
M. Naylor

factors which are unchangeable, e.g., genetics, and bolder in allow-

ing rate freedom for lifestyle choices. However, choices will have to
be made. Does the government allow insurers to price young drivers
who display very poor habits out of the market, or reward those who
undertake regular exercise so substantially that the unfit feel discrimi-
nated against? Because telematic feedback is tied to individual activities,
it is likely to generate a powerful nudge effect toward better individual
choices. Tying this feedback to small monetary inducements should
have a beneficial impact.
If the rates discrimination induces positive behavioral changes
which can be shown via research to contribute to reducing social ills
and therefore reduce government costs, e.g., by reducing car crashes
or reducing health costs, then regulators can be expected to be more
supportive of rates freedom than they have been previously. When
Progressive increased auto-insurance premium rates for its worst 5% of
drivers, (who cause a high proportion of crashes) there is no reaction
from regulators.
Customers have, in general, been supportive of linking discounts
to telematic feedback, so political reaction should be minimal for
well thought through proposals. After all, the current factors used for
underwriting are broad classes, and customers often feel aggrieved -
Why should a careful young driver pay more for auto-insurance than
a careless middle-aged driver? Telematic data would provide a detailed
breakdown of the individual’s risk factors and therefore provide moral
justification. Dynamic insurance, which gives feedback via changes in
rates, will create a strong sense of moral justice and an obvious reward
for good behavior. Complexity will arise, however, when big data anal-
ysis discovers correlations between multiple factors and beneficial out-
comes, and where the logic is unclear. What is the link between early
evening shopping and car crashes?
There are, however, many legal and ethics issues with the increas-
ingly sophisticated use of data. One issue is that customers need to be
able to understand what factors their premiums are based on and why
weightings on those factors are fair. There may need to be causation and
not just correlation. This can be very difficult to achieve if sales staff or
even underwriters find it hard to understand the models, and nearly
 9  Regulatory and Legal Issues    
265

impossible if dynamic multi-causal models are used. A related issue is

that insurers have to be able to show that banned discriminatory fac-
tors like ethnic origin are not proxied by a closely correlated factor.
Another issue is that insurers have to be able to show that data is being
protected, and this is difficult if data is sourced from or used by exter-
nal firms or connected ecosystems. There are particular issues if data
is sourced from social media, as is increasingly common. One known
complication here is that users often lie or are deliberately malicious on
social media postings.

Data Law
The creation, storage, and use of personal data raise many legal, ethi-
cal, and regulation issues. What right does each of the parties have in
respect of the raw, aggregated, and processed data; what uses should be
permitted; and in what circumstances? What rights exist to control the
use of data across ecosystem companies, when the customer is unaware
of their existence or data usage intentions?
A wide range of legal rights and obligations is developing in rela-
tion to data, based on traditional intellectual property rights, contract
law, and regulatory law. This means that use of data without the right
licenses or permissions can lead to large damages claims. Meanwhile,
regulatory data protection liability is being expanded.
Therefore, insurers need to make sure they have all the rights they
need to all the data they use and in all the ways they use it. In legal
terms, this means licensing it, processing it and using it correctly, and,
especially with personal data, obtaining the explicit informed consent of
the individual concerned in order to comply with data protection law.
This requires a structured approach to data use and governance.
It is useful to classify commercial data into four groups:

1. Metadata: data about the data. For example, not the details of a
phone call, but when, where, and for how long.
2. B2B Data: data about individual businesses, which are suppliers or
customers.
266    
M. Naylor

3. Anonymized B2C Data: data about individual customers but with

individual identifiers removed.
4. Personal Data: data about individual customers, which has individual
identifiers.

Nearly all big data analysis only requires one of the first three groups, as
outside of telematic or purchase data, it is rare for data to need to iden-
tify individuals. Most big data analytical systems don’t use individu-
alized data, they use data about ‘someone like you’, and so it is about
slotting individuals into a tighter category than currently. Data analy-
sis is about assessing correlations about how ‘someone-like-you’ suffers
adverse events. Only data involving identifiers will face legal issues.
Insurers will need data law managers. They will have to understand
the data analysis structure in terms of: (i) data inputs - Where does all
the data, the structured data sets (industry, marketing, and personal)
and unstructured data (social media, mobile, and Internet) come from?
(ii) data processing - How does the business intelligence software that
analyses the input data work? and (iii) data outputs - Where (internally
and externally) do the data outputs go, who uses them and how?
This requires close cooperation between the insurance company’s
legal team and its technology group. These groups will have to assess
the IT system structure maps in terms of the flow of information. The
structure map must enable tagging of all the characteristics of each type
of data and map it to all the relevant licenses, permissions, and con-
sents that attach to it. The legal team needs to understand the technical
vocabulary of IT architecture and the technology team needs to become
familiar with the concepts of copyright licensing, confidentiality, con-
tracts, and explicit informed consent under data protection law.
Four steps are required to ensure legal compliance: (i) A ‘deep dive risk
assessment’ examines current data use; reviewing and assessing it, then
reporting back to senior management on issues and solutions, (ii) A big
data strategy should then be created by a senior management team. This
should include members from all business areas and should clarify and
formalize the Data Strategy as a written statement of high-level objec-
tives, goals, and relevant considerations, (iii) A big data policy is then
created, which is a high-level plan showing who’s doing what, when and
 9  Regulatory and Legal Issues    
267

how, and (iv) A policy and procedures plan is then created, setting out
how data governance should be handled.
The insurance of drones raises a number of additional as yet unre-
solved issues, especially as regulation is currently unsettled. Apart from
physical damage to drones, drone operators can potentially be sued for
trespass, nuisance, and privacy invasion. Insurers will have to determine
how to insure the innocent use of drones but exclude the ill-intended
use. Given that drones with cameras can survey neighboring land even
when not passing above them, areas like trespass will not be easy to
define. In general, while there is no right to privacy if a person is in a
public place or in a private place which is easily viewable from a pub-
lic place, a drone can view a high-level apartment or see over a high
fence - this can easily be defined as an ‘unreasonable’ breach of privacy.
Potentially, a drone operator could be sued for invasion of privacy every
time they operate a drone in an urban setting. Failure of a drone could
mean damage to people or property below it.
Policy terms will have to be precise, and insurers will have to clearly
decide what areas they do not want to insure, yet insurers who do not
offer a wide enough definition will not find customers. Does any liabil-
ity from drones arise for house contents insurers?

Data Privacy Issues

Three approaches have been used to collect individualized data:
(i) A free market - whereby firms collect and sell whatever data they can
collect, (ii) an opt-out system - whereby data can be collected unless cus-
tomers actively refuse, and (iii) an opt-in system - whereby data is only col-
lected if customers positively agree. A mix of these three is used in most
countries, though the USA tends toward (ii) and Europe toward (iii).
Dynamic insurance is only possible if data is willingly shared. This
will only be possible if customers give their informed consent, which
they will only do if they trust the insurer, if there is established social
capital. An alternative is for data to be held by a neutral body and used
by insurers on a regulated basis.
268    
M. Naylor

The main issue is that legislation to enforce opt-in is difficult to

enforce in practice as data users often do not need to deal directly with
the public. Insurers are able to buy data sets from venders or receive
it from ecosystem members. Data vendors use schemes like free sur-
veys to gather data. Many Apps include cookies, the uses of which it
takes an IT expert to understand. Many Apps use a tick-in contract to
ensure legal opt-in, but 99% of users never read the terms and condi-
tions. Thus data law can be obeyed in letter but not in spirit.
The collection of data is‚ however‚ not as much of a concern as the
public often thinks as data privacy concerns and regulation mainly
revolves around personal data, and most data analysis does not involve
this. Thus, in general, individual privacy will not be compromised.
Data privacy systems will need to set rules around: (i) storage - what
data is stored and for how long, (ii) access - which type of employees
have access to what data elements, (iii) sharing - what types of data can
be shared within the insurer and with what external parties, (iv) merg-
ing - what data should be linked or amalgamated, and (v) reacting - how
each data element should be acted upon. These rules have to cover both
positive and negative information or behavioral predictions. For exam-
ple, insurers have to respond cautiously to buying patterns, which indi-
cate pregnancy or adverse health conditions, as the client may not have
informed friends or their spouse. The behavior of others with the client’s
social network is often vital to buy decisions, yet care is needed when
contacting a client because of the actions of their friends.
With regard to personalized data, there are two areas of concern. One
is that data analysts have shown that it is possible to combine non-iden-
tified data from a range of sources, including political affiliations in the
USA, and get close to identifying preference types by street or house. This
can then be linked to browser history to actively, in real time, understand,
and monitor personal preferences. For example, only 15 minutes of real-
time driving data is enough to identify individuals from a selection of
possible drivers. There will need to be regulation around this.
The other area of concern, which is important for dynamic insurance,
is telematics data. If this is linked to underwriting or services then indi-
vidual identifiers will be required. In addition, analysis of data may give
insurers substantially more information about a customer’s risk factors
than the customer pocesses. Insurers can then be seen as ethically bound
 9  Regulatory and Legal Issues    
269

to inform the customer of the risk factors they face or ways they can
reduce risk. This could be complex for genetic or other health factors
where the factor only shows a tendency rather than a high probability,
or where there is a complex and poorly understood but robust, correla-
tion between multiple factors.
Telematic, browsing, or other individualized data is thus a major
privacy issue and of deep concern to many people. If it identifies your
location in real time it is a private security issue. Health data may also
involve intimate and sensitive information. Used inappropriately these
types of data can be used to harass or blackmail people and thus this
area can cause general unease.
One suggestion is for individualized data to be held by a regulated
3rd party and provided to ecosystems under strict conditions. This
has strong merit. For individuals concern about insurer use of data
will probably not revolve around the data held but around the uses to
which it is put. The problem is that understanding this involves exam-
ining the validity of AI algorithms and 99% of the public have no
capacity to scrutinize those. For example‚ one suggestion is for govern-
ments to require insurers to provide to individuals all the information
held about them, so they can correct errors. However‚ customer access
to their own individualized telematic data may be problematic because:
(i) they can’t access it without data warehouse programs, (ii) its size will
exceed individual’s capacity to cope, (iii) it can’t be understood without
using data algorithms, (iv) big data analysis often involves indirect cor-
relation between two variables, which have no causal link but proxy for
an unknown third variable, thus causing confusion, and (v) it will not
tell a customer how the insurer uses the data, especially the weights the
insurer’s systems places on differing aspects of the data, which is com-
mercially sensitive. What will be of more use to a customer is to know
their ID data and a summary of the assessment of their risk, which the
insurer has discovered from their own analysis.
It needs to be noted, however, that the urge for privacy is not inher-
ent but cultural. Before the 16th century a lot of activities, which we
currently regard as private, like toileting or sleeping, were public or
done in groups. Today’s younger generations, who are growing up in
a globally connected world with significant moments posted online,
270    
M. Naylor

tend to regard freedom as not bound up in self-contained autonomy,

but in enjoying access to others and in inclusion in a global public life.
Research does support this, showing that there are generational differ-
ences in data privacy assumptions. The younger generations are thus
less likely to forego the benefits to be gained by agreeing to share data.
The assumptions which European Commission (2013) make about the
inherent need for data privacy may not be shared by these younger gen-
erations. The website Patientslikeme.com shows that at least some of the
public are prepared to make detailed health information publicly avail-
able in an effort to find a cure for their conditions.
Hacking of data is a serious concern. The possibility that hackers could
take control of health or car telematics is real and can be used to terrify,
blackmail, or for terrorism. This could lead to increased auto-theft by
hackers as well as hacker-caused auto-accidents. Insurer data security is
thus a serious concern, both from an operational and a legal angle. Insurers
will have to work closely with smart device manufacturers to ensure data
privacy is maintained. Does a homeowner’s failure to secure their smart
devices from hacking breach the conditions of their house or contents
insurance? Who is liable if a telematic device gets a health metric wrong?
Any unauthorized access to or public release of data or illegal use of
data will expose the insurer to legal liability and destruction of their
reputation. Given the need for increased trust outlined in earlier chap-
ters and the increased ease of customer switching, a serous data breach
could lead to bankruptcy.
Issues also arise around the use of autonomous digital agents by cus-
tomers. The interaction between these and insurer IT systems, with a
mutual exchange of data which is not actively supervised by humans,
requires robust procedural safeguards. Savvy customers are increasingly
using these agents to control access to personal data and‚ in some cases‚
attempting to charge firms for their use of data in analysis.

System and Flash Crashes

As IT systems become more complex they create the possibility for
unexpected but sudden cascades of failure. Complexity leads to unpre-
dictability. It is impossible for codes to be created which are bug-free, as
 9  Regulatory and Legal Issues    
271

no single person fully understands the millions of lines of code. Certain

combinations of events can lead to system freezes or adverse reactions‚
as has occurred in stock markets. Algorithms are not open to public
inspection and verification.
If these crashes are linked to telematics or to auto-drive cars then the
consequences could be serious. This will impact in two ways: (i) The
insurer admin system crashes, preventing the insurer from continu-
ing normal business, and (ii) telematics crash, individually, or as a net-
worked group. The insurer or its clients would then be liable for a mass
adverse event. A few car crashes caused by software errors might not be
fatal to the autonomous driving industry. But if autonomous cars are
also connected, via car-to-car systems, or car-to-infrastructure, then a
systemic problem could bring down a whole network of cars. These are
all insurer legal nightmares.

The Ethics of Insurance Bets

Insurance has traditionally been based on the concept of a ‘fair-bet.’
Neither you nor the insurer knows what your future holds, so premiums
are based on broad generalized groups. Insurance markets exist because
no one really knows whether or not some bad event will happen. We
may know the historical odds, but we can’t accurately predict the event
for particular instances. Pools of policy holders were thus traditionally
created with premiums sufficient to cover the payouts for accidents that
will eventually occur within the pool as a matter of statistics.
Moving to individualized premiums, based on real-time behavior
breaks this concept, the equity of this ‘fair-bet’ - you are charged your
actual risk level. This raises a question about ethics: is it problematic to
offer insurance bets when the outcome is increasingly certain? If one
party has much more information than the other party, ‘asymmetric
information,’ is an insurance transaction morally questionable? If insur-
ers can know everything about customers related to risk, and the insurer
knows that the individual’s chance of an a specified event is extremely
small, then it doesn’t seem fair to charge a customer anything. If the
customer knew they had no risk, they wouldn’t want to pay anything.
272    
M. Naylor

Increasing asymmetry of information may erode the moral justification

for insurance markets to operate as they currently do.
Traditionally insurance law and regulation have assumed that indi-
viduals know more about their personal risk than the insurer does, so
that law requires individuals to disclose relevant data. In an age of big
data, however, the insurer may know a lot more about the risk that
some individual faces than that individual does. The insurer can then
set up a biased bet. Regulation will be needed to ensure the insurer pro-
vides a summary of the risk factors which the customer faces.

Privacy Legislation/Regulation
The European Commission (2013) argued that the Internet of things
will require strict privacy protocols and extensive security. Creating
restrictive data regulation, however, can be a major competitive hin-
drance for a countrys’ insurance industry. For example, EU direc-
tive 95/46/EC (Data Protection Directive) requires that personal data
is only collected for ‘specified, explicit, and legitimate purposes’ and
requires that it must be ‘adequate, relevant, and not excessive’ in relation
to the purposes for which it is collected. Typically, regulations require
that data only be kept for as long as it may be lawfully used, only be
used for the purposes which the individual consented, and the insurer
must be satisfied as to the information’s accuracy.
These rules contradict the requirements of big data analytics, which
involves collecting the maximum amount of data from the largest range
of data sources and then using that data across a range of purposes, often
seemingly unrelated to the initial reason for the data collection. For
example; shopping data used in an algorithm to predict car crashes. The
EU has thus modified 95/46 by adopting the General Data Protection
Regulation, which relaxes these restrictions. A part of this, however,
involves stricter conditions on ‘profiling’ purposes, whereby individu-
als have a right not to be subject to a decision based solely on automated
algorithms. Profiling based solely on sensitive data, such as health data, is
prohibited. Profiling based on pseudonymous data is allowed. GDPR also
requires a more active consent process, including express consent for data
 9  Regulatory and Legal Issues    
273

to be processed. Organizations must be able to demonstrate the individual

really understands and is in agreement with how their data will be used.
The issue is that big data analysis inherently involves the merging of
data from a large number of individuals and the insurer than explor-
ing the data in new and innovative ways, which will not have occurred
to insurers at the time of gathering data. Data may also be gathered by
other firms in the ecosystem, or automatically via telematics, a lot of
which may be rough and uncertifiable as 100% accurate. The data pool
will also be too large to allow extraction of individualized data. It may
thus be impossible to ensure strict security and a balance will be needed
between privacy and the advantages of the coming technological revolu-
tion. Choices will have to be made between insurance innovation and
privacy. It needs to be understood that for big data, there is no need for
100% data accuracy as each item of data plays only a minuscule part in
creating a risk profile, unlike the current situation.
A general restriction on data use is thus likely to halt domestic insurer
innovation and lead to their replacement by disruptors from less strict
jurisdictions. The EU’s GDPR may thus prove difficult for insurers and
may hamper InsurTech within Europe. Data controls need to be only
applied to data which directly identifies individuals. There is increas-
ingly regulatory competition between countries, with jurisdictions
like Singapore proactively introducing regulation reforms to attract
InsurTech innovators.
More importantly in an age of big data and AI analysis, regulation
should not be about the use of pieces of data, but be focused on regula-
tors examining the algorithm code or AI settings, so the privacy or ethical
dimensions of the insurers’ IT system can be examined. After all, in an age
of big data, human analysts do not look at the raw data, what they do is
set the framework of AI decision making. Regulation relating to algorithms
is more important than regulation relating to data. Regulation is required
around the aggregation of data and its use to back-identify individuals or
vulnerable groups. Regulation is also needed around data breaches and the
on-sale of data, or its transfer after takeovers or bankruptcies.
Another issue is that current regulation assumes large vertically inte-
grated insurers who perform all roles. If disruption leads to extensive
disintermediation and therefore a fragmented market, there will need to
274    
M. Naylor

be regulation aimed at specific roles within the market. For example,

regulation covering; client acquisition, customer relationships, under-
writing, claims assessment, and payments.

Unavoidably Higher Risk Customers

As discussed in an earlier chapter individualized pricing and the asso-
ciated incentives will reduce the average risk and spread risk. This will
increase the group for whom insurance may be unaffordable. For those
for whom their higher risk is due to factors over which they have no
control, insurers or governments may have to consider how to protect
them. There three main ways to do this: (i) impose restrictions on the
ability of insurers to set premiums based on individualized risk factors,
(ii) develop market-based risk sharing schemes, and (iii) create a govern-
ment agency to act as an insurer of last resort.

General Regulation Recommendations

Regulation should be digital-friendly, technologically neutral, and
future proof. It needs to be principles based and flexible. Regulation
needs to stop being based on insurance type, but become activity based,
applying to all insurance activities and providers, regardless of size or
origin. Regulation has to be internationally agreed so that national regu-
lators have to exchange data and experiences. Data and privacy proto-
cols have to be international.
Regulators need to support innovations and allow failure, which is
against the traditional regulatory culture. To do this regulators need to:
(i) create an innovation ‘regulation sandbox,’ where insurers can trial
new ideas for a while outside existing regulations, (ii) create a team who
understands technology and the innovation process, (iii) lower licens-
ing barriers, (iv) help small innovators comply at minimal cost, (v) allow
heavily automated insurers to use third-party sign-offs on organization
competence requirements, (vi) allow ecosystems to be licensed as group
or allow the insurer to certify data or procedures provided by non-insurer
 9  Regulatory and Legal Issues    
275

parties, and (vii) weaken merger and cooperation regulations to allow eco-
systems to strengthen relationships or use pricing/procedure agreements.

Auto-Insurance Vs Product-Liability Insurance

In the short-term, major liability issue arises from a mix of auto-drive and
self-drive - how to partition fault if the car was in auto-drive but the driver
was inattentive? Level 3 and 4 cars will be a major issue - drivers will not
be paying attention to driving and cannot be expected to take over in an
emergency if the software defaults to them. Longer term, for level 5 cars the
turning point for insurers will occur when owners start to send cars off by
themselves, to pick up the children, or a pizza, or to find a car park. If there
is a crash when no one is in the vehicle, then it is obvious that the product
producer is to blame, via product liability. How can laws which assume a
‘driver’ is responsible for errors work if there is no one in the car driving?
Insurance for level 4 or 5 auto-cars will thus eventually move from
being covered by drivers under auto-insurance to being covered by
car-manufacturers, or software producers, under product-liability law.
Insurer cash-flow will drop as profit margins in product-liability cover
are significantly lower than retail insurance margins. There is an issue
for older auto-cars as telematic degradation can be expected over time.
Given that telematic sensors need to be keep clean and correctly aligned‚
insurers may need to set conditions around maintenance. Therefore,
there is a scope for liability to increasingly shift to telematic maintenance
companies. There will always be a need for residual auto-insurance to
cover damage by external objects, flood, bushfire, terrorism, etc.
Brookings Institute (2014) argues that the USA has a robust prod-
uct-liability regime used to adapting to technological challenges and
should cope well. ABI/Thatcham (2016) points out that other countries
may need to upgrade their product-liability legislation. In particular:
(i) product liability needs to be compulsory, (ii) liability for personal
injury under product liability needs to standardized as per current auto-
insurance cover, (iii) careful consideration is required as to the duration
of product liability cover, (iv) product liability has to include damage
to the product itself as well as other products, (v) product liability will
276    
M. Naylor

not cover external causes unrelated to the product, (vi) product liability
will not cover if poor maintenance is an issue or sensors have not been
cleaned. ABI/Thatcham thus argues that auto-drive insurance needs to
be covered by separate legislation and regulation to existing product-lia-
bility legislation. There are strong arguments in favor of this idea.
A major complication will come from level 3 or 4 cars where drivers
may switch between fully auto-driving and manual modes. If an auto-drive
car finds itself in a crisis not of its own making, and defaults to the driver
who finds themselves unable to cope with an sudden emergency‚ is the
product or the driver liable? This may come down to—‘would a reason-
able driver have been able to cope?’ ‘Was it reasonable for the AI to default
back?’ As mentioned earlier, it is clear that humans are weak at manually
driving cars, but are terrible at paying attention to what is happening when
the car does 80% of the driving. How much can the product be expected
to ensure the human is paying attention, so they can take over if need
be? If a human being repeatedly refuses to watch the traffic after repeated
warnings‚ is the car entitled to stop and refuse to auto-drive any further?
Therefore, legislation and regulation need to initially make a distinc-
tion between level 3 and 4 cars with advanced driver assistance systems
(ADAS) and level 5 fully automated driver technology (ADT), as only
in the latter case can any fault on the part of the driver be excluded. The
distinction between these two will fade, however, as manufacturers will
steadily upgrade ADAS cars so that they approach ADT capacity in all
but extreme cases.
Product liability will also be a complex area because multiple parties
are involved. Should a particular crash be covered by the car manufac-
turer‚ or the telematic supplier‚ or the AI driving-system supplier‚ or
the tire manufacturer, or the road marker supplier‚ or car maintenance
workshop, or the owner for poor cleaning? Manufacturers will not cover
post sale modifications, or for faulty telematic maintenance.
The defendants in auto-accident lawsuits will change. Rather than
going after drivers and their insurance companies, plaintiff personal
injury attorneys will target vehicle manufacturers, software companies,
producers of vehicle sensors, maintenance services, or even for coun-
cils for not maintaining smart roadside telematics. Accident claims will
grow much more complex, with more subrogation claims, as the parties
try to shift liability to others in the vehicle service stream.
 9  Regulatory and Legal Issues    
277

Note, however, disputes about the facts of a crash will be easier to

solve, because of the large amount of real-time data collected by the car
on its actions, the actions of the other cars, and the environment. Thus,
while some litigation may become more complex, overall a high per-
centage of crashes will be handled by data exchange between the two
insurance company computers.
Adjuster investigations for these crashes and accidents are increasingly
going to shift from disputes about fact controversies, like who really had
the green light or right of way, to disputes about software glitches or
collision avoidance systems and what kind of malfunctions in those sys-
tems caused the collision. The auto-insurance adjuster profession as it
exists today, could thus become obsolete. Rather, there will be product-
liability adjusters who will need a much different skill set than tradi-
tional auto-insurance adjusters possess.

Issues Around Autonomous Drive Programming

There are major as-yet-unsolved legal and ethical problems for insurers,
producers, or programmers of self-drive cars. Many unthought of ethi-
cal issues arise for software coders. For example, how does the software
faced with an unavoidable impact, decide between minimizing harm
to its passengers vs minimizing harm overall, e.g., what does the soft-
ware embedded in the two cars jointly decide about an impeding head-
on crash if one car has one occupant and the other has six? Does the
software avoid the crash by sending the one occupant car off the edge?
What if the car has to choose between hitting two different pedestri-
ans? This is called the ‘trolley problem’ which predates self-drive cars.
We need to teach AI systems some ethics.
Most manufacturers, like Mercedes, have announced that they will
set up their software rules so the car occupant is always given prior-
ity regardless of the level of harm to others. Manufacturers have no
choice, as who would buy an auto-drive car which may decide to sacri-
fice you? This will require regulation setting the parameters for autono-
mous car ethics, even though the issues could get very complex. Future
lawsuits for software producers could be media minefields. Given
that automobile trade is international there is an urgent need for an
278    
M. Naylor

international regulation agreement. In 2016, the Federal Automated

Vehicles Policy started the process by disseminating a set of proactive
safety guidelines to provide safety through autonomous innovation, rec-
ognizing that ‘automated vehicles hold enormous potential benefits for
safety, mobility, and sustainability’ but raise many legal issues.
Issues arise if an unoccupied car is involved in a crash. How do the
police deal with the car? How do the police deal with a driver-less car if
its software is misbehaving? What if a fleet of cars is hacked and used as
a weapon? The answer is that police will obviously need a software over-
ride to control programs. Note that for fully automatic cars, there will be
no need for an occupant to have a driver’s license or be able to drive or
to be sober. This means that police need the ability to correct any issue
with a malfunctioning car and get the occupant to their designation.
Another issue is that auto-drive cars are currently set up to be very
passive in traffic, obeying rules, trying to avoid crashes, and always giv-
ing way. Experience with Google cars indicates that manual drivers
recognize this and learn that if they drive aggressively they can always
bully the Google car to give way.1 This could be frustrating to owners,
so they may start to demand the ability to adjust the car’s passivity set-
ting. Insurers have to be aware of this and adjust premiums accordingly.
Owners who are bullied will have the data to send to police to show
the manual driver’s rule breaking. Regulation may be required around
modifications in this area.
Another major legal issue is the possibility of cars being hacked and
taken over. In theory, someone with knowledge of coding could con-
trol a rich person’s car, lock the doors, drive to a remote location, and
demand a ransom. The perpetrator could be anywhere in the world,
outside domestic jurisdiction. Burglars could also hack car software to
tell if a family is at home.
It needs to be remembered that crashes will still occur with auto-
drive cars as the technology will never be perfect. Even with perfect
programming, self-driving cars rely also on sophisticated hardware.
Improperly maintained equipment, such as mud covering a sensor or a
sensor misaligned by a big bump in the road, could mean incomplete or

1A crash involving a Google car occurred because a bus driver tried to force the car to give way.
 9  Regulatory and Legal Issues    
279

wrong data, and that could cause an accident. Redundant systems and
sensors reduce those odds, but the odds are still not zero.
There are indications that the public may be less forgiving of
these crashes even if the overall crash rate is a lot lower than the manual
drive rate, as increasingly they expect more of software that they do of
humans. This may lead to media controversies and liability issues. Tesla
has already faced several auto-drive law suits.
To enable a self-drive system to work successfully a substantial
amount of data is required. To get to that Level 5 autonomy goal,
manufacturers will collect data as cars drive in level 2 or 3 modes, com-
paring driver reactions to that of machines, known ‘Shadow Mode.’
Currently only Tesla and Progressive have enough data to enable self-
drive at level 4 or 5 by 2020. Insurers who want to survive will have to
negotiate access to these data sets.

References
ABI/Thatcham. (2016). Pathway to Driverless cars: Proposals to support advanced
driver assistance systems and automated vehicle technologies, Association of
British Insurers.
Brookings Institute. (2014). Product Liability for Driverless Cars: Issues and
Guiding Principles for Legislation, Centre for Technology Innovation.
European Commission. (2013). Digital Agenda for the Future: A Europe 2020
Initiative, Brussels, Belgium.
 10
Consequences for Insurance Workers

Introduction
There has been much discussion and fear of the impact of disruptive
technology on the level of employment. A lot of this discussion has,
however, been uninformed. It is useful to note that every technologi-
cal change in history, from the invention of the wheel, to the spinning
machine, to the petrol engine, to the computer, has raised employment
questions, yet there are still enough jobs to go around. During the ini-
tial industrial revolution this was called ‘the machinery question.’
In general, the conclusion from research into past episodes of disrup-
tion finds little net long-term impact on the overall level of employ-
ment. This is because the disruptive changes mean a reduction in the
price of many goods or services and this increases demand for those
goods, and for substitute goods. Overall demand for workers remains
unchanged, with employees in nicer jobs with higher incomes. There
is, however, a distinction between successful survivor companies who
respond to disruption by actively retaining, retraining, and reallocating
staff, and the laggards who respond by reducing cost and retrenching
staff, and thus slowly shrinking into obviation.

© The Author(s) 2017 281

M. Naylor, Insurance Transformed, Palgrave Studies in Financial Services Technology,
DOI 10.1007/978-3-319-63835-5_10
282    
M. Naylor

It needs to be noted that just because a human doing an activity can

be replaced by technology, that doesn’t mean that will be. Five factors
are involved: technical feasibility, costs to automate, relative scarcity,
skills, and cost of the workers who would be replaced. Many activities
which could be replaced won’t be for a while.
The question is more one of transition, as many jobs may have to
be drastically rearranged and some workers may have to change indus-
tries or face substantial reductions in earnings. Not every employee may
have the capacity to make a smooth transition. While the outcome is
speculative for insurance workers, there are several principles which are
of relevance.
Kelly (2016) argues that there is a recurring pattern to worker reac-
tion to the threat of computer replacement;

1. A robot/computer cannot possibly do the tasks I do.

2. [Later]—OK, it can do a lot of those tasks, but it can’t do everything
I do.
3. [Later]—OK, it can do everything I do, except it needs me when it
breaks down, which is often.
4. [Later]—OK, it operates flawlessly on routine tasks, but I need to
train it for new tasks.
5. [Later]—OK, it can have my boring old job, because it is obvious
that was not a job humans were meant to do.
6. [Later]—Wow, now that the robot is doing my old job, my new job
is much more interesting and pays more!
7. [Later]—I am so glad a robot/computer cannot possibly do what I do
now.

Mechanical vs. Analytical

The faster growth in software development than mechanical devel-
opment has complex consequences. In the 1980s, most of the occu-
pational disruptions due to IT innovations occurred in jobs which
involved a definable set of routine procedures, like assembling a car.
These could be easily replaced by a mechanically simple robot obeying
 10  Consequences for Insurance Workers    
283

simple software. Thus, the blue-collar workers were affected more than
the white-collar workers. The increasing use in the immediate future of
smart robots may in contrast be beneficial to Western blue-collar work-
ers as reduced marginal cost and increasing demand for flexibility will
induce producers to switch production back from emerging countries.
Management ranks were until recently, largely left untouched. While
typists were impacted from the 1960s, and book keepers from the
1970s, it was only from about 2000 that software started to affect the
ranks of lower middle managers. Even here, the rule has been that soft-
ware has been focused on the routine and by provision of better infor-
mation has expanded the role of analytics. In general, the overall process
has been good for office workers, making jobs more interesting and bet-
ter paid.
The rise of AI software has recently, however, started to impact on
higher level administrators. In the future, as software tackles ever more
complex tasks, higher and higher levels of management occupation
classes will be affected. For example, IBM has used what it learnt devel-
oping its Jeopardy software to expand into offering legal research and
health diagnosis services. They have started to make this expertise avail-
able to everyone by renting software on-the-cloud by the minute.
Conversely, jobs like office cleaning or gardening, which involve rea-
sonable complex manual dexterity, which robots currently find difficult,
should be replaced at a far slower pace, at least until the cost of renting
autonomous robots reduces below the minimum wage.

Fluff Not Stuff

Given that most physical goods have low-income elasticities of demand,
the world is consuming ever more intangible material rather than physi-
cal, fluff not stuff. Thus, most of the increase in future income will
be reflected in an increase in service employment, especially areas like
health care or entertainment, rather than industrial employment. Jobs
for cafe waitress are probably secure while demand for animators and
YouTube clip producers will boom. Who knew that you could get rich
by home-making short clips about makeup in your own bedroom?
284    
M. Naylor

The USA has been particularly successful at expanding into inno-

vative new services. Progressive Policy Institute (2013) found that the
creation of smart phone Apps had by 2013 resulted in the employment
of over 750,000 extra people in the USA, even though it was only five
years since Apple launched its App store. For countries which support
it, this increase in employment due to software demands will vastly
exceed any losses due to the use of robots.
Note, however, that the drastic reduction in the cost of international
communications means that workers in the West will have to offer
something extra in terms of creativity or cultural connectivity then
workers in Bangalore or Chengdu can offer.

Complexity
The level of difficulty in replacing a human activity by software or
robotics is determined by a mix of the mechanical aspects of the job, the
analytical complexity, and the human interactions. It is easy to under-
stand that truck driving or being a plane pilot has a limited future, as
it is already largely mechanized and involves regular routines. Already
major mines as well as ports are using autonomous vehicle fleets.
Conversely, courier delivering may expand (though the deliverer may
not do much driving), as it involves complex last step deliveries, irregu-
lar routines, and customer interactions. Call center workers have a lim-
ited future, as 90% of calls are routine and voice recognition software
can handle them, without those accent issues, or the need for breaks.
Sales or persuasive calls are unlikely to be automated.
The important aspect is not the complexity of the job, but the pre-
dictability of the activities. Jobs with set rules are fairly simple to auto-
mate from the software angle. The main difficulty involves mechanical
innovations, which are making slower in progress. An area like retail is
more complex. Cashiers are rapidly being replaced in supermarkets, as
are warehouse and shelving staff. ‘Shopping’ itself is, however, changing
from a necessity to an entertainment, with high street shops becoming
demonstration centers linked to resultant online purchases or auto-
mated cashiers. Engaging with customers to motivate them to buy or
 10  Consequences for Insurance Workers    
285

chatting to them in the café queue is unlikely to be automated. A lawyer

may seem to have a complex job, but a lot of legal analysis is finding the
relevant information and then applying well-specified rules, which AI
software tends to do better than humans. From a software angle, the law
firm cleaner’s job involves more complex, unstructured, activities than
the lawyer’s job does. In addition, since the lawyer is paid more, there is
also more incentive to replace them.
Many clerical jobs involve the organization, storing, and retrieving
of information. These are tasks that computers handle particularly well,
and so the demand for clerical labor has fallen. It is useful to note that
while all decision making, and therefore most jobs, involves information
processing of some kind, computers are not better than humans at every
kind of information processing.
Levy and Murnane (2004) draw some useful distinctions between
three types of activities. (i) Simple activities which require ‘rules-based
logic’ can be easily delegated to machines. A problem is broken down
into a sequence of statements, the truth or falsehood of which can be
unambiguously assessed. Many activities using data can be broken
down into a sequence of decisions, culminating in a final decision.
These kinds of tasks have already been largely replaced by software.
(ii) Some activities may seem more complex as they involve ‘pattern rec-
ognition,’ by which is meant the ability to assess a situation and deter-
mine if it fits a previously encountered pattern. Until recently, these
activities were thought to be too complex for computers. However,
AI can handle most of these, as long as the required patterns are clear.
Software only struggles when ambiguity is present. In the work of auto-
mobile mechanics, computer diagnostics can identify many problems
previously encountered by car owners. Novel problems would require
human intervention. (iii) Harder is ‘expert thinking’ which involves
assessing a body of knowledge and building the experience necessary for
specialized pattern recognition. This is unlikely to be impacted in the
medium term.
The key question to be asked is - are changes to the capacity of ana-
lytical software able to be handled by incremental change within exist-
ing management structures or does it involve disruptive change which
requires new management paradigms?
286    
M. Naylor

Because of the expanding ability of software to handle many profes-

sional level occupations, especially those which involve pattern recogni-
tion, Martinez (2016) argues that in the future software will no longer
fill selected hard gaps in a human workflow; instead, humans will fill the
gaps in an IT workflow which computers find hard. Marc Andreessen,
co-creator of Netscape, argues that ‘in the future, there will be two types
of jobs: people who tell computers what to do, and people who are told
by computers what to do.’
How much occupational disruptions will there be elsewhere? The dras-
tic change could be most disruptive in areas which can be handled by
analytical AI software. This can tackle a lot of jobs which now require a
university level education, but involve fewer issues with mechanization.
Even in these very complex cases, because computers will access to an
exponentially growing database of similar cases worldwide and can search
and analyze these faster than humans can, software will become better
at recognizing complex patterns based on inadequate information than
human experts are. Software is making an increasing number of deci-
sions around loans, mortgages, and replacing tax accountants. Examining
x-rays or MRI scans is likely to be increasingly automated, as the latest
software is better than humans. Medical software is now equal to doctors
at diagnosing illnesses and by 2020 will exceed even specialists.
Before the Darpa auto-driving ah-ha moment it was assumed that
only jobs involving repetitive actions were at risk, after the moment it
was obvious that jobs involving complex analysis are at risk. Why can’t
software do 90% of the job of a lawyer or accountant or doctor? The
software industry itself will be disrupted, as programs are proving more
effective and reliable than humans in writing generic code, in debug-
ging, and in creating large programs. A key reason is that humans are
unable to cope with tracing logic strands over millions of lines of code.

Degree of Creativity or Human Interaction

The activities most immune to a takeover by software are those which
require ‘complex communication’ - the ability to convey and understand
nuances and subtle differences in meaning between statements that may
 10  Consequences for Insurance Workers    
287

sound very similar. In face-to-face or telephone communication, this

skill may involve reading meaning into voice tone or body language.
Jobs differ in the extent to which they involve complex social skills.
Humans in general prefer human company. Software is still facing
large challenges in the area of social interactions, with particular dif-
ficulties in recognition of human emotions and the ability to respond
appropriately to these. Software has, for example, so far been unable
to detect sarcasm or irony, though programmers are working on this.
Thus, human interactions with software will seem formal and ‘stiff’
for at least another decade. Occupations where persuasion or selling
or empathy is involved have a high level of required social skills which
are very hard for a computer to replicate, at least for face-to-face com-
munication. Thus, occupations involving high levels of human interac-
tion or persuasion or social intelligence are unlikely to be at immediate
risk.
Occupations involving creativity are also unlikely to be impacted.
Software can only predict based on the data it has and can only be
creative to the extent that it can find hidden trends in huge data
sets. Software cannot predict new products, nor can it assess whether
humans will find a new product ‘cool.’
Thus, the kitchen hand will be replaced but the waiter won’t, and a
barrister is less likely to be replaced by a program than a solicitor is.
This is why an insurance broker or financial adviser is less likely to be
replaced than an underwriter. Personal Assistants are less likely to be
replaced, nor are wise-cracking bar-tenders. While a small but growing
percentage of routine business and sports journalism in the USA is now
computer written, interviewing and investigative or political journalism
will still require people.
In the longer term, however, this will change as digital avatars are
improving in their ability to recognize emotions and respond appro-
priately. IPSoft’s Amelia, for example, can be taught using employee
training manuals and scenario training to continually improve its cli-
ent interactions and is able to handle e-mails, Skype, and phone calls,
recognizing natural languages expressions, searching files for answers,
and knows when to transfer queries which it can’t handle to a human.
She then learns from the advice given by that human. She is able to
288    
M. Naylor

recognize most emotions from faces or text, even analyzing the rhythm
of your key strokes, and is able to respond in the most emotionally
appropriate manner. She is also fluent in 20 languages and works 24/7,
never getting bored by repetitive tasks.

Types of Jobs
Kelly (2016) argues that jobs can be broken down into four categories;

1. Jobs that Humans can do but Machines can do better - Humans can
weave cloth, but automated looms are better. While handmade cloth
is still produced and is prized for its imperfections there is little rea-
son to value imperfections in driving software or X-ray analysis.
Machines improve these products.
2. Jobs Humans can’t do but Machines can - Humans can’t produce com-
puter chips or search the entire web to find a precise page. Most of
the jobs software now does are jobs humans could never have done.
3. Jobs we didn’t know we wanted done - One hundred years ago nobody
asked if they could watch moving pictures while in their carriage, or
if they could find imaginary creatures on a small screen while walk-
ing. Each innovation creates entirely new classes of occupations to
fulfill needs we didn’t know we had.
4. Jobs only Humans can do, at first - The one thing software can’t (cur-
rently) do is decide what humans want to do. Industrialization has
enabled us to feed and clothe ourselves with less than 20% of the
workforce. This has enabled an ever greater proportion of people to
decide to do other things, like musicians, yoga teachers, fan-fiction
authors, makeup bloggers, game developers, night-club reviewers, or
travel bloggers. When each of these tasks is taken by a robot, other
areas will open up.
 10  Consequences for Insurance Workers    
289

Increasing Income Inequality

During the initial stages of the looming revolution, workers whose skills
are in short-supply, like data analysts, may find their income sharply
increasing, while workers unable to adjust may suffer hardship. Given that
a substantial proportion of these affected workers will be middle-income
professionals this has major implications for social stability and politics.
The high level of skill required to take advantage of the new opportu-
nities imply that more developed countries will initially gain more than
less developed and that countries which aggressively welcome the new
changes required by technological innovations will gain a lot more in the
long run than countries which restrict the changes. McKinsey (2015),
however, notes that developing countries will have the opportunity
to leap intermediate stages of technology and not face the problems of
resistance from sectors hurt by the decline of the old. This is similar to
the way many developing countries without landline phones have moved
straight to mobile phones, and have developed more innovative phone-
based payment and other systems than exist in developed markets.

The Overall Job Market

Of course, jobs in general won’t disappear as many commentators’ fear.
Economists call the idea that technological changes mean the disappear-
ance of work in the general the ‘lump-of-labor fallacy’ (which is the idea
that there is only so much work so if machines do more, workers do
less). It is a fallacy because work as such never disappears.
What actually happens is that IT intensive workers will earn more,
and thus spend more, and this extra spending will provide demand for
displaced workers. There will thus be the equally rapid creation of com-
pletely new occupations. Overall there will be increased GDP growth,
and therefore, increased demand for goods and services. In the end,
there will be far more wealth, far higher average wages, and demand for
nearly all workers who want a job. What is more difficult is that some
workers will need to change sectors or change location or be forced to
enter occupations they have no natural talent for.
290    
M. Naylor

A lesson from history is that automation of activities within a sec-

tor does not automatically translate to a lack of demand for workers in
that sector. This is because the sharp drop in the price of the activity
should create a substantial increase in demand for that activity, and thus
a demand for experienced workers to supervise the software.
It is vital to understand that ‘jobs’ don’t disappear, what happens is
that particular work ‘activities’ disappear. An example of this would be
typing, an activity which used to be very common, but which comput-
ers have minimized. Typing, of course, replaced scribes. Yet demand for
clerical workers has remained strong.
Some activities will shrink, some activities will be unaffected, whilst
some activities will grow in demand. The consequence for individual
jobs is dependent on the specific bundle of activities which make up
their work. Things may thus not be so nice for some individual work-
ers. It also needs to be remembered, however, that if costs drop drasti-
cally for a product due to software, then its price will drop, and demand
should rise. If price or income elasticity is high enough, then demand
for workers in a heavily impacted area may well increase.
What is unprecedented about the looming waves of disruptive
innovation is the size and speed of the disruption, as it will occur at
a far faster pace than the two earlier industrial revolutions did. Frey
and Osborne (2013), in an in-depth study, assessed that about 50%
of all jobs are at high risk of being eliminated by 2030, with 30% of
entire occupational classes going the way of typing pools or supermar-
ket cashiers. McKinsey (2016) estimated that by 2025 about 60% of
all current work activities could be replaced by automation with an
average saving in excess of 30%. It is hard to survive in a job when
the cost base of your software competitor is halving every year. The
impact will be felt across society, with white-collar jobs at particular
risk. CBRE/Genesis (2014) states that some experts predict that up
to 50% of current job areas will either disappear or be radically trans-
formed by 2025. McKinsey (2015) argues that up to 45% of all work
activities can be automated right now. CitiGPS (2015) argue that 54%
of finance and insurance jobs are at high risk, while another 17% are
at medium risk.
 10  Consequences for Insurance Workers    
291

In contradiction to these reports, Arntz et al. (2016) focused just

on the impact of movable robots and broke work down into activities
which repetition versus those which involved group work or face-to-face
interaction, and estimated that only about 9% of jobs faced an immedi-
ate risk of more 90% of their activities disappearing. They note how-
ever that jobs may disappear because employers rearrange the remaining
activities into fewer jobs, or decide to forgo the face-to-face interaction
which is routine in some areas, but which customers may choose to
forego. An example of this is the use of grocery self-checkouts.
Overall, workers will have to adjust, moving from their current
area, at which they may be quite skilled and well paid, to a job with
may involve a drastically different skill set, a drastically different work
environment, and possibly substantially reduced pay. Older workers
unwilling to adjust may face early retirement. This is similar to the dis-
appearance of the 1960-style typing pools or phone operators, but at a
greatly increased scale.
The scale of the looming jobs transformation thus may prove very
socially disruptive as required new skill sets may be well outside the dis-
placed workers’ comfort zones. Incoming young employees can expect
to have their role radically transformed every decade or so.
The change may occur faster than workers can be re-skilled or re-edu-
cated. Autor and Dorn (2013) argue that the work force is being polar-
ized with demand for the elite and for manual service workers, and less
demand for the middle. MGI (2013) argues that over the next decade
or so sophisticated algorithms will replace 140 million knowledge work-
ers worldwide. I would argue that what history does show is that even if
skilled workers do lose their jobs, as long as their skills are generalizable,
they are more likely to find re-employment than lower skilled workers.
The revolution is also international, which implies that jobs lost by
a worker in one country may be easier picked by a worker in another
country than a worker in the same country. Worker flexibility and
innovativeness will become a vital attribute to the success of countries.
Countries with inflexible work environments may lose entire industries.
292    
M. Naylor

Lessons from the Professions - the importance

of programmable rules
The high-level jobs most at risk will be in job areas where routines,
complex as they are, can be reduced to a set of programmable rules.
The jobs least at risk will be in areas which involve mental innovation,
or intense human interaction, or which have a wide array of activities.
Stock-market floor traders have largely disappeared, to be replaced by
traders skilled in creating trading programs. Banking has also under-
gone a large transformation in work activities, though far larger changes
are to come. Software is unlikely to replace barristers, as persuasion is
their key skill, though the paper-shuffling juniors who accompany them
may disappear. Solicitors are at high risk.
A large proportion of accountancy work is already disappearing to
software, with successful accountants rapidly moving from routine
account preparation to overall business management advice. Expert esti-
mates are that demand for the activities accountants traditionally do
may drop by 50% or more over the next decade. Survivors will morph
in business consultants and financial advisers. Surgeons are beginning
to be replaced by far more precise robots. The work of general doctors
is likely to be replaced by computerized expert diagnosis systems, with
estimates of up to an 80% reduction in demand. Survivors will focus
on offering sympathy and psychological aspects of their service. They
may, however, find that a skilled nurse can operate the diagnosis system
at a far lower cost and will thus screen clients, leaving only the more dif-
ficult cases or client reassurance to the doctor. Software is more likely to
replace a highly skilled radiologist, who does a narrow range of routine
activities, rather than their PA, who does a wide array of activities.
It is important to understand that impacts can be complex. For exam-
ple, Markoff in 2011 found that law firms faced with legal analysis soft-
ware initally drastically reduced their demand for law clerks, as the software
could far more effectively search case databases using language analysis to
identify general concepts, during a better job at a fraction of the cost, and
with no risk of getting bored and missing things. The increasing language
ability of software even allows them to read documents for tone and spot
unusual conversations. However, the drastic drop in the price of ‘discovery’
 10  Consequences for Insurance Workers    
293

led to a substantial increase in client demand for it, and since the software
still needs to be supervised, demand for law clerks is currently actually
rising by 1.1% per annum. I would argue that a similar trend is likely to
occur for accountants, if they creatively use the increased richness of data.
University academics are at risk in areas of routine teaching and mark-
ing, with software being able to adjust the teaching pace and amount of
problems asked based on student feedback and performance. Conversely,
academics capable of innovative research are unlikely to be replaced.
The Economist (2016) pointed out that in the USA there has been a
trend since 1990 for non-routine jobs to grow in demand while routine
jobs have stagnated. In the future, routine jobs should start to shrink.
The issue is that it is nearly impossible to predict what types of jobs will
be created though we can be certain that many new types of jobs will be
created. After all, who would have predicted 20 years ago that you could
make a living being a professional video game player or YouTube star?
Careful assessment of internal insurer administrative functions in
terms of which exhibit the type of programmable rules required for
the role to be capable of being handled by software in the short-term
and artificial intelligence (AI) systems in the long-term, indicate that
approximately 60% of all insurance administrative functions are poten-
tially replaceable by software within the next five years and up to 80%
within the next decade. The key point is that if a function is program-
mable and costs of software are dropping by at least 10–15% per year
then that function will eventually be handled by software.
It is important to note that by ‘programmable’ I don’t just mean able
to be written into a codeable preset sequence of steps. Modern advances
in Artificial Intelligence mean that software can learn by doing, so that
even activities which seem complex and require intuition can now
be classified as programmable if an AI system can be correctly guided
through a learning process. Amazon has shown that AI systems which
control networks of machines are generally more efficient than average
humans and the efficiency gap is increasingly increasing. An important
new area is the development of natural language capacity so that answer-
ing phones, writing reports, or creating video presentations are now
areas which are programmable. This implies that professions may be
more at risk of disruption in the future than the average low-wage job.
294    
M. Naylor

McKinsey (2015) argues that ‘steeper declines will occur in more satu-
rated markets, products with declining business volumes, and, of course,
the more predictable and repeatable positions, including those in IT. The
broader corporate functions including these roles will lose jobs overall.
Other roles, however, will experience a net gain in numbers, especially
those concentrating on tasks with a higher value added. Some activities
will be engines of job creation—these include marketing and sales sup-
port for digital channels and newly created analytics teams tasked with
detecting fraud, creating ‘next best’ offers, and smart claims avoidance’.
John Cusano, in Accenture (2013), argues that ‘technology will change
many insurance jobs and even eliminate some. But a much bigger trend is
the support that technology will provide, enabling professionals to spend
less time on routine administrative tasks and more time adding value;
engaging with customers, solving problems, making better decisions.’
To meet these challenges, insurers will need to source, develop, and
retain workers with skills in areas such as advanced analytics and agile soft-
ware development; experience in emerging and web-based technologies;
and the ability to translate such capabilities into customer-minded and
business-relevant conclusions and results. McKinsey (2015) argues that
there will be a 25% drop in insurance jobs overall by 2025, with minimal
change in product development and sales support, a minimal drop in IT, a
25% drop in operations staff and a 45% drop in administration staff.
I would argue that an overall drop in insurance sector employment is
unlikely, as new opportunities at successful insurers will grow fast. It needs
to be remembered that during the initial industrial revolution employ-
ment in the hardest hit sector, weaving, quadruped between 1830 and
1900, despite mechanization. This is because the cost of cloth dropped by
98% so demand within the clothing sector boomed. Where insurers may
struggle is in attracting sufficient data-skilled employees, as competition
for STEM graduates is exceeding supply and surveys show that less than
3% of graduates regard insurance as an attractive sector. Insurers are not
seen as offering engaging activities or being ‘cool employers’.
One of the differentiating factors between insurance incumbents who
will survive and those who wouldn’t is that the non-survivors will focus
on cutting costs by reducing staff, while the survivors will focus on add-
ing value by reallocating staff.
 10  Consequences for Insurance Workers    
295

References
Accenture. (2013). Closing the gap: How tech-savvy advisors can regain inves-
tor trust.
Arntz, M., Gregory, T., & Zierahn, U. (2016). The risk of automation for jobs
in OECD countries: A comparative analysis. OECD Social, Employment,
and Migration (Working Papers, #189), OECD.
Autor, D & Dorn, D. (2013). ‘The growth of low skilled service jobs and the
polarization of the US labor market’, American Economic Review, 103(5):
1553–1597.
CBRE/Genesis. (2014). Fast forward 2030: The future of work and the
workplace.
Citi GPS. (2015). Technology at Work: The future of Innovation and
Employment. Citi GPS Global Perspectives and Solutions.
Economist. (2016). The Return of the Machinery Question. Special report, June
25, UK.
Frey, C. B., & Osborne, M. A. (2013). The future of employment: How suscepti-
ble are jobs to computerisation? (Working Paper). UK: Oxford University.
Kelly, K. (2016). The inevitable: Understanding the 12 technological forces that
will shape our future. USA: Viking.
Levy, F., & Murnane, R. J. (2004). The new division of labor: How computers
are creating the new job market. USA: Princeton University Press.
Markoff, J. (2011, March 4). Armies of expensive lawyers replaced by cheaper
software. New York Times.
Martinez, A. G. (2016). Chaos monkeys: Inside the Silicon Valley money machine.
London, UK: Ebury Press.
McKinsey. (2015). Four fundamentals of workplace automation. McKinsey
Quarterly, November.
McKinsey. (2016). Automating the insurance industry. McKinsey Quarterly,
January.
MGI. (2013). Disruptive technologies: Advances that will transform life, business
and the global economy. Tech. Rep. McKinsey Global Institute.
Progressive Policy Institute. (2013). 752,000 App Economy jobs on the 5th anni-
versary of the App Store. www.progressivepolicy.org.
 11
Impacted Occupations

Actuaries and Underwriters
Two of the occupations likely to be immediately impacted are ­actuaries
and underwriters. Actuaries will find 95% of their training on small
data increasingly irrelevant and will have to rapidly retrain in big data
skills. In this area, they will face sharp competition from graduates in
the new Masters in Data Analytics programs currently being launched.
Underwriters have been listed in most reports as one of the ‘ten
likely-to-be-impacted occupation classes.’ This is because the job has
traditionally been composed of analyzing data for trends which may
impact on risk levels, selecting key indicators, and matching these to
the selected characteristics of applicants for cover. These are routine
and rule-based so involve areas which software can easily replace. The
need to underwrite all clients and customize at a minimal cost means
that humans cannot be involved. Automated underwriting software is
already being extensively introduced, as noted in SwissRe (2013).
It is important to note, however, that things are not this simple, as
decreasing demand for the current work done by underwriters does
not automatically translate to a lack of demand for underwriters.

© The Author(s) 2017 297

M. Naylor, Insurance Transformed, Palgrave Studies in Financial Services Technology,
DOI 10.1007/978-3-319-63835-5_11
298    
M. Naylor

This is because the rapidly decreasing cost of using software to under-

write clients means that the amount of underwriting undertaken will
explode, as nearly all clients will be underwritten, and in a more inten-
sive fashion. The new style underwriting will become dynamic and real
time so that rates will change based on client activity rather than just
the rate just being set at contract creation, so that as the actions of cli-
ents change, they can be moved between risk classes and differing pre-
miums charged for differing activities. These considerations imply that
there will remain amble jobs in the area of underwriting.
The employment issue is that these new underwriting jobs will
demand strong computing/mathematics/AI skills as underwriters will
generally not directly underwrite but will be involved in overseeing and
improving the underwriting programs, as well as setting the rules for
when software marks a potential client as unusual and thus requiring
human consideration. New style underwriters will thus have to under-
stand big data and algorithms, as well as skills in creating user-friendly
interfaces, and display the creativity to use the ever-expanding richness
of data in ways which are novel and client-relevant. Whether individuals
currently employed as underwriters can make the transition is a moot
point. I would argue that because underwriters have experience with
customers and with products, it is essential that they are available to
guide new big data graduates in creating useful insights from algorithms.
Aviva UK, by contrast, has substantially reduced the u ­ nderwriters it
employs and increased data analysts.

Administrative Staff
One of the first areas to be impacted will be administrative systems,
with the administration functions of insurers likely to be severely
impacted. By 2025, a substantial proportion of insurance administra-
tion staff are likely to find the activities they carry out in their jobs
drastically changed or gone, as the rapidly dropping costs of computeri-
zation impacts on administration processes.
The best estimates are that 95% of current administrative activities
within insurance companies follow programmable rules and therefore
 11  Impacted Occupations    
299

will ultimately be handled by software. It is not a matter of going digital

or going paperless, it is creating integrated end-to-end software systems
capable of handling all client requests at minimal cost for all but the
most complex of client situations. Analysis shows that about 70% of all
current job activities in insurance administration will undergo profound
disruption, with these staff expected to adapt to fundamentally different
roles and expectations.
Note that this is best examined in terms of ‘activities’ rather than
‘jobs,’ with employees seen as performing a range of activities, which
will each be disrupted at differing times to differing degrees. Thus, soft-
ware will not replace jobs but will replace some activities. Insurers will
have to continually rearrange employee roles to expand newer activities
and reduce shrinking activities.
However, again, this does not mean that these workers will face
unemployment, as these systems will still need to be supervised, and
demand in other areas of work will open up, for example engaging in
social media. After all, the disappearance of typing pools has not led
to a decrease in office employment, and there has been an increase in
banking sector employment as computerization has spread. The abil-
ity of employees to interact with and oversee software processes will be
vital, as will the ability to interact with big data and social media.
This disruption will include a number of fairly senior management
jobs, and not just junior positions, as many senior roles involve analyti-
cal activities which can be replaced by AI agents. Morgan Stanley/BCG
(2014) estimates that automated administrative processes alone will
cut insurer costs and thus premiums by around 25% by 2019, while at
the same time as speeding up response times to clients. The impact is
likely to be closer to 60% by 2025. Note that this doesn’t mean elimi-
nating human response as most digital natives expect the availability of
humans or advisers as one of the many channels available.
Many of the surviving administration jobs will involve a high level
of interaction with social media or CRM systems, as insurers need to
evolve a deep understanding of digital natives and social capital, and
then either create a dedicated social media presence and strategy or
quickly become irrelevant. Given the demand for rapid 24/7 responses
and intense monitoring of social media, insurers will not be able to
300    
M. Naylor

reduce response staff by the same proportion as customer calls reduce.

The nature of those jobs will, however, fundamentally change as most
insurer staff will have to work closely with software, so that extensive
retraining to a higher level of technical skill will be required.
Most industry observers argue that jobs which depend on analyzing
statistical trends will die, as they will be replaced by expert system AI
programs. CBRE/Genesis (2014) thus lists underwriters, fraud asses-
sors, claims clerks, and appraisers, as among the top 20 most-at-risk of
all occupations in the next few years. Even software coding jobs may
shrink as high-level software takes over.
I would argue that these expectations are wrong as a more astute
observation would that the nature of insurance data jobs will drastically
transform, with a rapid expansion of analytical jobs overseeing AI pro-
grams as well as dealing with the resultant outputs. The range of new
insights available will actually increase demand for analytical activities.
There will have to be a corresponding expansion in strategy jobs, in
terms of trying to understand the insights from the rapidly expanding
big data systems and the application of these, especially roles involved
in converting the new insights into profitable services. This will imply
faster revision of IT-based administration systems, and a large expansion
of software related jobs.
One of the largest changes will thus be the heavy increase in the need
for regular staff retraining. The idea of staff being trained in higher edu-
cation and then being suitable for a job for life is already archaic. Staff
can be expected to experience a drastic upheaval of their role at least
once a decade. The newness of most of the challenges means that there
will be a scarcity of already trained staff. This implies that employers
will need to allocate substantially increased resources to staff training
and an expansion in training roles. The survivors will be those compa-
nies which have deep and wide continuous retraining programs. Even
new graduates will need to be retrained within the first decade as their
knowledge will date fast. The ability to learn will be more important
than deep knowledge.
If the insurance employer responds to the increasing opportunities to
engage with clients, then the increased demand for staff in the expand-
ing areas will mean that overall employment may increase rather than
 11  Impacted Occupations    
301

shrink. Insurers who proactively move from product selling to services

supply should employ more staff. It would be a major mistake for insur-
ers to view the digital transformation as an exercise to reduce staff num-
bers, rather than a process of transforming staff skill sets. Incumbents
need to remember, however, that the new entrants will have a need for
experienced staff and will try to attract the most competent employees
with a more exciting work environment.
The key issue for staff is thus their capacity to adjust to a new job
and new skill sets. Those staff who are flexible enough to pick up new
skill sets in new areas as they are invented will thrive. Given the pace of
change, employees can expect a substantial change in their ­employment
activities at least once a decade. Those staff who are incapable of adapt-
ing will lose their jobs. Employers who do not ensure their staff feel
confident enough to retrain will face bankruptcy.

Robo-advice and Brokers/Advisers
Overview

SEI (2015) claims that software innovations aimed at the financial and
insurance planning/advice industry are now being created faster than
the industry can adapt to. The limiting factor in technological change
for advisers/brokers is thus now the ability of advisers/brokers to adopt
their culture to use the innovations. Firms which adopt faster will have
an increasing competitive advantage.
Arthofer (2016) argues that insurance agents/brokers are uniquely
vulnerable to tech-based external entrants because (i) they are aging and
thus less in touch with modern customer demands, (ii) the next wave of
insurance customers will be millennials, who expect an omni-channel
response, (iii) entrants have an integrated, front-end-to-back-office IT
system, (iv) customers are being expected to take more responsibility
for their purchases, which they do by going online, (v) goods ­similar to
insurance are already sold online in many markets, (vi) online d­ isruptors
provide a better education experience, and (vii) online disruptors use
302    
M. Naylor

their CRM systems to provide superior customer service. Arthofer

argues that most of the older advisers/brokers will be unable to adjust to
this radically different world.
What does this mean for financial and insurance brokers and advis-
ers? Lessons can be taken from other professions are illustrative.
Financial market traders now do very little actual trading, and floor
traders have disappeared. Market specialists instead supervise high-
speed trading programs. Since this requires IT skills, a substantial num-
ber of previously highly paid traders have been forced to find other
occupations.
Similarly, modern accounting has very little to do with creating and
analyzing accounts, as the routine inputting of data and analysis of the
output is now done by software. Accountants in general have responded
successfully to the IT challenge by moving into offering general business
advice, learning to use the increased data analysis capacity to enhance
their capacity to examine business trends, and give sophisticated advice.
Accountants have not had to become skilled in IT as the software is user
friendly.
Lawyers have yet to be affected as strongly but Markoff (2011) argues
that since most legal advice involves knowing existing law and analyzing
these to find patterns, the common lawyer is in trouble as 80% of legal
advice could be provided by software. Lawyers will have to respond like
the accountants and shift into using the greater data analysis capacity to
offer substantially more sophisticated advice built on the insights deliv-
ered by the software.
Buyers of financial/insurance products who wish to undertake sim-
ple purchases or receive generic advice will get this online from well-
structured Web sites, especially ones which offer computerized feedback
to client questions via verbal or facial response systems. Therefore, bro-
kers or advisers who want to survive will have to differentiate themselves
from what software systems can offer. In particular, they will have to
understand which aspects of their advice follows programmable rules,
and can therefore be replaced by software, and which aspects are too
complex or psychological to be replaceable. Involvement in activities
which create client trust is the most important concept.
 11  Impacted Occupations    
303

The current state of the broking/advice industry in terms of the IT

transformation is however lagging as badly as insurers are. For example,
most brokers use their Web page merely for quotes and purchases. The
majority do not understand the level of sophistication in adviser-based
software response which is required. SEI (2015) found that financial
and wealth advisers are currently underestimating the potential impact
of AI systems, with very few comprehending the ability of future AI
systems to offer competent generic advice to 90% of investors. This is
because they are evaluating the current rough state of robo-advice soft-
ware rather than the rapid exponential pace of future developments. In
terms of Figs. 1.1 and 1.2, in the introduction, adviser software is not
yet at the inflection point, so while we are disappointed by the relatively
slow pace of current developments, we will be shocked by the rapid
pace of future changes. The industry needs to prepare for the high qual-
ity of what is to come, rather than guessing forward based on what cur-
rently exists. It may be a consolation for insurance advisers to know that
while their sector needs change, the extent of the required change is less
than that required for insurance companies.
Increasingly sophisticated software using voice recognition will
be able to offer generic advice and answer routine questions over the
phone or a Web site without clients being aware they are talking to soft-
ware. These programs will even be able to cold call prospects, and offer
responsive virtual-reality ‘talking heads,’ so that Skype calls can be done
by software. Obviously, the market share of direct sales Web sites will
grow. Most routine clients will thus have little need to talk to advis-
ers, or pay their fees. An area where skilled software will be vital is in
handling social media, as brokers and advisers will not have the time to
intensively monitor all their customer channels. Therefore, queries and
comments will need to be first screened by software and answered by
software, with pre-set exceptions passed on to the adviser.
One downside for the established western firms is that program-
mers in Bangalore could become a major player in the lower-end advice
market. Another downside is that technology firms will be able to
­
produce advanced AI software and enter the market. Thus, a substan-
tial share of the providers of robo-advice will probably be new market
entrants, firms with an IT background, who can easily program a set of
304    
M. Naylor

rules based on an extensive questionnaire. As costs drop, banks will find

it easier to install an AI booth than retain branch advisers. Since banks
tend to offer more rule-based advice than personalized advisers do, they
may find their clients more susceptible to the IT competition.

Key Concepts for Adviser Survival

The first key concept for adviser survival is that in a well-run adviser
practice, the adviser will have already delegated the routine tasks, like
answering clients phone enquires, or tracking transactions, or calculat-
ing sums producing reports, to lower paid assistants. They will have left
themselves the more productive, more skilled, more interactive areas,
like meeting clients and convincing them of the importance of follow-
ing the adviser’s recommendations using persuasive skills.
Thus, the adviser’s staff who are currently involved in areas involv-
ing routines which follow programmable rules are more likely to be
replaced by software than is the senior adviser. The Adviser can use
quality admin software from 3rd party platforms to automate most of
their administrative tasks, including routine client contact and regula-
tion requirements. This should enable them to cut per-client cost by
60–80% while increasing service quality because they are freed from
routine administration, therefore increasing the frequency of mean-
ingful engagements with clients. The upside for adviser survival is that
it will be easier for smaller practices to survive by replacing admin or
reception staff with software. Most routine regulatory activities can be
handled by software. The key here is an integrated end-to-end IT sys-
tem which integrates everything which happens in the adviser office so
each staff member has what they need on screen, so clients feel their
queries are intelligently answered by lower paid staff, or can interact
via the Web site. Reducing per-transaction cost, increasing customer
engagement, and limiting adviser time to high-value activities are keys.
A second key concept is that while previously customers would turn up
not knowing much about products or prices, and therefore needed edu-
cating, increasingly new customers have done extensive online research
and so know a lot more about products and prices and so are more
 11  Impacted Occupations    
305

focused on the advice aspect - the ‘how do I get to where I want to be’
aspect rather than the ‘what can you offer me’ aspect. Hard-sell skill sets
are becoming less valuable than softer persuasion skills. The focus will
be on lifecycle advice, not product sales.
A third key concept is that as insurance becomes telematic and data
intensive as well as embedded in an ecosystem, advisers will have a role
in servicing those telematic devices, in explaining to clients how they
can use feedback from telematics to reduce dynamic commissions, and
in cross selling additional products or services from related ecosystem
providers. Advisers are used to the idea that they offer services and not
just products. The key strength the adviser has to offer the ecosystem is
their close and personal contacts with clients, and the associated abil-
ity to influence them and engage the client in an emotionally success-
ful life-process. Data analytics will offer detailed feedback on the best
advice or sales techniques. This will lead to far higher persistence rates
for adviser clients. Note that the adviser’s IT system will need to work
with clients’ data agent Apps.
A fourth key concept is that the successful survivors of the indus-
try during the coming IT revolution will be those who ensure finan-
cial planning is about personalized advice and associated interpersonal
skills, and not about sales or transactions or standardized advice. This
is because generic advice follows programmable rules. Aspects of the
job which depend on personality skills will survive better than those
aspects which depend on mathematical skills. Successful advisers will
proactively embrace software advice systems to enhance their advice
and their capacity to deal with clients by offering better service. It is
well established that the key part of being a successful financial or insur-
ance adviser is not the giving of the advice, but ensuring that clients
are ­convinced to actually make the sacrifices required to implement the
advice - the human psychological skills.
A fifth key concept is that AIA/Beddoes (2015) shows that ­insurance
advisers demonstrating that they can manage claims by ensuring
­hassle-free reasonable settlements is vital to ensuring client retention.
Since many claims take place at a time when clients are less able to self-
manage, advisers are able to add a personal touch which will result in
multiple referrals - good advisers’ claims management is much more
306    
M. Naylor

than just money, it is also dealing with all the other aspects of the
­clients’ financial affairs. Since many brokers and advisers don’t a­ dvertise
this area of their work, new clients often undervalue this advantage
of advised insurance. Good advisers will ensure all elements of the
­insurance chain run smoothly. Successful claims handling is impressive
to ­clients and should generate strong social media ratings.
A sixth key concept is that advisers’ currently only service about the
top 10% of households by income. The reduction in per-client service
costs produced by adviser software will enable advisers to expand their
offering down into the middle-income brackets. These lower wealth cli-
ents who are currently not serviced due to the lack of fees generated, but
can be if per-client costs are slashed. This will expand adviser’s potential
market size at least 3 or 4 times. This can be done by using a predomi-
nately robo-advice approach which is integrated with restricted human-
delivered advice. By working alongside quality robo-advice software,
advisers will be able to offer a low-touch version of personalized service
which is now only available to the richest clients to these middle-income
clients. The high-net-worth clients, whose financial lives have com-
plexities not amendable to programmable rules, will still prefer human
financial advisers and can now be offered a high-touch, high-tech
approach.
It will also be easier for advisers to segment clients by costs and prof-
its, and offer differentiated services. Fidelity’s 2012 annual broker and
adviser sentiment survey showed that tech-savvy advisers managed on
average $8M more in funds than their colleagues. This difference will
exponentially grow as more tech-savvy advisers offer a more inclusive
and dynamic experience to an increasingly diverse array of clients, while
the less tech-savvy collapse under a blizzard of admin and regulation
requirements.
A seventh key concept is that the cost of services like portfolio selection
or investment selection or policy application will drop more than other
areas. Investment-only advisers will be impacted more than full-service
advisers.
An eighth key concept is that clients will get comprehensive informa-
tion on adviser costs and will be better able to hold them accountable.
Clients will see all their accounts in one place and will see all load fees
 11  Impacted Occupations    
307

and be able to compare high-cost mutual funds with low-cost index

funds under a range of scenarios. These costs can be then used to evalu-
ate the worth of an adviser. Quality advisers not tied to a one-product
supplier will do better than chain advisers
A ninth key concept is that a quality customer relationship man-
agement (CRM) system is vital. There is a need to integrate multi-
ple ­communication channels in a manner which is easy to use for both
­customers and advisers. For example, some customers may prefer ­contact
via text messages, and these need to appear on one system alongside
e-mail or phone messages, so the adviser does not have to be constantly
checking multiple sources. Thus, the CRM system needs to cover: per-
sonal interviews, phone calls, e-mails, Web site self-service, Web site life-
chat, Skype, mobile Apps, text messaging, online forums, short videos,
and social media, and it needs to collect information and provide sum-
maries automatically via key word searches. The CRM then needs to out-
put the adviser’s response in the customer’s preferred channel. What the
adviser wants is to get their message to as many people as possible, for the
lowest cost per client, and yet make the client feel like they are getting
quick and customized attention, in the way they want to communicate.
There also needs to be an effective interchange of information
between staff so the lower cost assistants or software applications can
do their job, without constant adviser oversight. The CRM should give
feedback to the adviser on which channels specific types of custom-
ers prefer, why, and alert the adviser to evolving trends. The adviser
can experiment with different client approaches and see which results
in the best client outcome. The CRM needs to be open to review by
­customers and improvement suggestions. Anticipating customer needs
and ­proactively addressing them are a vital part of a superior customer
experience. Software providers will need to supply insights gained from
the AI analysis of CRM outcomes across their network.
A tenth key concept is the need to create multi-generational client and
adviser strategies, with differentiated channels, so that the children of
current clients can be retained. SEI (2015) reports that less than 6% of
US wealth advisers have a multi-generational strategy, resulting in 94%
of firms having a dubious long-term future. A key part of this strategy is
nurturing the next generation of advisers. Note that the style and ethnic/
308    
M. Naylor

gender composition of these advisers will have to match the increasing

diversity of the younger generations. An upside is that today’s recruits are
the first generation of advisers to start out as trained financial planners
rather than being converted from sales-related occupations. The CFP
qualification is an essential first step, with a specialist degree increasing
common. Being tech-savvy will be a required basic competency.
An eleventh key concept is that while earlier generations of advisers
have had to deal with inefficient and non-integrated adviser admin soft-
ware, the new generation will benefit from high quality modular 3rd
party software, which is largely bug-free and adviser/client friendly. The
adviser will be able to have this personalized to their practice very easily.
Advisers, however, will have to get used to frequent upgrades and pres-
entation changes in their software, instead of the current tendency for
advisers to stick with the same package for up to a decade.
The adviser will have to offer multiple channels to clients, as differ-
ent aged clients may have differing preferences in terms of channels.
Younger clients, for example, may prefer more software-based systems.
A key here will be to ensure that even 100% digital clients get some
human contact so a relationship is built up, and as wealth builds they
get used to human-based advice. Advisers can’t assume that they can sit
back and wait for younger digital clients to ‘mature’ into behaving as
older clients do - they will have to be cultivated.
The heart of this software will be ensuring that the client’s goals
are clear and that clients receive personalized feedback about progress
toward achieving them. This has to include feedback based on behav-
ioral finance concepts using real time telematic data-flow and spending
as an integrated aspect of encouraging good client behavior. This should
ensure deep and real-time client engagement.
Omni-channel access will allow advisers and clients to work together
to establish goals and mechanisms, which is called ‘co-planning.’
Currently, however, less than 10% of US wealth managers offer an
effective client portal which is omni-channel, integrated, personalized,
offers collaborative real-time document tools, and is linked to real mar-
ket data. Very few offer tools such as allowing clients to actively stress
test their current and planned portfolios under various potential market
scenarios, or run insurable risk scenarios. All of these areas are creatable
 11  Impacted Occupations    
309

by a quality software provider, and will be part of superior robo-advice

packages within the next decade, so advisers who do not regularly
upgrade their CRM systems will be at risk of being ‘ubbered.’
A twelfth key concept is that passive investing will increasingly dominate
over active. The cost of conducting transaction will also drop by at least
100-fold. Thus, advisers who rely on transaction-related fees for profit will
struggle, while advisers who focus on client-behavior-based fees will survive.
A thirteenth, and potentially most important, concept is that current
robo-advice solutions tend to assume the problem is one of informa-
tion, so focus on providing this more efficiently - ‘if we have more
knowledge and better tools we will all achieve financial success.’
However, as any adviser knows, the major problems clients have are
actually behavioral. Stage 3 or 4 robo-advisers will start to incorporate
knowledge in these areas, including real-time feedback to investors
based on telematics or spending. Because investors may feel embarrassed
to share failing in this areas with humans they may accept this feedback
better from software. Advisers will have to embed these systems in a way
which assures clients that their failings are confidential.
A fourteenth key concept is that it is faulty to think there is a dichot-
omous choice between human and robo-advice. Successful advisers
will use a multiple of customer contact channels, human/Web/video/
phone/robo, etc., with the choice partly determined by the adviser on a
revenue-cost basis, and partly determined by clients. Advisers, for exam-
ple, may provide generic information via robo-channels once the client
has been motivated by human contact, or may use a vibrant robo-
web-presentation to increase client numbers. A robo-adviser may help
to do the initial screening for return-risk profiling, with a voice acti-
vated talking-head taking the client through a questionnaire, with the
capacity to answer additional client queries. Given that portfolio con-
struction is increasingly being automated, this could link easily into
suggestive portfolios, and then book an adviser meeting.
The switch between channels does not have to be linear, as clients
can be given the choice at each stage to switch to a human adviser via
a Skype or chat box. Robo-software can analyze client voice patterns
in response to questions, to determine stress and comfort, and alert a
human at preset levels. A human adviser would oversee the robo-derived
310    
M. Naylor

choices and intervene at key stages. The key is for the adviser to strategi-
cally set up an advice process which minimizes cost, maximizes client
engagement, and optimizes client outcomes. The client interface has to
be seamless, high quality, and engaging.
A fifteenth key concept is that advisers have generally done a better job
of creating close customer relationships than insurers have. Accenture
(2013) found that 60% of surveyed US advisers have at least weekly
contact with clients through social media. Advisers have been particu-
larly active in use of social media. This puts them in a sound position
to offer their expertise in this area to external entrants as well as existing
suppliers. Certainly, quality advisers have a far smaller transformation to
make than insurance companies do.
Accenture, however, also argues that there are two communica-
tion issues, (i) advisers tend to seriously overestimate client investment
knowledge so that what advisers regarded as clear communication was
perceived by clients as promotional, and (ii) advisers tend to overesti-
mate clients’ willingness to take investment risk. These issues create a
channel for advisers to offer useful education via social media which will
be welcomed by clients. Contacts with clients need to be two-way and
group based, so advisers can answer questions.
A sixteenth key concept is that the idea that ‘millennials and digital
natives do not value advisers and will only purchase on-line’ is faulty.
Evidence clearly shows that the digital savvy are happy to talk to human
advisers, and are happier to pay for this than baby boomers if they are
approached correctly. The issue has been that advisers in general are
used to dealing with baby boomers and do not know how to relate to
millennials and digital natives. The new generations want the same
thing the baby boomers want - a relationship with a competent profes-
sional they can trust. They just want to build it on a computer first. The
older generations can often be quite tech-savvy.
However, younger clients have different customer experience expec-
tations - in particular their activities with the online world mean that
they expect excellent service at a low price with extremely fast delivery,
via a channel of their choice. Their first impulse is not to phone to talk
to a person, but to do extensive net searches first, including reviews of
the adviser on social media. Cold callings, direct marketing, and refer-
rals are uncommon. Online articles and blogs are highly regarded.
 11  Impacted Occupations    
311

Education and client stories should be a large part of that. Advisers who
do not work in that space and do not offer open two-way reviews, or
generate numerous online reviews, do not exist in the new world. If an
adviser’s business is not visible on a millennial’s mobile phone, it doesn’t
exist. The methods advisers have successfully used to attract baby boom-
ers will not work.
Zaptitude (2016) surveyed millennials in Australia and found that
they felt disengaged from the financial service industry, especially advis-
ers. Millennials felt that advisers ‘spoke an alien language’ which patron-
ized them and pushed them away. Zaptitude found that the most
important qualities desired in an adviser by millennials were ‘a sense
of humor’ (68%), ‘simplifies the complex’ (66%), and ‘casual/relaxed’
(48%). ‘Technical/detailed’ was only at 29% and ‘wearing a suit’ at 26%.
Zaptitude argues that ‘financial service professionals need to stop assum-
ing they know what is best for the public and start to ask them instead.’
For Millennials and digital natives, direct quotation firms are the
only ones to vaguely appeal to the new generations; yet these platforms
only generate a low-quality relationship, use insurance as a commodity,
and therefore not do have sticky customers. They can be easily bested by
quality advisers.
Advisers need to avoid the commodity-price space and create a deep
and rich omni-channel experience based on quality advice and quality
communication. Part of this is not using a simple price quotation sys-
tem, as this ignores quality issues. Adviser Web sites should start with
an advice-based questionnaire to ascertain client needs and therefore the
most appropriate range of products, before price is mentioned. Advisers
need to mention products the client doesn’t need or how to cut cost and
reduce risk. This is also vital as it establishes a trust relationship. Advisers,
who are skilled at relationships, will find it easier to establish these omni-
channel experiences than large transaction-based insurers will.

Change in Activities

There is thus no reason to expect an overall substantial reduction in the

size of the adviser/broker industry as long as the industry is prepared
to respond proactively, and understands how to respond strategically to
312    
M. Naylor

the rise of robo-advice. Effective use of technology will allow advisers to

source clients from a wider geographical region, or even internationally,
and thus specialize in a market segment.
In fact, as internal software allows advisers to service clients at a low
marginal cost and therefore at a lower price, overall demand for advice
should increase substantially. One key is to ensure that younger, lower
income, clients who will initially use generic robo-advisers, get used to
being advised and therefore switch to human advisers as their wealth
rises.
The aspect of financial/insurance planning which will be most
­difficult for programmers to replace will be the client psychology aspect,
the understanding of client weaknesses, the understanding of how to
persuade clients to act in their own long-term interests, and how to
help them overcome those weaknesses in order to become financially
­successful. It is well established that most clients need external prodding
before they will act in their own best financial interest. The key element
of being a professional financial adviser is not providing a plan, but in
ensuring that clients have the willingness and motivation to carry it out.
Robo-advice software is unlikely to be able to use psychological
knowledge to provide this style of advice in the intermediate future.
Therefore, advisers will have to place a premium on interpersonal or
creative or abstract-thinking skills. Advisers will have to define very
carefully how their advice is different to, and superior to, the standard,
generic, advice produced by AI software.
It would be a mistake, however, to assume that AI systems will not in
the future be able to handle many of the psychological aspects of advice
once talking heads are available. Advisers will have to run to stay ahead
of an increasingly fast improving competitor.
Successful advisers will have to structure their processes so they are
able to attract clients away from software systems, as younger investors
will start by using software advice at the stage when they have limited
investible funds. These investors may then have a tendency to stick with
those systems, even when they reach the stage when they have enough
investible funds to be attractive to advisers, normally at about $100,000
in funds, as they will find it a cultural change to switch from a software
 11  Impacted Occupations    
313

adviser to a human one. If advisers attract younger clients by offering

a free robo-advice platform with generic advice, then they can increas-
ingly offer users fee-based human advice as their wealth grows.
Current advisers who effectively use technology with an advanced
data presentation system have found that clients understand their
­financial situation better and trust the adviser’s advice more. Clients feel
more ownership of the advice, and are thus more likely to implement
it, because they have the opportunity to examine their own financial
progress rather than the adviser lecturing to them. Note however that
adviser software needs an advanced data presentation system and the
adviser needs to know how to use the various data slices effectively.
A major issue is that the current robo-adviser systems, immature as
they are, are superior to the Web sites of most advisers and wealth man-
agers in terms of graphical presentation, personalization, and customer
motivation and acquisition. Successful advisers will have to urgently
learn Web presentation skills if they are to survive against fast improv-
ing robo-advisers.
One obvious area of human advantage is during bear markets, when
nervous clients may abandon AI systems which do not offer psychologi-
cal reassurance, or rearrange assets sufficiently to retain liquidity. The
use of social media by advisers needs to be handled with care, as there
are potential pitfalls. While carefully selected links and postings can add
to client understanding and engagement, the adviser needs to be care-
ful not to link articles which could lead to faulty actions by clients if
misunderstood.
Note that while there will be a legitimately wide range of adviser
responses to advances in technology, with some firms using it inten-
sively for most clients, while others will use it as a supplement to a
human-focused practice, advisers will have to pay close attention to
ensuring each part of their business is cost-competitive with the rap-
idly falling cost of robo-advisers to avoid being disintermediated. The
best advisers will be those who successfully blend the advantages of the
human approach with the advantages of the technological approach to
offer clients an omni-channel personalized service which is superior to
today’s service at a lower price, as well as enabling them to service more
314    
M. Naylor

clients, serve them better, and serve them with less time. They can ser-
vice clients in China as easily as they can clients down the road. Less
able advisers will face increasing competition from online sites or AI
software. Advisers who cannot distinguish themselves from software will
exit the industry, as the costs of their IT competitor could be dropping
by 30% a year. It is probable that a wide gap will open between the best
tech-savvy advisers and the rest.
The Economist (2014) argues that the IT revolution threatens to
drastically widen the income and wealth gap between the small number
of highly paid IT workers and the stagnating rest of us, as most workers
struggle to cope. This is obviously good for advisers who deal with the
wealthy, as long as their Web interface appeals to the highly IT liter-
ate. An effective IT internal admin system will enhance the abilities of
­advisers able to work with the new tech-wealthy.
The threat of generic advice from robo-advisers will initially impact
most strongly on bank or large chain advisers. This is because the major-
ity of these advisers follow an advice framework handed to them by the
institution which is designed to maximize sales. The process does not
empower the advisers to form a professional view of what’s in the cli-
ent’s best interest. Advisers are also encouraged to accept the client’s self-
diagnosis of what they need. The focus is on the outcome, the product
distribution, not on whether the advice is fulfilling the client’s needs.
There is no trust relationship established.
The problem is that most clients do not know what is in their best
interest - that is why they need ‘advice.’ What clients need is a pro-
fessional advice process which follows a guiding set of principles to
­accurately diagnosis the client’s condition and determine the best set of
solutions, even if they do not involve a product, or involve a competi-
tor’s product, or a non-product strategy like budgeting or savings, or
involve advice about what products they do not need. As the adviser
environment changes to become more advice and customer focused,
banks or generic chain-firms will thus find it harder to compete with
robo-advisers than will adviser firms which already offer a professional,
client-focused, service. Individual advisers will have to be strategic about
which firms they work for and need to ensure that their firm has the
capacity to move with the waves of change.
 11  Impacted Occupations    
315

Another key concept for networks of advisers or banks or insurers

who offer advice, is the importance of social capital. Advisers and
adviser networks are going to have to accept that digital native cli-
ents will review their services and Web platforms online in detail and
that adviser practices which do not offer a sophisticated, efficient, and
friendly service, will be rated down and will struggle to attract new cli-
ents. Advisers need to actively encourage unedited reviews by clients.
Advisers will have to use a mix of channels across a range of social
platforms and modes. They will have to advertise and participate in
social media. Adviser practices will have to actively monitor their social
capital and have proactive programs to increase it. Use of informality
and humor will have to be selectively used.
They will have to invest heavily to keep up-to-date with the latest
trends in Adviser software while offering a client-focused interface which
is sophisticated enough to appeal to the IT savvy and yet easy-to-use so
that it appeals to the less IT savvy. Their software systems need to be able
to offer a 24/7 service, even when no humans are present. Any advisers
who fail to consistently uphold the highest level of client service will find
their rating on social networks dropping, with a subsequent loss of clients.
Insurance advisers and product suppliers will need to simplify their
products and ensure the related fees and commissions are transparent,
as robo-advisers will offer simple products with straightforward fees.
Advisers need to be able to be proud of the way they are paid, so they
can discuss this on social media. Any hint of their advice being influ-
enced by commission and not in the client’s best interests could devas-
tate their social capital. Note that it is the response to complaints which
matters, not the mistakes themselves.
Given that the adviser market will become inherently international,
there will be a need for regulation to dynamically change as the adviser
market adapts to robo-competitors, especially if these are foreign based.
It will be unfair to domestic advisers if they have to obey expensive
rules while foreign Web-based advisers don’t. This may require extensive
international regulatory integration.
The financial advice profession has demonstrated over the last few
decades the capacity to respond positively to industry, regulatory, and
market changes, and should have the capacity to respond to robo-advice
and digital native clients.
316    
M. Naylor

Diagnostic Robo-advice Systems

The majority of current robo-advisors try to reproduce the traditional

financial advice process. Specifically, they install a detailed questionnaire
on their Web sites, just as if you had gone to the office of a financial
planner. They then try to understand your current financial situation;
your goals, age, income, risk aversion, etc. They then recommend
choices; how much to save; how much to invest; how much insurance
you need, etc. Finally, they implement each choice for you. This is the
standard CFP six-step process. This suits the majority of traditional cli-
ents who do not have much knowledge or confidence.
This approach may, however, not suit the likely first clientele of robo-
advice systems, the do-it-yourselfer. Those investors have instead already
made their investment choices. What they need is for the robo-advisor
system to examine those choices and find improvements. This approach
can be described as diagnostic. For these clients, robo-advisers should
not use the traditional questionnaire but will instead ask investors why
they have made their current choices. Why are you 65% invested in stocks
when your risk profile indicates you should be reducing risk? Why have you
chosen an aggressive growth fund and not a conservative? Why aren’t you
using tax-deductible funds? Here are two reasons why that might be best
practice. Does either apply to you? Here are some suggestions to reduce your
insurance risk.
SEI (2014) argues that there are eight actions which advisers have to do:

1. Co-plan with clients: Clients have become more astute, so advisers

need to work with clients rather than act as an expert who imposes
advice from on-high.
2. Adopt goal-based reporting: Advisers should embrace total life goals
rather than pure financial and assess client progress toward those
broad goals, rather than money saved. Understand behavioral invest-
ing issues.
3. Hone the value proposition: Assess exactly what advantages being
human offers over software, and focus on that part of the business,
and automate the rest. Being a niche player may be effective as the
potential market is now international.
 11  Impacted Occupations    
317

4. Age and ethnic diversify: Ensure the firm has as wide a range of pos-
sible clients as possible, by hiring diverse younger advisers or those
from potential target markets, as investors most identify with a like
person.
5. Optimize operations: Examine each part of the firm’s processes and
use the least cost staff possible and look to eventually running firm
admin by a software package. Ensure customer contact systems use
multiple channels and offer clients at least as good an experience as
rivals.
6. Assess income structure: Ensure the structure of fees/commissions will
survive banks and fund managers offering free Web-based transac-
tions. Know what is best practice.
7. Become a techno-adviser: Ensure the firm is aware of the latest soft-
ware and is at the forefront of adviser-friendly CRM and admin
software. Ensure all staff are able to optimize all systems. Create a
market-leading client interface.
8. Outsource: Outsource everything except your client relationship and
value proposition. Don’t try to become a chief technology officer - it’s
not your core skill set.

Revenue Issues

The basis of fee charging will change as investment companies will use
robo-advice software to enable investors to trade for free or at very low
basis points. Many banks or Internet firms may offer robo-advisers for
free as a ‘loss-leader.’ Thus, advisers will have to move to a fee for advice
model rather than one based on assets-under-management.
Current robo-advisers tend to charge a percentage of assets, typi-
cally between 25 and 75 basis points (0.25–0.75%). They find it hard
to charge any ongoing fee unless they provide ongoing advice. This is
the challenge for a diagnostic robo-advice service - how to charge ongo-
ing rather than one-off, or episodic, fees. There are two approaches that
could be taken. First, a diagnostic robo-adviser could be embedded in a
wider service. It might be bundled with the services of an SMSF admin-
istration provider. Secondly, the diagnostic service could be provided at
318    
M. Naylor

low cost by a private wealth management firm, private bank or even an

industry fund as a way of establishing contact with wealthy investors.
For those advisers who do run robo-advisers, they will have to think
carefully about charging. While the financial advice sector is mov-
ing toward fees for service, many still charge a fraction of funds under
advice. Even advisers who claim to charge a fee for service often start by
asking the size of funds to be advised on and then work backward to a
nominal fee per hour.
What very few advisers do is start by working out fixed (non-time
related) and marginal (time related) costs per client and then charging
based on a profit margin above that. This can be seen as distinguishing
between the costs required to keep the business open even without cli-
ents vs the extra costs that each type of client creates. In general, advis-
ers tend to have substantially higher marginal costs than fixed costs. In
contrast, robo-advice systems tend to have higher front fixed costs (for
creation) and then very low marginal costs. What this implies is that
extra (predominately) robo-clients can be taken on for little extra cost.
Clients need to be segmented into low vs high-touch. Low-touch clients
can be offered robo-mainly advice with low fees, and additional charges
for contact with humans. These clients are useful as high fixed and low
marginal costs also mean that there will be substantial decreases in per-
client cost as the number of clients increase.
This may imply cost advantages to larger networks, but these should
be offset by smaller firms buying robo-adviser software from cloud-
based 3rd party vendors. These vendors will also be able to supply sys-
tem maintenance services and unlimited cloud-based processing and
storage. These will probably be the same vendors which supply the
adviser’s administrative systems, so that the results of robo-advice can
be inputted directly to client files. This will enable advisers to handle
a large book of less-wealthy clients as long as they set restrictions on
human-based contacts. Given that these clients are currently unprofit-
able but substantially exceed wealthier clients in number, this will mul-
tiply the client base and spread fixed costs.
 11  Impacted Occupations    
319

Overview

The IT advances will substantially change the world of financial and

insurance advisers, but probably not as much as they will change insur-
ance companies. Advisers/brokers who have approached insurance as a
volume-based transaction business will find that they cannot compete
on cost with increasingly advanced software. Advisers who focus on
advice and psychological areas will thrive as adviser software both reduces
administrative costs and increases client engagement. The passion an
adviser can project to a client is very hard for software to re-create.
It is very likely that the admin cost of offering advice will drop drasti-
cally, maybe to 10% of its current level, so advisers will be able to han-
dle more clients with drastically less paperwork. This reduction in cost
should mean a sharp drop in client fees and a correspondingly rapid rise in
demand for advice. Advisers who are able to ride the wave of change should
find themselves with substantially increased incomes and happier clients.

References
Accenture. (2013). Closing the gap: How tech-savvy advisors can regain investor
trust.
AIA/Beddoes. (2015). The advice challenge: Understanding what clients value,
Australia.
Arthofer. (2016). Insurance brokerage is broken. TechCrunch, May 2.
CBRE/Genesis. (2014). Fast forward 2030: The future of work and the workplace.
Economist. (2014). The third great wave. Special Report, October 4th.
Markoff, J. (2011, March). Armies of expensive lawyers, replaced by Cheaper
Software. New York Times. Science, 4.
Morgan-Stanley/BCG. (2014). Insurance and technology: Evolution and revolu-
tion in a digital world. USA: Boston Consulting Group.
SEI. (2015). The Next Wave of Financial Planning. White Paper. www.seic.com/
enUS/advisors/14834.htm.
SwissRe. (2013). Automated underwriting: Threat or opportunity?. New York.
Zaptitude. (2016). Investing in Millennials, Research Report. www.zaptitude.com.
 12
Conclusion

There is a perfect storm of technological advances which are about to

combine and disrupt both the insurance industry and the adviser/bro-
ker industry. The majority of current companies and advisers are tech-
nological laggards and will require immense transformation before they
will be in a position to combat these threats. The industry is thus wide-
open to disruptive external entrants who have leading-edge skills in
technology.
Insurers have to understand that, while parts of the technology
explored in this book seem extreme to insurance managers, nearly all
the technology covered already exists, either in companies in other sec-
tors, or within small new entrants, or in beta form. There is no question
about ‘what will occur’ - there is only a question as to the timing. The
insurance industry in 15- to 20-year time will be very different to today
and will be continuing to change at an ever-faster rate. The big transfor-
mation will probably come from the entry of tech-savvy external firms,
firms which will do things completely different - the ‘ubers.’ The major-
ity of existing firms will probably go bankrupt due to their inability to
transform themselves into a client-centered model, while survivors will
be so transformed that they may be unrecognizable.

© The Author(s) 2017 321

M. Naylor, Insurance Transformed, Palgrave Studies in Financial Services Technology,
DOI 10.1007/978-3-319-63835-5_12
322    
M. Naylor

Insurance as an industry needs to be reimagined. Insurance is fun-

damentally a series of services: application, policy maintenance, capital
funding, customer relations, claims, which are currently packaged and
sold as a ‘product.’ Insurance needs to be transformed into a series of
risk data-based value-added services.
The ability to staff, manage, lead, and transform increasingly auto-
mated organizations will be the key future competitive advantage. The
biggest change is that insurance will stop being a price-based commod-
ity and become a lot more flexible, and focused on real-time dynamic
services. While policies may be simplified, the possibility of individual-
ized underwriting and real-time dynamic premiums mean that clients
will feel a lot more engaged with insurance and insurers. Insurers who
can build trust and social capital, and can find valuable business eco-
systems, will create substantial competitive advantage and cross-selling
opportunities. Their future will be rosy. Insurers who persist in using
traditional approaches will find themselves in Kodak’s shoes.
Insurers need to respond with a proactive structured approach involv-
ing five elements: (i) a clear long-term strategy, (ii) mechanisms to cap-
ture areas where future value is created, (iii) investment in core technical
and change management capacities, (iv) creation of an adaptive, flexibil-
ity culture, and (v) a clear road map from here to there, which is clearly
understood across the organization.

The More Distant Future

As speculative and extreme as the above advances in IT may seem, the
insurance industry needs to understand that an even more extreme
future awaits humanity after about 2040. The exponential change
curve becomes very steep then. Insurers need to adjust as fast as pos-
sible to the more immediate future I have outlined so that they are in
the best position to adjust to the more extreme future which will occur
after that. I will only outline a sample of speculative future possibilities.
Some of these technologies already exist in a trial stage, other are still in
the imagination stage. Possible future technologies include:
 12 Conclusion    
323

1. Hand chip-implants - microchips embedded in your hand, so that

when you shake hands with a similarly chipped human, the two
chips pass on each other’s essential information. There are already
more than 10,000 humans with embedded chips, though most are
low-tech.
2. Stomach micro-chip - an inbuilt bio-med reader in your stomach
which feeds back information to medical professionals on what you
are eating and how it is being digested. Research shows that differ-
ences in stomach bacteria can mean that drugs work differently for
different patients. Already done on a temporary basis via swallowed
chips, and in clinical trials.
3. Networked contact lens - a contact lens which will record every-
thing that you see as well as display information over your eyeball.
Samsung has recently patented this. These and networked glasses are
likely to replace mobile phones.
4. ID forearm chip - an embedded forearm chip which allows you to
scan items and verify your ID at buildings or security. This would
allow you to get access to your gym or car. A Swedish office already
offers this to tenants to gain entry.
5. Voice-changing windpipe - a networked artificial windpipe which
allows you to change the sound of your voice or to speak in any for-
eign language. It is covered in stem cells to avoid reject issues. Still on
the drawing board but regarded as doable.
6. Brain-chips - a networked chip embedded in your brain which gives
you access to information in real time. You can talk to a client while
being fed information about them or any topic without the other
person being aware. Some chips have already been embedded on a
trial basis.
7. Organ replacement - the combination of 3D bio-printing and stem cells
will allow most body parts to be replaced without any rejection issues.
8. Exotic bodies - it is obvious that some people will use 3D bio-printing of
body parts and genetic engineering to augment themselves with extra
limbs or weird facial features. Already happening in a low-tech way.
9. Self-aware software - nobody is able to estimate how smart an AI sys-
tem needs to be before it becomes self-aware and starts demanding
the right to be heard, to claim ‘cyborg-rights,’ to separate itself from
324    
M. Naylor

its programmers, and possibly to register itself as a business and ask

for insurance. However, this will occur at some stage, and given an
exponential curve, sooner than we think. Some industry experts pre-
dict that neutral networks will reach the same level of complexity as
the human brain by 2025, be 10x as complex by 2030, and 200x as
complex by 2040.
 Useful Sources

Abbot, D. (2014). Applied predictive analytics: Principles and techniques for pro-
fessional data analyst. USA: Wiley.
ABI/Thatcham. (2016). Pathway to driverless cars: Proposals to support advanced
driver assistance systems and automated vehicle technologies. Association of
British Insurers.
Accenture. (2013a). Running business for growth. Insurance-blogs.
Accenture. (2013b). Closing the gap: How tech-savvy advisors can regain investor trust.
ACE Insured. (2013). Reputation at risk: ACE European risk briefing. UK.
Actuaries Institute. (2016). The impact of big data on the future of insurance.
Australia: Green Paper.
Adner, R. (2002). When are technologies disruptive? A demand-based view of
the emergence of competition. Strategic Management Journal, 23, 8.
AIA/Beddoes. (2015). The advice challenge: Understanding what clients value.
Australia.
Arntz, M., Gregory T., & Zierahn U. (2016). The risk of automation for jobs
in OECD countries: A comparative analysis. OECD Social, Employment and
Migration Working Papers, #189, OECD.
Arthofer. (2016). Insurance brokerage is broken. TechCrunch, May 2.
Autor, D., & Dorn, D. (2013). The growth of low skilled service jobs and the
polarization of the US labor market. American Economic Review, 103(5),
1553–1597.

© The Editor(s) (if applicable) and The Author(s) 2017 325

M. Naylor, Insurance Transformed, Palgrave Studies in Financial Services Technology,
DOI 10.1007/978-3-319-63835-5
326    
Useful Sources

Bain & Company. (2013). For insurance companies, the day of digital reckoning.
Bain & Company. (2015). Global digital insurance benchmarking report 2015:
Pathways to success in a digital world.
Barras, R. (1986). Towards a theory of innovation in services. Research Policy,
15, 161–173.
Boobier, T. (2016). Analytics for insurance: The real business of big data. UK: Wiley.
Boston Consulting Group. (2013). Telematics: The test for insurers. BCG
Perspectives.
Boston Consulting Group. (2014). Bringing big data to life. BCG Perspectives.
Boston Consulting Group. (2015a). Robo-Taxis and the new mobility. BCG
Perspectives, April 21.
Boston Consulting Group. (2015b). Revolution in the driver’s seat: The road to
autonomous vehicles. BCG Perspectives, April 21.
Boston Consulting Group. (2015c). Navigating a world of digital disruption.
BCG Perspectives, July 24.
Boston Consulting Group. (2015d). Man and machine in industry 4.0.
September 28.
Boston Consulting Group. (2015e). The double game of digital strategy. BCG
Perspectives, October 16.
Boston Consulting Group. (2016a). How to avoid the big bad data trap. BCG
Perspectives, October 2.
Boston Consulting Group. (2016b). How digital switchers are disrupting US
auto insurers. July 26.
Breznik, L. (2012). Can information technology be a source of competitive
advantage? Economic and Business Review, 14(3), 251–269.
Brookings Institute. (2014). Product liability for driverless cars: Issues and guid-
ing principles for legislation. Centre for Technology Innovation.
Brynjolfsson, E., & McAffe, A. (2011). Race against the machine: How the digi-
tal revolution is accelerating innovation, driving productivity, and irreversibly
transforming employment and society. Digital Frontier Press.
Brynjolfsson, E. R., Hitt, L., & Kim, H. (2011). Strength in numbers: How does
data-driven decision-making affect firm performance? Social Science Review
Network, April.
Canas. T. (2015). What will be the Uber of insurance? Insurance Thought
Leadership, PWC, July 22.
Capgemini Consulting. (2015). Strategies for the age of digital disruption.
Digital Transformation Review, No 7: February.
Capgemini/Efma. (2016). World insurance report.
 Useful Sources    
327

Carr, N. (2003). IT doesn’t matter. Harvard Business Review, May, 5–12.

Carr, N. (2004). Does IT matter? Information technology and the corrosion of
competitive advantage. Cambridge, MA: Harvard Business School.
Castriotta M., Floreddu, P. B., Di Guardo, M. C., & Cabiddu, F. (2014).
‘Disentangling the strategic use of social media in the insurance indus-
try: A value co-creation perspective’, chap. 4 in Social Media in Strategic
Management: Advanced Series in Management, 63–86.
CBRE/Genesis. (2014). Fast Forward 2030: The future of work and the
workplace.
CEB. (2013). TowerGroup wealth management client experience survey.
Celent. (2014). The internet of things and property/casualty insurance: Can an old
industry learn new tricks?
Charan, R. (2015). The attacker’s advantage: Turning uncertainty into break-
through opportunities. Public Affairs, February.
Charitou, C. D., & Markides, C. C. (2003). Responses to disruptive strategic
innovation. MIT-Sloan Management Review, USA.
Charitou, C. D., & Markides, C. C. (2003). Responses to disruptive strategic
innovation. MITSloan Management Review, 44(2), 55–63.
Chartered Accountants ANZ. (2016). Disruptive technology: Risks and
opportunities.
Christensen, C. M. (1997 & 2011) (4th Ed). The innovator’s dilemma. Harper
Business/Harvard Business School, USA.
Christensen, C. M., & Overdorf, M. (2000). Meeting the challenge of disrup-
tive change. Harvard Business Review, March–April.
Christensen, C. M., Anthony, S. D., & Roth, E. A. (2004). Seeing what’s next:
Using the theories of innovation to predict industry change. USA: Harvard
Business School.
Christian, B., & Griffiths, T. (2016). Algorithms to Live By: The computer science
of human decisions. New York, USA: Henry Holt & Company.
CII. (2012). Future risk: How technology could make or break our world. UK:
Chartered Insurance Institute.
CII. (2015). Big data and insurance: A conversation. UK: Chartered Insurance
Institute.
CII/EY. (2016). Insurer of the future. Confederation of Indian Industry.
Cisco. (2011). The internet of things: How the next evolution of the internet is
changing everything.
Citi GPS. (2015). Technology at work: The future of innovation and employment.
Citi GPS Global perspectives and Solutions.
328    
Useful Sources

Clemons, E., & Hitt, L. (2000). The internet and the future of financial services:
Transparency, differential pricing and disintermediation. The Wharton School.
Cognizant. (2017). The work ahead: Seven key trends shaping the future of work
in the insurance industry.
Cohan, M. (2012). Product performance based business models: A service
based perspective. Hawaii Conference on System Science. http://ieeexplore.
ieee.org/abstract/document/6149476/.
Corporate Insight. (2016). The innovation advantage. White Paper, New York,
http://www.corporateinsight.com.
Cusumano, M. A., Kahl, S. J., & Suarez, F. F. (2014). Services, industry evolu-
tion, and the competitive strategies of product firms. Strategic Management
Journal, 36, 559–575.
Danneels, E. (2004). Disruptive technology reconsidered: A critique and
research agenda. Journal of Product Innovation Management, 21(4),
246–258.
Dataiku. (2016). Data science for banking & insurance. USA.
Deloitte. (2013a). The out-of-sync advisor: Applying disruptive innovation to serve
non-customers of wealth management advice. Deloitte Centre for Financial
Services/University Press.
Deloitte. (2013b). Innovation in insurance: The path to progress, Forward focus
series. USA: Deloitte University Press.
Domingos, P. (2015). The master algorithm: How the quest for the ultimate
learning machine will remake our world. USA: Basic Books.
Dosi, G. (1982). Technological paradigms and technological trajectories: A
suggested interpretation of the determinants and directions of technical
change. Research Policy, 11, 147–162.
Economist. (2014). The third great wave, special report, October 4th.
Economist. (2016). The return of the machinery question, special report, June
25th, UK.
Energesse. (2015). Future solutions in customer experience for health insurers
white paper.
Equinix. (2014). Challenge to change: Three-part series. Acord/Global Futures &
Foresight.
Equinix Insurance Post. (2016a). The future of insurance technology: Customer
data and analytics. UK: White Paper.
Equinix Insurance Post. (2016b). The future of insurance technology: IT out-
sourcing and the cloud. UK: White Paper.
European Commission. (2013). Digital agenda for the future: A Europe 2020
initiative. Brussels: Belgium.
 Useful Sources    
329

EXL. (2016). Robots, smart homes and ride shares: Innovations opportunities in
the age of disruption for P&C insurers. USA: EXL white paper.
EY. (2013). Insurance in a digital world: The time is now. New York: Ernst-Young.
EY. (2014). The future of underwriting: A transformation driven by talent and
technology. New York: Ernst-Young.
EY. (2015). Leading insurers address reputation and its risks. New York:
Ernst-Young.
Floreddu, P. B., & Cabiddu, F. (2014). Multi-channel strategy and sustain-
able competitive advantage: A dynamic capacity approach, working paper,
University di Cagliari.
Floreddu, P. B., Cabiddu, F., & Evaristo, R. (2014). Inside your social media
ring: How to optimize online corporate reputation. Business Horizons, 57,
737–745.
Forbes. (2014). Internet of things by the numbers: Market estimates and forecasts.
Foreman, J. W. (2013). Data smart: Using data science to transform information
into insight. USA: Wiley.
Foster. (2012). The customer experience index.
Franks, B. (2014). The analytical revolution: How to improve your business by
making analytics operational in the big data era. USA: Wiley.
Frey, C. B., & Osborne, M. A. (2013). The future of employment: How suscepti-
ble are jobs to computerisation? Working Paper, Oxford University, U.K.
Gartner.com. http://www.gartner.com/technology/research/methodologies/.
General Electric. (2012). In P. Evans & M. Annunziata (Eds.), Industrial inter-
net: Pushing the boundaries of minds and machines. USA.
GenRe. (2014). Marketing to the millennial mindset. USA.
Gibbs, W. W. (2014). The clock-watcher. Nature, 508.7495 (Apr 10, 2014): 168–170.
Gilbert, C., & Bower, J. (2002). Disruptive change: When trying harder is
part of the problem. Harvard Business Review, 80(5), 94–102.
Glasmeier, A. (1991). Technological discontinuities and flexible production
networks: The case of Switzerland and the world watch industry. Research
Policy, 20, 468–485.
Gomes. (2014). Driving in circles. Technology, October 21.
Halliburton, C., & Poenaru, A. (2010). The role of trust in consumer relation-
ships. ESCP Europe Business School.
Horvath, S. (2013). DNA methylation age of human tissues and cell types.
Genome Biology, 14(10), 3156.
IAIS. (2017). FinTech developments in the insurance industry. Basel,
Switzerland: International Association of Insurance Supervisors.
330    
Useful Sources

Iansiti, M., & Levien, R. (2004). The keystone advantage: What the new dynam-
ics of business ecosystems mean for strategy, innovation, and stability. Boston,
USA: Harvard Business School Press.
IBM. (2006). Insurance 2020: Innovating beyond old models. USA: IBM
Institute for Business Value.
IBM. (2007). Insurance 2020: Now what? Exploring initiatives for innovation.
USA: IBM Institute for Business Value.
IBM. (2015). Capturing hearts, minds and market share: How connected insurers
are improving customer retention. USA: IBM Institute for Business Value.
Insurance Networking News. (2016). Strategies: Policy administration and
underwriting. UK.
International Transport Forum/OECD. (2015). Urban mobility system upgrade:
How shared self-driving cars could change city traffic.
Johne, A., & Davies, R. (2000). Innovation in medium-sized insurance
companies: How marketing adds value. International Journal of Bank
Management, 18(1), 6–14.
Joseph, M., Stone, G., & Anderson, K. (2003). ‘Insurance customers’
Assessment of service quality: A critical evaluation. Journal of Small Business
and Enterprise Development, 10(1), 81–92.
Kamiya, S., Schmit, J. T., & Rosenberg, M. A. (2010). Determinants of insurers’
Reputational risk, Society of Actuaries, US.
Kanter, R. (1990). When giants learn to dance. USA: Free Press.
Kanter, R. (2006). Innovation: The classic traps. Harvard Business Review,
84(1),
Kelly, K. (2016). The inevitable: Understanding the 12 technological forces that
will shape our future. USA: Viking.
Komlos, J. (2016). Has creative destruction become more destructive? The B.E.
Journal of Economic Analysis & Policy: 20160179.
KPMG. (2014). An industry transforming: Insurers and intermediaries in a fast
changing, digital world, white paper.
KPMG. (2015a). Automobile insurance in the era of autonomous vehicles: Survey
results, June.
KPMG. (2015b). Big Data: The big factor driving competitive advantage,
Frontiers in Finance.
KPMG. (2016). Empowered for the Future: Insurance reinvented, KPMG
International.
Kunreuther, H. C., Pauly, M. V., & McMorrow, S. (2013). Insurance and
Behavioral Economics: Improving decisions in the most misunderstood industry.
UK: Cambridge University Press.
 Useful Sources    
331

Lepore, J. (2014). The Disruption Machine, Annals of Enterprise.

Levy, F., & Murnane, R. J. (2004). The New Division of Labor: How computers
are creating the new job market. USA: Princeton University Press.
Lloyds Insurance. (2014). Autonomous Vehicles - Handling over Control:
Opportunities and Risks for Insurance, UK.
Markides, C. (2006). Disruptive innovation: In need of better theory. The
Journal of Product Innovation Management, 23, 1.
Markoff, J. (2011). Armies of Expensive Lawyers, Replaced by Cheaper Software.
New York Times, Science, March 4.
Martinez, A. G. (2016). Chaos Monkeys: Inside the Silicon Valley Money
Machine. London, UK: Ebury Press.
McGahan, A. M. (2000). How Industries Evolve. Business Strategy Review,
11(3), 1–16.
McGill, R. (2008). Technology Management in Financial Services. London:
Palgrave-McMillan.
McKinsey. (2012). The Social Economy: Unlocking value and productivity
through social technologies: consumer financial services. McKinsey Global
Institute.
McKinsey. (2015a). The Internet of Things: Mapping the Value Beyond the Hype,
McKinsey Global Institute.
McKinsey. (2015b). An Executive’s Guide to Machine Learning, McKinsey
Quarterly, June.
McKinsey. (2015c). Four Fundamentals of Workplace Automation, McKinsey
Quarterly, November.
McKinsey. (2016a). Automating the Insurance Industry, McKinsey Quarterly,
January.
McKinsey. (2016b). An Incumbent’s Guide to Digital Disruption, McKinsey
Quarterly, May.
McKinsey. (2016c). The Economic Essentials of Digital Strategy, McKinsey
Quarterly, March.
McKinsey. (2016d). Transforming life insurance with design thinking, May.
McKinsey. (2016e). Making Digital Strategies a reality in Insurance, September.
McKinsey. (2016f ). Shifting Gears: Insurers adjust for connected-car ecosystems,
May.
McKinsey. (2016g). The Need to Lead in Data and Analytics, April.
McKinsey. (2016h). A Digital Crack in Banking’s Business Model, April.
McKinsey. (2016i). How Companies are using Big Data and Analytics, April.
McKinsey. (2017a). A Future that Works: Automation, Employment, and
Productivity, January.
332    
Useful Sources

McKinsey. (2017b). The Case for Digital Reinvention, February.

MGI. (2013). Disruptive Technologies: Advances that will transform life, business
and the global economy. McKinsey Global Institute: Tech. Rep.
Mitchell, T. M. (1997). Machine Learning. USA: McGraw-Hill.
Morgan-Stanley, B. C. G. (2013). Wearable Devices: The internet of things
becomes personal. USA: Boston Consulting Group.
Morgan-Stanley, B. C. G. (2014). Insurance and Technology: evolution and revo-
lution in a digital world. USA: Boston Consulting Group.
Morgan-Stanley, B. C. G. (2015). Insurance and Technology: The emerging role
of ecosystems in insurance. USA: Boston Consulting Group.
Morgan-Stanley. (2016a). Cybersecurity: rethinking security. USA: Blue Paper.
Morgan-Stanley. (2016b). Insurance – Property & Casualty – Digital disruption
in small business insurance’, North America Insight, June 29.
MyCustomer/Microsoft. (2015). Customer Experience in the Insurance Sector:
Insight Report,’ UK.
Nahai, N. (2013). Webs of Influence: The Psychology of Online Persuasion, FT
Press.
Neirotti, P., & Paolucci, E. (2007). Assessing the Strategic Value of
Information Technology: An analysis on the Insurance Sector. Information
and Management, 44, 568–582.
Nicoletti, B. (2016). Digital Insurance: Business Innovation in the Post-Crisis
Era. London: Palgrave-MacMillan.
Nordhaus, W. D. (2007). Two centuries of productivity growth in computing.
The Journal of Economic History, 67(1), 128.
Pearson, R. (1997). Towards an Historical Model of Services Innovation” The
case of the insurance industry, 1700–1914. Economic History Review, L, 2,
235–256.
Perdue University/ Benchmark Portal. (2014). Service Centre Industry Report.
Peters, T. J., & Waterman, R. H. (1982). In Search of Excellence. New York,
USA: Harper & Row.
Porter, M., & J. Heppelmann. (2015). How smart connected products are
transforming competition, Harvard Business Review, November.
Progress Corticon. (2014). Flexibility and Control in Insurance Operations,
White Paper, www.progess.com.
Progressive Policy Institute. (2013). 752,000 App Economy jobs on the 5th anni-
versary of the App Store, www.progressivepolicy.org.
PWC. (2012). Insurance 2020: Turning change into opportunity.
PWC. (2013). PwC’s NextGen: A Global generation study.
PWC. (2014a). Reinventing Life Insurance.
 Useful Sources    
333

PWC. (2014b). Insurance 2020: the digital prize – taking customer connection to
a new level.
PWC. (2015a). Insurance 2020 & Beyond: Necessity is the mother of reinvention.
PWC. (2015b). Insurance Banana Skins.
PWC. (2016). Turning risk into opportunity: The changing face of Financial
Services: 19th Annual Global CEO Survey.
Rao, A. (2015). The Rise of Robo-Advisers. Insurance Thought Leadership, PWC, June.
Rifkin, J. (2011). The Third Industrial Revolution: How lateral power is transform-
ing energy, the economy, and the world. London, UK: Palgrave-MacMillian.
Rifkin, J. (2014). The Zero Marginal Cost Society: The internet of things,
the collaborative commons, and the eclipse of capitalism. London, UK:
Palgrave-MacMillian.
Sabatier, V., Craig-Kennard, A., & Mangematin, V. (2012). When
Technological Discontinuities and Disruptive Business Models Challenge
Dominant Industry Logics: Insights from the drug industry. Technological
Forecasting & Social Change, 79, 949–962.
Saporito, P. L. (2015). Applied Insurance Analytics: A framework for driving more
value from data assets, technology, and tools. USA: Person.
SAS. (2013). Predictive Claims Processing: Transforming the insurance claims life-
cycle using analytics, SAS White Paper.
Satell, G. (2012). Business Models and the Singularity, May.
Schryen, G. (2013). Revisiting IS business value research: What we already
know, what we still need to know, and how we can get there. European
Journal of Information Systems, 22, 139–169.
SEI. (2015). The Next Wave of Financial Planning, white paper, www.seic.com/
enUS/advisors/14834.htm.
Shirky, C. (2010). Cognitive Surplus: Creativity and generosity in a connected
age. New York: Penguin Press.
Siegal, E. (2016). Predictive Analytics: The power to predict who will click, buy,
lie, or die. USA: Wiley.
Stevens, D. (2015). Prepare to be Disrupted’, 12 March, www.riskadviser.com.
Stewart, R. B. (1981). Regulation, Innovation, and Administrative Law: A
conceptual; framework. California Law Review, 69(5), 1256–1377.
Summers, L., & DeLong, J. B. (2001). The ‘new economy’: Background, histori-
cal perspective, questions, and speculations. USA: Economic Review of Federal
Reserve Bank of Kansas.
Susskind, R., & Susskind, D. (2015). The Future of the Professions: How tech-
nology will transform the work of human experts. Oxford, U.K.: Oxford
University Press.
334    
Useful Sources

SwissRe. (2013). Automated Underwriting: Threat or Opportunity? New York.

SwissRe. (2014). Insurance and the Rise of Drones, New York.
SwissRe. (2016a). How 3D printing will reshape the insurance landscape. New
York: Centre for Global Dialogue.
SwissRe. (2016b). Sigma 2016, New York.
SwissRe. (2016c). The Future of Motor Insurance: How car connectivity and
ADAS are impacting the market. Berlin: SwissRe/Here.
Talla, V., Kellogg, B., Ransford, B., Naderiparizi, S., Gollakota, S., & Smith, J.
(2015). Powering the Next Billion Devices with Wi-Fi, ArXiv:1505.06815.
Target Group. (2015). Distribution Channels in Insurance: Understanding the
Customer Viewpoint.
Teece, D. J., Pisano, G. G., & Shuen, A. (1997). Dynamic Capacities and
Strategic Management. Strategic Management Journal, 18(3), 509–533.
TowerWatson/MergerMarket. (2013). ‘Surviving the Perfect Storm: The outlook
for Insurance M&A in EMEA’, UK.
Verint. (2016). Customer-Centricity in Financial Services: The rules of engage-
ment. USA: White Paper.
Wernerfelt, B. (1984). A Resource-based View of the Firm. Strategic
Management Journal, 5(2), 171–180.
Willetts, D. (2012). Ties that bind: The important of science and technology to
the insurance industry, Thinkpiece no 89, Chartered Insurance Institute,
October.
World Economic Forum. (2015). The Future of Financial Services: How disrup-
tive innovations are reshaping the way financial services are structured, provi-
sioned and consumed.
Yates, J. (2005). Structuring the Information Age: Life Insurance and Technology
in the Twentieth Century. Baltimore, USA: John Hopkins University Press.
Yoder, J., & Rao, A. (2015a). Insurance at a Tipping Point, Insurance Thought
Leadership, PWC, June 20.
Yoder, J., & Rao, A. (2015b). Reinventing Life Insurance, Insurance Thought
Leadership, PWC, April 15.
Yovanof, G. S., & Hazapis, G. N. (2008). Disruptive Technologies, Services,
or Business Models? Wireless Pers Commun, 45, 569–582.
Yu, D., & Hang, C. C. (2010). A Reflective Review of Disruption Innovation
Theory. International Journal of Management Reviews, 12, 435–452.
Zaptitude. (2016). Investing in Millennials, Research Report, www.zaptitude.
com.
 Index

A C
Administrative processing cost Change management 97–100, 116,
49–51, 70, 104, 298–301 134–135, 145–148
Automated claims 111–114, 201, Chatbots 33–34
206, 277 Cloud computing 18–19
Automated underwriting 103–111, Complex adaptive systems 26–31
297–298 Artificial Intelligence (AI) 25–31,
Automobile insurance 111, 168, 130–134
181–188, 204–205 artificial neutral networks 28–29
Autonomous vehicles 7–8, 175–181, complexity science 25–31
215–217 Cultural change 114–118
Customer issues 58–68, 125–130,
150, 159
B
active engagement 127–129,
Big Data 19–21, 69–76, 80–83
147–149, 150–153, 164–170,
Block-chain Payment systems 36–37,
192, 198–199, 202
140
customer churn 81
Business ecosystems 158–164
customer pain points 60–61, 111,
113, 126–127, 152, 194

© The Editor(s) (if applicable) and The Author(s) 2017 335

M. Naylor, Insurance Transformed, Palgrave Studies in Financial Services Technology,
DOI 10.1007/978-3-319-63835-5
336    
Index

Customer Relationship E
Management (CRM) systems Employment impact 50, 281-294,
24, 65, 127–130, 307 297-319
customer trust of insurer 73, 88 Ethical issues 73-74, 79, 84, 167,
insurance value proposition 263-265, 268-269, 271-274,
59–61, 125, 192–194 277-279
omni-channel engagement 65–66, Exponential change 4-7
126, 148–149, 307–309
trigger points 129
F
Cyber insurance 156–157, 205
Financial advisers 304-315
Flexible programming 143-145
D
Data/product Silos 74, 115, 117, 135
G
Data security 133, 203, 205, 270
Gartner 5-step Hype Cycle 9-10
Data Storage 18–19
Data visualization systems 57
DARPA 7–8, 37 H
Digital Agents 35–36, 79 Hardware innovation 25
Digital avatars 33–34, 287–288 Health insurance 78, 168, 193–203
Digital Natives 21–24, 65–66, 79, Hyperscaling 31–32
126
Disintermediatization 67–68, 247
I
Disruptive change management 5–6,
Image recognition 27–28
48, 93, 217–218, 221–227,
Individualization 79
232–237
Internal management systems
disruptive entry 228–231
134–139
disruptive innovation 93–99, 136,
Internet of Things (IoT) 15–18
242–245
Intelligent Process Automation (IPA)
disruption survival pathways
53, 132–133
248–249
Insurance production costs 49–51,
disruption point 232–234
104, 142
The double game 236–242
Insurance telematics 17–18
stages of disruption 235–236
IT system creation 140–142,
3D printing 38
160–164
Dynamic administrative systems
IT productivity paradox 53
105–106
Dynamic capacity 154-155
Dynamic insurance 77, 102, 106-
107, 258-259
 Index    
337

L S
Legacy systems 54–58, 115, 121–125 Service speed 23–24
Legal & regulation issues 79, 105, Servicization 51–52, 101–102, 258
121, 134, 164, 196, 219–220, Social capital 24, 150, 164–170, 199,
263–279, 315 203, 315
Social media 64, 165–166
Social networks 23–24, 136–138,
M
150, 166
Machine Learning 21, 75, 132
Software agents 26
Millennials 21–24, 125, 149, 191,
310–311
T
Telematics 15–18, 76–78, 246
N
automotive 77, 179, 204–205,
Non-disclosure 106, 272
213–215, 278
medical 78, 196–198
P networking 176–179
P&C insurance 112, 169, 189–190 Third wave of technological change 3
Per-to-peer insurance 155–156 Transformation pathways 118–121
Policy creation 171–172
Predictive analytical systems 26
V
Visual recognition 33–34
R Voice recognition 32–35
Ratemaking 77, 80, 84, 163, 172
Risk pools 104
Robo-advice 85–89, 301–319
Robotics 37