Tech Trend Radar

2025
What innovations will transform insurance in 2025?
Explore our expert assessment
Munich Re Page 2/111
Tech Trend Radar 2025

Introduction

The Tech Trend Radar 2025, a collaborative initiative of By leveraging insights from various business and strategy
Munich Re Business Technology and ERGO IT Strategy, functions within Munich Re Group, and those actionable
provides essential information on technology-driven trends and objective insights from Gartner® research, the radar offers
relevant to insurance companies. Prepared by insurers and a comprehensive overview of the latest tech trends, their
for insurers, it aims to raise awareness, spark new discus- maturity, and their relevance for the insurance industry,
sions, and unlock new business opportunities. helping insurers make informed decisions and drive growth.
AI
-A

e
ed Medicin
u

nce
gm

tics
lthcare
en

er ie
an

Analy
t ed

it y
m

Exp
AI Medicine
AI
So

t iv
Digital Hea
hy Hu

Personaliz
Se

vioral
f tw

ec
te d
Leg
arc

nn
are

g
nec
Healt

Co
e

t in
hE

al T

B eha

ers
En

pu
us
ngi

C on
e

t av
gin

om
it o
ch
nes
ee

Me

y
tC
i qu
AI

it
r in

al
G

en
Ag

Ub

Re
en

AI
bi
ns
en
er

ed
G Am
tio
at

ts

ov

nd
AI c
iv

De er ra
te
e

m na AI nt
e
AI

Ex
oc nc g e sI
ra
tiz e Ed ou
at om
Sy
nth io t on pt
o
y
n Au ry
eti C om
Qu
ant
cD r& E c on
um at a be at a
Co Cy ed
D
I n su
ra n c mp
ut i tr aliz
e AP ng c en ce
I St De lian
and
ardi C omp
zat i om at e d
Artific on Au t
ial Ge
neral
Intell ts
igenc l Ass e
e Digit a
Spatial Inte efense
lligence
Deepfake D

Data & AI Non-Human Identity

Hold Assess Trial Adopt
Digital Imm
une System

Rede
fin ing In
dustr
ies
Cro
p In
t elli
g en
Ne ce
wE
n er
gy
Tra at S
d eS cal
e
ec
El ret
ec Pr
tri ote
H cV c t io
yd eh n
ro i cl
g en es
H

Ec
um

on
om
an
Au

oi

y
to

d
Cl

no

Ro
im

b
at e

ot
ou

s
sM
Re

ob
s il

ilit
ien

y
ce
Munich Re Page 3/111
Tech Trend Radar 2025

Contents

Introduction 2 Data & AI 9

Foreword 4–5 Spatial Intelligence​ 12
Our approach 6 Artificial General Intelligence 14
How to read this report 7–8 Insurance API Standardization 16
Quantum Computing 19
Synthetic Data 21
AI Democratization 23
AI Governance 24
Impacts on the value chain 101–102 Generative AI 26
Our solutions 103–109 AI Agents​ 29
Team & Contributors 110 AI-Augmented Software Engineering 32
Imprint 111 AI Search Engines 33
Legal Tech 35

Healthy Human 37
Personalized Medicine 40
AI Medicine 43
Digital Healthcare 46
Behavioral Analytics 48

Connected Experience 50
Metaverse 53
Ubiquitous Connectivity 55
Ambient Computing 56
Extended Reality (XR) 57
Edge AI 59
Autonomous Interactions 60

Cyber & Crypto 62

Decentralized Data Economy 65
Automated Compliance 67
Digital Assets 69
Deepfake Defense 72
Non-Human Identity 74
Digital Immune System 76

Redefining Industries 78
Crop Intelligence​ 81
New Energy at Scale 84
Trade Secret Protection 86
Electric Vehicles 88
Hydrogen Economy 91
Humanoid Robots 94
Autonomous Mobility 97
Climate Resilience 99
Munich Re Page 4/111
Tech Trend Radar 2025

Foreword

Message from the CEO

Technology is a driving force behind insurers’ business suc- Understanding new technologies, their applications and risks,
cess and has long since ceased to be just a supporting factor. and developing customized insurance solutions for clients
The latest edition of our Tech Trend Radar highlights the key based on technological expertise and sound judgement is
trends shaping our industry, for example AI Agents, Spatial more crucial than ever for achieving or maintaining a leading
Intelligence or Digital Healthcare. competitive position.

Technological advancements not only have the potential to We want the expertise gathered in this Tech Trend Radar to
transform our industry processes and dynamics. At the core of help our clients navigate new business opportunities, find the
our business, they can act as a catalyst for the emergence of best route to increased operational efficiency and ultimately
completely new risks, which challenges us and at the same create more value for everyone on board.
time opens up opportunities. The same technologies that bring
efficiency and innovation often also create new vulnerabilities I invite you to explore Tech Trend Radar 2025 and join us on
and liability challenges. this journey.

In the past, the dual nature of technology has always given Sincerely yours,
rise to new insurance solutions, such as car insurance
with extended warranty for electric vehicles, cyber insurance
or AI performance guarantees.

Joachim Wenning,
Chair of the Board of Management
Munich Re
Munich Re Page 5/111
Tech Trend Radar 2025

Foreword

Tech Trend Radar 2025 –

Insurance in the age of intelligence
In 2025, Artificial Intelligence (AI) is a success-critical techno-
logical dynamic for the insurance industry, though hardly the
only one. To thrive in this fast-changing environment, insurers
need a clearly defined long-term strategy and a laser-sharp
focus on how technology can help them get ahead.

Our 12th edition of the Tech Trend Radar serves as a guide to

navigating these complexities, while also shedding light on the
forces currently shaping the industry.

Investments in technology – be it AI Agents, Non-Human Beyond the optimization of insurance processes through AI
Identities, or Spatial Intelligence – should not be made for and automation, this year’s Tech Trend Radar highlights major
novelty’s sake but to enhance operational efficiency, and to shifts in the industry. The essential ability to access and
streamline all processes, from sales and underwriting to claims standardize data – particularly through sources like electronic
handling. At the same time, an acute awareness of market health records (EHRs) in the life and health sector and ACORD
dynamics is crucial. What appears revolutionary today will in property and casualty business – is transforming under-
become a commodity tomorrow, making it ill-advised to lock writing and paving the way for more accurate and informed
into expensive software solutions or develop everything risk assessments. At another level, the energy transition,
in-house. Instead, the focus should lie on investments that with new opportunities from renewables to energy storage
deliver a lasting competitive advantage, while partnerships and a growing electric vehicle infrastructure, is unlocking new
can be leveraged for those capabilities that won’t help set opportunities in specialty and motor insurance.
the company apart from its competitors.
At the same time, cyber risks are growing in step with tech-
nological progress, underscoring the need for robust protection
(which also stands to benefit from technological advances).
And rapid evolution always pose challenges for insurers: the
need to understand the complex associated risks as quickly
and thoroughly as possible, and to mitigate them through
innovative risk transfer solutions.

Nevertheless, insurers are well positioned to continue

balancing their ability and expertise, so as to align techno-
logical ambitions with strategic foresight.

Warm regards,
Martin Thormählen & Daniel Grothues

Martin Thormählen Daniel Grothues

Chief Technology Officer Chief Architect
Business Technology Primary Insurance
Munich Re ITERGO
[email protected] [email protected]
Munich Re Page 6/111
Tech Trend Radar 2025

Our approach

Process Result

Step 1: Screening Total trends

Analysis of trend developments

Compiling developments and new trends

for 2025 with external analysts’ reports and
internal market know-how, following
three golden rules:

1.  Technologies that potentially disrupt the

insurance industry.
2.  Technologies that potentially change
business models.
3.  Technologies that potentially have an
influence on the value chainand on internal
and external processes within the insur-
ance industry.

Step 2: Aggregation and trend selection Selected trends

Definition of trends

Aggregating data from screening processes

and defining the most relevant trends catego-
rised in five primary fields – and screening for
“outdated” trends.

Further drill-down and validation with market

data and identification of corresponding use
cases.

Disclaimer: References to companies do

not constitute or imply endorsement of any
company or organisation.

Step 3: Evaluation 5 Trend fields Hold Assess Trial Adopt

Assessment of impact and relevance 36 Trends

Classifying trends according to their level of (11 new,

relevance for ERGO and Munich Re. 4 transformed,
9 adopted,
Analysing trend impact along the insurance 5 niche,
value chain and deriving opportunities and 2 expired trends)
risks.
Munich Re Page 7/111
Tech Trend Radar 2025

How to read this report

ne

Trend fields
w
AI
-A

e
ed Medicin
ug

nce
tics
new
me

lthcare

er ie
nt

an

Analy
new
ed

it y
m

Exp
AI Medicine
AI
So

t iv
Digital Hea
hy Hu

Personaliz
Se

vioral
f tw

ec
te d
Leg
arc

nn
are

g
nec
Healt

Co
e

t in
hE

al T

B eha

ers
ne

En

pu
us
ngi

C on
w

t av
gin

om
it o
ch
nes
ee

Me

y
tC
i qu
AI

it
w
ne
r in

al
G

en
Ag

Ub

Re
en

AI

bi
s
en
er

ed
on

Am
G ti
at

ts

ov

nd
AI ac
iv

De er er

te
e

m na AI t
AI

Ex
oc nc e s In
g
ra
tiz e Ed ou
m
Sy at no to
nth io
n to yp y
eti Au Cr om
Qu cD & c on
ant at a er at a
E
um
C yb D
I n su Co ze d
ra n c mp
ut i nt rali
e AP
I St
ng
D ece lian
ce
and
ardi C omp
zat i om at e d
Artifi on Au t
cial G
enera
l Intell ts
igenc l Ass e
e Digit a
new Spatial Inte
lligence efense
Deepfake D

1 Data & AI Non-Human Identity new

Hold Assess Trial Adopt
Digital Imm
une System
2
Rede
fin ing In
dustr
ies
Cro
p In
t elli
g en
Ne ce
wE new
n er
gy
Tra at S
d eS cal
e
ec
El ret
ec Pro
tri t ec
H cV t io
yd eh n
ro i cl
g en es
H

Ec ne
um

on w
om
an
Au

oi

y
to

d
Cl

no

Ro
im

b
at e

ot
ou

s
sM
Re

ob
s ilie

ilit
nc

1 2
e

ne
w
ne

The report is structured Adopted, transformed, niche,

w

Trends are provided with

along five trend fields a trend maturity or expired trends

− Data & AI Adopt − Adaptive AI (transformed into − Privacy Enhancing Tech (niche)
− Healthy Human Start taking full advantage of this AI Agents) − Digital Identity (adopted)
− Connected Experience technology − Data Centric AI (adopted) − Blue Economy (part of Munich Re
− Cyber & Crypto
− Knowledge Graph (adopted) Marine Radar)
− Redefining Industries Trial
− Data Mesh (adopted) − Business Resilience (adopted)
First initiatives should be under-
wayin the most affected business − Human Enhancement (niche) − In Space Economy (niche)
areas − Robotic Health (niche) − Autonomous Things (transformed
− Biomanufacturing (niche) into Autonomous Mobility)
Assess − SuperApps (adopted) − Smart Cities (expired)
Consider what it maymean for − Automotive IoT (adopted) − Robotic Revolution (transformed
your business
− Data Fabric (adopted) into Humanoid Robots)
− Edge Computing (transformed − Connected Motor Insurance (adopted)
Hold
Add it to your watch list into Edge AI) − Carbon Capture (expired)
Munich Re Page 8/111
Tech Trend Radar 2025

How to read this report

Orientation

Navigation and links

All pages are linked in the table of contents.

The arrow → icons allow you to navigate through

the brochure.

Underlined text indicates links that will take you

to the corresponding website.

Colour coding

Each trend is assigned a color according to the

following scheme:

Data & AI
Healthy Human
Connected Experience
Cyber & Crypto
Redefining Industries
Munich Re Page 9/111
Tech Trend Radar 2025
→ Data & AI
→ Healthy Human
→ Connected Experience
→ Cyber & Crypto
→ Redefining Industries

1 Data & AI
Munich Re Page 10/111
Tech Trend Radar 2025

Data & AI

Trend evolution and impact on the value chain

high medium low

Trends Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

& Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Data & AI
Spatial
hold
Intelligence
Artificial
General hold
Intelligence
Insurance
API Stand- adopt
ardization
Quantum
assess
Computing
Synthetic
hold
Data
AI Democra-
adopt
tization
AI
adopt
Governance
Generative
adopt
AI

AI Agents trial

AI-Aug-
mented
trial
Software
Engineering
AI Search
adopt
Engines

Legal Tech trial

Munich Re Page 11/111
Tech Trend Radar 2025

Data & AI

Data, AI, and the future

of insurance: Navigating the
intersection of technology
and expertise

Not a day goes by without AI capability Generative AI is transforming the way how we classify and
extract data from documents supplied by brokers, surveyors
updates or breakthroughs in infra- and insureds. As predicted by Andrew Ng in 2024, the technol-
structure, foundational models, AI-based ogy has advanced to AI systems with rudimentary reasoning
and memory capabilities, which allows insurers to leverage AI
systems and new AI investment for more complex tasks in e.g. underwriting triage and claims
announcements. And many of these can triage. And it has also boosted legal technology, helping law-
yers and legal teams in all industries to create and review legal
be translated directly into impacts for documents.
our insurance industry: how customers
The hype around AI Agents will plateau in 2025, when insurers
search for insurance products and realize that they first need to differentiate between plain auto-
services will be disrupted by AI-based mation, AI workflows, and truly autonomous AI Agents. The
current language-based LLMs, LMMs and AI Agents can’t
solutions – which means that relying on understand concepts like 3D or gravity. Accordingly, we have
classical search engine optimization now added Spatial Intelligence to our radar and are curious to
see how this technology will evolve. With all the excitement
(SEO) will no longer be enough. surrounding AI in the media, at home and in our offices, we
again address the paramount importance of practicing and
monitoring AI Governance.

The trend field was deliberately named Data and AI to empha-

size three aspects: that without a solid foundation of data,
neither pricing nor AI solutions can be meaningfully grounded
and developed; that without evaluation data, AI solutions
cannot be sufficiently validated; and most importantly, that
AI is not needed if partners exchange digital data directly,
which also brings the trend of insurance API standardization
back into focus.
Munich Re Page 12/111
Tech Trend Radar 2025

Data & AI

Spatial Intelligence
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Hold & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Enhanced risk assessment through advanced geospatial data analytics − Data privacy concerns with increased use of geospatial data
− Proactive disaster management with real-time geospatial data − Dependence on technology for risk assessment and management
integration decisions
− Customized insurance products for unique regional communities and
businesses

Overview The power of geo-spatial data

AI has progressed from speech and text to 2D image and The insurance industry is undergoing a significant trans-
video reasoning and generation, and is now expanding to formation in risk assessment and risk management, driven
create spatial intelligence, which refers to the ability to under- Zby technology. Spatial intelligence, the ability to understand
stand and navigate three-dimensional spaces and environ- and interpret spatial relationships between different entities,
ments. The next frontier involves developing a “Large World is emerging as a critical technology trend. This trend has been
Model” that not only comprehends the real world in 3D but fuelled by advances in geospatial analytics, satellite imagery,
also incorporates physical laws like gravity. However, the and Artificial Intelligence (AI). As natural disasters become
timeline for the availability of this model as a foundation for more frequent and severe, insurers will adopt innovative
spatial computing remains uncertain. Insurers will be able solutions that provide a clearer understanding of the risks
to not only visualize and interpret spatial relationships but to associated with specific properties.
develop more accurate pricing models and better risk assess-
ments. As catastrophic events become more frequent, Leading experts in AI research, such as Fei-Fei Li, inaugural
integrating spatial intelligence offers a path to better risk Sequoia Professor at Stanford University who also established
management and operational efficiency, ultimately leading the famous ImageNet, emphasize the importance of spatial
to improved customer satisfaction and reduced losses. intelligence in enabling AI systems to process visual data,
make predictions, and act on those predictions in real-world
environments. Li emphasizes the importance of spatial intelli-
gence in enabling machines to interact meaningfully with
humans and their environment. Her work focuses on develop-
ing AI models capable of understanding three-dimensional
spaces, bridging the gap between visual perception and physi-
cal interaction (more here). Li asserts that achieving true spa-
tial intelligence will open up new applications in various fields,
including healthcare, robotics, and urban planning.

Researchers at Texas A&M University have also explored the

potential of spatial reasoning capabilities in AI, demonstrating
the need for enhanced algorithms to improve machine under-
standing of complex three-dimensional transformations. Their
findings suggest that while current AI models show promise
in visual learning tasks, they still struggle with higher-level
spatial reasoning tasks compared to humans (more here).
Munich Re Page 13/111
Tech Trend Radar 2025

In recent years, the insurance industry has made remarkable

progress in leveraging cutting-edge technologies such as
Artificial Intelligence (AI) with Natural Language Processing
(NLP),Large Language Models (LLMs), and Multimodal
Models. In addition, advances in aerial imagery and drone
technology have greatly enhanced our ability to collect and
analyze data.

However, despite these achievements, the ultimate goal of

achieving true spatial intelligence remains elusive. The devel-
opment of a comprehensive “world model” that fully integrates
and understands the complexities of our physical environment,
including spatial relationships, gravity, and dynamic interac-
tions, is still on the horizon. While we have made significant
progress in various areas, creating a sophisticated world model
that can accurately simulate and predict real-world scenarios
remains a challenging and ambitious goal. Initially relying on Conclusion
historical data and broad geographic zones to assess risk,
the industry is now moving toward more precise models that Spatial intelligence is expected to greatly impact the insurance
incorporate real-time geospatial data. Technologies such as industry by enhancing risk assessment capabilities, stream-
satellites, drones, and real-time data sets are providing insur- lining underwriting processes and improving claims manage-
ers with unprecedented visibility into the risks surrounding ment efficiency. By integrating insights from expert research
facilities. Recent advances have enabled insurers to quickly into AI's evolving understanding of spatial relationships with
assemble comprehensive data sets. For example, the inte- perspectives from other experts in the field, insurers can better
gration of radar-based satellite imagery with drone-captured address the challenges posed by climate change and cata-
visuals as well as creating digital twins allows insurers to strophic events.
conduct detailed property assessments much more efficiently
than traditional methods. However, it is important to recognize that we are still in the
very early stages of truly understanding spatial relationships
in the context of both human cognition and machine learning.
Outlook for spatial intelligence True spatial intelligence – where machines fully understand
in insurance complex environments as humans do – remains a goal for
future research and development. As we move towards this
The future outlook for spatial intelligence in insurance is goal over time, improving our understanding of spatial intelli-
promising, but also acknowledges that we are still in the early gence will be key to unlocking new opportunities not only in
stages of truly understanding and fully modeling spatial the insurance sector, but in many other areas where effective
relationships, including gravity, dynamics, and 3D. As climate decision-making relies on accurate interpretations of space.
change continues to reshape the risk landscape, insurers
equipped with advanced analytical tools will lead the way in
providing effective coverage solutions. The global insurance
market is expected to grow significantly over the next decade,
driven by increasing demand for innovative products that
address emerging risks (RMP).

In addition, as consumers demand more personalized services

and proactive risk management solutions, insurers that lever-
age geospatial intelligence will be well positioned to meet
these expectations. Fei-Fei Li's work on AI emphasizes the
importance of machines understanding physical space
through advanced algorithms capable of extrapolating visual
data into three-dimensional environments. This trend aligns
closely with the insurance industry's need for improved risk
assessment capabilities. By leveraging AI alongside geospatial
property intelligence, insurers can gain unprecedented insight
into location-specific risks. By integrating geospatial property
intelligence into daily operations, insurers can navigate the
complexities of climate change with greater confidence and
precision.
Munich Re Page 14/111
Tech Trend Radar 2025

Data & AI

Artificial General Intelligence

Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Hold & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Potential to revolutionize most industries − AGI could act in ways that are unpredictable and/or harmful
− Improved information extraction and analysis if not properly controlled.
− Solution for technical tasks like coding − It may be difficult to align AGI systems with our human values.
− Automation and optimization can reduce costs and − Some experts worry that AGI could pose existential risks to humanity.
improve overall efficiency.

Overview The definition of AGI

There is no universally accepted definition of Artificial General When it comes to the topic of AI, the question as to when we
Intelligence (AGI), but it is associated with an AI system that will reach Artificial General Intelligence (AGI) has been asked
can exhibit human-like capabilities in most domains. While we increasingly of late. As there is no universally accepted and
don’t know when or if we will reach AGI, AI progress in recent accurate definition of AGI, it may be more helpful to think of
years has been faster than expected, and there is confidence AGI as a milestone in the development of AI. This development
that it will continue for some time. started with simple rule-based and statistical systems and
progressed to machine learning (ML) systems.
It is assumed that we will see five levels of AI until we reach
AGI, and that each level will enable different use cases. These The next milestone came in the form of narrow AI systems, i.e.
levels can also help us to understand where we are in the systems that have human-like capabilities in a very narrow
development of AGI and how fast we are moving. In particular, domain like playing chess, but are useless in other domains.
they may help to anticipate what will come next and, as a The next expected milestone is artificial general intelligence
result, how to prepare for upcoming opportunities. (AGI) where an AI system can exhibit human-like capabilities
in most domains. Further in the future, artificial super intelli-
gence (ASI) is envisioned, where an AI system would exhibit
super-human capabilities in most domains.

To make the AGI milestone more precise, recent definitions

of AGI focus either on the kind of capabilities, like OpenAI’s
approach, or the level of capability, like Google’s approach.
We will use here the leaner approach from OpenAI. Finally,
the discussion is moving away from a one-dimensional
human–AI comparison to a more nuanced multi-dimensional
comparison, which is often referred to as the “jagged frontier”.
Munich Re Page 15/111
Tech Trend Radar 2025

The path to AGI

If we follow the OpenAI definition of AI, we should see
five levels of AI before we reach AGI at the fifth level:

Level 1 - Conversational AI: The system can interact in

conversational language with people.

Level 2 - Reasoning AI: The system can perform basic

problem-solving tasks through reasoning.

Level 3 - Agents: The system can take actions and use tools
on a user’s behalf to complete tasks.

Level 4 - Innovators: The system can develop innovations

and improve existing processes.

Level 5 - Organizational AI: The system is capable of

performing the work of an entire organization. Emerging risks
Right now, it is believed that we have reached level 2 and that As AI systems become more and more advanced, it becomes
we entered level 3 in 2025, much faster than most people more and more difficult to ensure that the output and actions
would have believed five years ago. The question as to whether of the AI system are aligned with certain principles, which is
and when we will reach level 4 and 5 is open for debate. Each referred to as the alignment problem. An ambiguous prompt
of these levels offers different opportunities for companies. may lead to unintended consequences, especially if the system
Level 1 enables information systems, where either the system is at Level 3 or higher, meaning that it can take actions. Some
answers a question directly or retrieves relevant information. people are even arguing that these systems could give rise to
Level 2 enables reasoning, helping to solve technical problems existential risks. In order to avoid unintended consequences,
(e.g. coding or medical diagnosis/treatment). Level 3 enables guardrails, strictly controlled access to data and systems, and
agentic process automation and information analysis (e.g. proper architecture is essential. Understanding and monitor-
deep research by Google, OpenAI and Perplexity). ing the reasoning steps for these systems and the conversation
between agents will also become key.

Current technological paths

to improve AI Implications for companies
The first and initial strategy to improve current AI systems is AI, especially at the level of AGI, is a powerful tool that will fun-
to increase Training Compute (TC) i.e. the compute resources damentally change the way we work and do business. While it
during model training. While some progress is being made offers many opportunities, these will have to be balanced with
here, especially combined with higher quality data, it seems appropriate risk measures, as with any other powerful technol-
that returns are diminishing. In more recent times, Test Time ogy. The insurance industry has a specific role to play here with
Compute (TTC) has been used, where we do not increase regard to risk assessment and pricing. It will be crucial for all
the compute for training the model but rather for inference, i.e. companies to understand that each of the five AI levels ena-
the model is given more time and resources to “think” before bles different use cases and that timing is critical to creating
it answers. One example of this paradigm is the “o” family of economic value. The five AI levels can be used to understand
models from OpenAI. This currently seems to be more cost- where we currently are and how quickly the technology is mov-
effective than the first strategy. The third strategy involves ing, but in particular, they can also help to anticipate what will
using a set of specialized models that collaborate instead of come next, how to prepare and how to realize upcoming
a single general-purpose model. opportunities.

Another strategy is to allow AI systems to use tools (e.g. code

execution), dramatically increasing performance in certain
domains (e.g. math) with the potential to reduce hallucinations
if grounding through internal data or internet access is
allowed. There are other more elaborate strategies, especially
to increase the data available or to improve the architecture of
the AI models.
AGI development is moving fast.
Understanding it helps us to prepare
and realize upcoming opportunities.
Andreas Nawroth
Leading Expert AI & Quantum
Munich Re Page 16/111
Tech Trend Radar 2025

Data & AI

Insurance API Standardization

Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Adopt & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Significantly better customer experience due to faster response times − Initial implementation efforts require careful execution, including
− Improved efficiency and savings by automating data entry and cybersecurity protection
reducing errors − An API-driven tech stack creates pressure on the underlying legacy
− Critical interoperability capability for seamless data exchanges with systems, as business partners demand “always on” integration with
new distribution partners no downtimes for maintenance efforts

Overview The lack of data standards

Insurance data standards are key to breaking down the silos The traditional insurance industry has a significant and grow-
between distribution partners, primary insurers, brokers and ing problem in 2025: The relevant data is still mostly shared
reinsurers. Insurance API standardization creates a common between the various participants in the insurance value chain
way for insurance systems to communicate, improving inter- in a very outdated fashion. Distribution partners like MGAs
operability, efficiency and security. It makes collaboration and brokers, as well as tied agents (humans), capture data,
faster and easier, whether it involves transferring risk, handling create PDF files and submit them to their insurer via email or
claims or developing new products. In the long run, this stand- portals. Yet another distribution partner also creates PDFs but
ardization will lead to significant cost savings, improved cus- uses a different document template. The ecosystem is con-
tomer experience and providing greater flexibility to innovate. stantly readjusting, testing and validating the exchange of
In essence: it is ultimately the implementation of a universal data, as it lacks a common data exchange format that is
language that everyone in the insurance industry can use to accepted by most players.
share information.
Technological advances are now making the situation worse:
the hype surrounding Large Language Model (LLM) and Large
Multimodal Model (LMM) systems makes insurers blind to
innovating in the industry with a first-principles approach.
Instead of pushing for agreement on a common data standard
for the industry, most insurers are spending millions of euros
on extracting data from incoming documents, creating
human-in-the-loop (HITL) systems, and generating additional
overhead through their efforts to ensure that the new AI-based
systems are highly accurate and ethically acceptable, and
because they need to ensure that the systems don’t deteriorate
over time.

Like the AI industry, the insurance industry also needs to apply

first-principles thinking: what are the mandatory key aspects
of risk transfer, accounting and claims handling? What is the
minimum level of processes, people, data and technology
required to exchange data between clients, intermediaries like
brokers and agents, and (re)insurers? What is only increasing
complexity and therefore adding avoidable costs? How can we
avoid becoming even more dependent on technology providers?
Munich Re Page 17/111
Tech Trend Radar 2025

The solution
Data standards have the most profound impact on people,
processes, technology and data.

Efficiency impact on companies Efficiency impact on technologies

When adopted company-wide, data standards streamline Adopting data standards allows companies to implement
operations, enhance team collaboration and increase produc- solutions faster and at lower cost by using third-party products
tivity. They reduce the time spent on data collection, allowing instead of proprietary ones. As standards become more com-
teams to focus on higher-value tasks and thereby boosting mon, vendors will be motivated to adhere to them, fostering
talent retention and satisfaction. Adopting data standards competition based on value and feature enhancements.
allows companies to direct their top talent towards high-value
opportunities like customer acquisition and business develop- Efficiency impact on analytics and AI
ment, rather than mundane tasks like data entry. This shift Data collection and sharing via data standards will also
enables a focus on strategic initiatives. improve overall data accuracy and consistency. This will not
only allow more accurate reporting and AI-based predictions
Efficiency impact on processes but also enable insurers to innovate on the next level of feature
Data standards streamline operations, reduce transaction engineering.
capacity needs, and enable companies to achieve straight-
through processing (STP). They create a common understand-
ing of data, allowing it to be collected once and reused across
different business areas. This facilitates semi- or even ful-
ly-automated systems where information flows seamlessly
from the first interaction to the end of the transaction without
the need for manual intervention.

Figure: Insurance Stakeholder Map

Insurance stakeholder who will benefit from data standardization.

Insured

Digital Agent &

Player Broker

MGA Insurer
Data
Standard

Claims
Reinsurer
TPA

Insurance
Regulators
associations

Source: Munich Re
Munich Re Page 18/111
Tech Trend Radar 2025

Real-life examples
ACORD data standards
The Association for Cooperative Operations Research and
Development (ACORD) is a globally active standards-setting
body for the insurance industry, as well as for related financial
services industries. Many global reinsurers, insurers and bro-
kers are members of the so-called Ruschlikon initiative, where
insurance data standards are discussed and defined.

LMA
The Lloyd’s Market Association is advocating for the adoption
The future of API Standardization
of ACORD data standards throughout the insurance risk in insurance
placement and processing lifecycle.
As the insurance industry continues to embrace the digital
BiPRO e.V. transformation, the role of standardization will become
BiPRO e.V. is a registered association in Germany that pro- increasingly important.
motes data and exchange standardization and implements IT
solutions for process optimization between brokers, agents Future trends may include:
and insurance companies. More here. − Regulatory mandates: Governments and regulatory bodies
may introduce mandates requiring API standardization to
FIDA ensure consumer protection and market stability
On December 4, 2024, the European Council agreed to make − Increased collaboration: Industry stakeholders may collabo-
consumers’ financial data more accessible in order to foster rate more closely to develop and adopt universal API stand-
transparency and comparability. Read here. ards, promoting a more integrated ecosystem
− Global standardization efforts: International organizations
working to establish global data and API standards in order
to facilitate cross-border insurance operations and services

Data and API standardization is the key to unlocking the full

potential of digital transformation in the insurance industry.

By promoting interoperability, efficiency and innovation, stand-

ardized APIs will be a key means of helping insurers meet
evolving customer expectations and navigate the complexities
of the modern digital landscape. While challenges exist, col-
laborative efforts and global initiatives should pave the way for
a more connected and efficient insurance ecosystem.
Munich Re Page 19/111
Tech Trend Radar 2025

Data & AI

Quantum Computing
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Assess & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Quantum Computing (QC) could solve complex problems in weather − Timing of quantum advantage difficult
and climate predictions, battery development, finance, pandemics, − QC may endanger current technologies such as IT-security protocols
logistics, cyber and AI. and blockchain
− First early applications in estimating extreme risks or asset pricing − Access to hardware and stability of early quantum devices may
− High probability of substantial advancement in QC technology over the become a bottleneck
next few years

Overview Unlocking quantum potential

Quantum computing (QC) has made significant strides in The unprecedented power of Quantum Computing (QC) could
recent years. It has seen notable advancements in qubit counts trigger a new wave of technological advancement over the
and stability. Despite progress, quantum computing still faces next five to ten years. Quantum computing has made signifi-
significant hurdles. Current quantum computers do not yet cant strides in recent years, evolving from a niche field to one
offer a tangible advantage over classical computers for most with substantial potential for solving complex problems. It is
applications. The field continues to attract substantial invest- currently based on technologies, like superconducting circuits,
ment, especially in the US and China. While quantum comput- trapped ions, neutral atoms, photons or even more exotic tech-
ing is not yet ready to replace classical computing, its rapid nologies. Recent advances are the successful introduction of
development and the growing interest and investment suggest error correction, earlier than expected and the transition from
a bright future ahead. (noisy) physical qubits to logical qubits. Also AI is influencing
the progress of QC. GenAI progress is on one side overshad-
owing the substantial progress of QC, on the other side, com-
panies like NVIDIA and DeepMind, have come up with ideas,
how to integrate GenAI with QC.

Quantum computers use so-called qubits, which, contrary to

classical bits, can represent 0 and 1 at the same time. Using a
process called “entanglement”, the qubits are combined to
a quantum computer. Due to his special combination of the
qubits the computational capabilities can increase exponen-
tially. This exponential increase in computational power offers
many opportunities. It holds the promise to solve complex
problems in long-term weather and climate predictions, bat-
tery development, finance, drug research, logistics, cyber and
AI. First results are pointing to early commercial applications
in estimating extreme risks or asset pricing using Monte Carlo
methods or quantum machine learning. In 2021, ten leading
German companies, including Munich Re, set up the Quantum
Technology & Application Consortium (QUTAC) with the
objective of accelerating the transition of QC from research
labs to useable industrial applications. Meanwhile 14 compa-
nies are part of QUTAC.
Munich Re Page 20/111
Tech Trend Radar 2025

The timing of quantum advantage for different use cases is

not easy to predict. But given the exponential benefits, the
different technological paths available, the current progress
and the strong investment trends, first useful applications are
expected by 2030 (more here and here). While the potential
for QC is high, QC may also create security risks for current
technologies like encryption.

Access to hardware and stability of early quantum devices may

also become a bottleneck for industrial applications. Overall,
while quantum computing is not yet ready to replace classical
computing, its rapid development and the growing interest
and investment suggest a bright future ahead.

“Though AI is currently
the main focus, Quantum
Computing is a critical
technology, that will have
as much impact as AI.”
Andreas Nawroth
Leading Expert AI & Quantum
Munich Re Page 21/111
Tech Trend Radar 2025

Data & AI

Synthetic Data
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Hold & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Enables creation and analysis of data sets without compromising − The complexity and nuances of real-world data might not be fully
sensitive information captured
− Synthetic data can increase the quality and robustness of model − Challenge of ensuring the representation of the real-world scenarios
training − Potential misuse or mis-representation leading to ethical dilemmas
− Reduces costs and risks of traditional data collection

Overview Synthetic representation of the

In the realm of rapidly advancing developments in generative
real world
AI models, data becomes a key differentiator in AI model train- Synthetic data, which mimics real-world datasets promises
ing and cross-domain usage. Conventional data collection and insurers potential to unlock new efficiencies and capabilities.
usage often pose significant risks to sensitive and identifiable It enables privacy-preserving data sharing, ensuring that sen-
information. Synthetic data, created by sophisticated algo- sitive customer information remains protected while fostering
rithms and statistical models, offers a game-changing solu- new solutions and collaboration across teams and external
tion. By replicating the characteristics of real-world datasets partners. It allows the acquisition of annotations, which are
without compromising sensitive information, synthetic data is difficult or impossible to obtain in the real world. By generating
a novel way to approach data-driven innovation, mitigating anonymized yet statistically identical datasets, synthetic data
data privacy, scarcity, and bias limitations. reduces regulatory burdens and enhances data accessibility.

A potential game changer for insurers

Applications for synthetic data is evolving across industries,
including insurance, finance, healthcare, automotive and retail.
Insurers could benefit in manifold ways. In risk modelling,
synthetic data offers the ability to simulate rare events, such
as natural disasters or economic shocks that are underrepre-
sented in historical datasets. This enhanced capability allows
for better preparation and more precise underwriting prac-
tices, especially in regions with limited historical data. When
it comes to fraud detection, synthetic data can be applied to
generate diverse scenarios and to identify anomalies and
fraudulent claims with greater accuracy.

These datasets enable more robust defense against evolving

fraud techniques. Furthermore, synthetic data can be also
generated to enhance the customer experience by creating
highly personalized synthetic marketing materials, such as
targeted messages, avatars and videos for campaigns.
Munich Re Page 22/111
Tech Trend Radar 2025

GenAI as the powerhouse of

synthetic data
GenAI models can learn complex patterns and relationships
within data, allowing for the generation of synthetic datasets
that are virtually indistinguishable from real-world data.
Synthetic data is also being used increasingly to fine-tune
AI models, such as large language models and computer vision
systems. The AI company Groq, Inc, features models trained
entirely on synthetic data that outperform current competitors
with a traditional approach in this field.

Outlook for Synthetic Data in 2025

The synthetic data ecosystem features a mix of established
players and innovators. Startups like Mostly AI are driving
advancements in data generation tools. Cloud providers such
as Google and Amazon are integrating synthetic data capabili-
ties into their platforms, enabling scalable adoption and
exploring the new frontiers of digital twins.

Collaboration between insurers, technology providers, and

researchers will be essential to fully realize the benefits of
synthetic data and to establish it as a fundamental pillar of
the industry’s future.
Andreas Schumacher
Project Manager, Artificial Intelligence

Conclusion
Synthetic data holds the potential to drive operational effi-
ciency and customer satisfaction. Nevertheless, challenges in
ensuring data accuracy, quality and maintaining stakeholder
trust remain, and potential for misuse could also emerge. The
rise of advanced generative models are enabling the creation
of highly realistic and diverse synthetic data. This, in turn, is
fuelling the development of more accurate and reliable insur-
ance models.
Munich Re Page 23/111
Tech Trend Radar 2025

Data & AI

AI Democratization
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Adopt & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− AI Democratization broadens AI access by reducing dependence on − Using externally hosted models creates a risk of intellectual property
technical experts (IP) loss, requiring tailored guidelines
− AI Democratization enables small to mid-sized AI use cases − Over-reliance on easily accessible AI tools may lead to misuse or errors

Overview Empowering people to use AI

AI Democratization aims to make Artificial Intelligence (AI)
effectively and responsibly
more accessible to non-technical staff in the insurance indus- AI can deliver significant competitive advantages, from per-
try, such as claims handlers and underwriters, through user- sonalized pricing to predictive claims management. However,
friendly tools like no-code platforms and Large Language AI adoption has been anything but uniform – often confined to
Models. This approach accelerates problem-solving, improves technical experts, it can create silos that slow innovation and
decision-making, and boosts productivity across organiza- limit value creation. Functional experts – such as underwriters,
tions, while robust training ensures the ethical, safe, and effec- claims handlers, and customer service representatives – play
tive use of AI technologies. critical roles in the insurance value chain but often lack the
tools or knowledge to effectively harness AI in their workflows.

AI Democratization breaks down these barriers, equipping

employees across the organization with the tools needed to
leverage AI. No-code/low-code platforms allow non-technical
staff to develop and deploy AI solutions, accelerating problem-
solving and removing bottlenecks. Embedded AI integrates
automation and predictive analytics into workflows, enabling
employees to make faster, data-driven decisions. In turn, LLMs
enhance productivity by summarizing claims, analyzing poli-
cies, and generating personalized communications at scale.

However, broader AI access is accompanied by certain risks,

e.g. misuse, over-reliance, and amplified bias in decision-
making. Externally hosted GenAI models also introduce
IP risks, requiring clear governance. Robust training programs
and compliance measures are essential to mitigating these
risks. Employees must receive both technical training and
guidelines on using AI ethically. Without such safeguards,
insurers run the risk of unintended consequences that could
erode trust and create operational vulnerabilities.
Munich Re Page 24/111
Tech Trend Radar 2025

Data & AI

AI Governance
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Adopt & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risk

− Interplay between offensive and defensive AI Governance allows a bal- The regulatory landscape around AI is being developed and some points
ance to be struck between compliance and AI advances in an enter- have not yet been fully clarified.
prise.
− Adaptive AI Governance frameworks allow the changes in an ever-de-
veloping legal environment to be accommodated.

Overview Key challenges of designing AI

AI Governance is a part of enterprise governance dealing with Governance
enabling AI development and purchase in a compliant way.
This includes fostering AI adoption (offensive AI Governance) In last year’s Tech Trend Radar, we described the emerging
and compliance and risk management-related measures regulation landscape and its complexity, and proposed a four-
(defensive AI Governance). The combination of offensive and step approach to establishing AI Governance on an enterprise
defensive governance is the key to promoting AI advancement scale. The focus was on the design phase, where a lot of deci-
in a compliant way. The regulatory landscape around AI is still sions on the structure of AI Governance (e.g. centralized vs.
challenging, which is being partially covered by the appear- decentralized approach and degree of regulation) are made.
ance of AI-related standards and codes of practice. This This year, we will outline specific key challenges that the mar-
requires adaptive AI Governance frameworks. ket faces when specifying AI Governance frameworks and give
concrete recommendations on how to overcome them.

Smooth interplay between offensive

and defensive AI Governance is key
Offensive governance in an enterprise traditionally covers
enabling the achievement of business goals, e.g. business
revenue or process optimization. In terms of AI Governance,
the offensive part might take the form of setting up develop-
ment platforms, employee training sessions, and (centralized
or decentralized) support in implementing AI use cases.
Defensive governance, on the other hand, concentrates on
managing AI-induced risks and ensuring compliance. Speci-
fically, for AI Governance this might be implemented as AI
compliance gates and enterprise-wide binding guidance on
AI, as well as the degree of regulation in the different stages
of the AI system lifecycle.
Munich Re Page 25/111
Tech Trend Radar 2025

As with other emerging technologies, businesses initially focus

on the offensive side of AI Governance, identifying use cases
and deploying the most promising ones in production. Simi-
larly to other tools like Big Data or Cloud Computing, the regu-
latory landscape lags behind the technological advancements,
allowing quite a broad field for experimentation. As AI slowly
became a standard part of the value delivery, a complex inter-
play of regulations around AI and predictive models, most
notably, the EU AI Act, also evolved. That led businesses to
establish defensive AI Governance structures. This became
even more relevant in 2025 when some parts of the EU AI Act,
e.g. prohibited AI practices and General Purpose AI Model
requirements, became binding.

When establishing an AI Governance structure, companies are

faced with the need to balance the offensive and defensive
sides of AI Governance to strike a balance between allowing
the rapid introduction of AI capabilities into business pro-
cesses and, at the same time, guaranteeing compliance with
regulations and managing risks connected with the use of AI
in production. This is probably the biggest challenge and, at
the same time, the major opportunity in governing the use of
AI in an enterprise. There is, unfortunately, no one-size-fits-all
solution, and it is also probably a moving target requiring con-
stant re-evaluation of the setup. This strongly depends on the
AI strategy and risk appetite of an individual enterprise. Fur-
thermore, the key to success is constant alignment between
defensive and offensive AI Governance. An example of a bal-
anced approach might be to allow for a broad field for experi-
mentation at an early stage, followed by a stricter compliance
gate before putting AI into production.

“The key to successful AI The regulatory landscape around AI is

Governance at enterprise still developing and needs clarification
scale is now to find a balance
between defensive and The regulatory landscape around AI is still being established
and continues to develop. One of the major challenges is the
offensive AI Governance,
need for clarification of the practical implications of the laws.
thereby ensuring compliance One example is the definition of an AI system in the EU AI Act.
and promoting AI advances
at the same time.” The need for clarification is being partially covered by position
statements and the emergence of related standards, e.g. the
Dr. Alina Nizamutdinova standard for AI developed by the National Institute for Stand-
Project Manager AI ards and Technology in the US. We expect to see more clari-
fication in the future, which at the same time will make the
implementation of requirements more actionable, but will also
narrow the action space for that.

An adaptive AI Governance framework

is a necessity
The developing regulation landscape, as well as the rise of
AI-related standards, implies a need for adaptive AI Govern-
ance frameworks. A robust AI Governance framework must be
able to respond quickly to compliance-related changes, while
also overseeing new AI trends. In addition to flexibility by
design, continuous monitoring and adaptation, a strong feed-
back mechanism from internal stakeholders paired with
benchmarking against industry contribute significantly to
resilient and, at the same time, adaptive AI Governance.
Munich Re Page 26/111
Tech Trend Radar 2025

Data & AI

Generative AI
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Adopt & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Improved risk assessments and personalized pricing models, which − Insurers must navigate evolving regulatory environments (e.g., EU AI
allow insurers to offer tailored policies that meet individual customer Act), which can be challenging, costly and time-consuming
needs − AI models are becoming increasingly specialized, making integration
− Optimization of operational aspects like customer interaction and interoperability key aspects to prevent rapid obsolescence
and claims handling, which leads to higher efficiency and customer
satisfaction

Overview Adoption of Generative AI in the

The integration of Generative AI into the insurance industry insurance industry
is yielding promising results, with many companies success- Over the past year, Generative AI has made significant strides
fully harnessing its power to enhance aspects like risk insights, in transforming the insurance industry, building on the founda-
efficiency, and customer experience. Recent advances in train- tion established in previous years. This progress has made the
ing data, specialized models, and reasoning approaches have technology more accessible and its applications more clearly
significantly improved the quality and capabilities of Genera- defined, paving the way for increased adoption and innovation
tive AI, making it an attractive solution for insurers seeking to in the sector.
boost their efficiency and effectiveness and stay competitive.
Specialized models have emerged, demonstrating the technol-
ogy’s ability to tackle specific insurance-related tasks, includ-
ing insight generation, chatbot features, and the automation of
routine processes. Generative AI-powered chatbots can e.g.
optimize customer service, while automated data analysis can
support underwriting and claims processing. As the technol-
ogy continues to evolve, its impact on the insurance industry is
likely to become more pronounced, particularly in areas where
data-driven decision-making is critical.

The adoption of Generative AI models is set to bring signifi-

cant improvements to the insurance industry. Insurers can
expect to see enhanced policy customization, improved fraud
detection and increased operational efficiency by leveraging
these technologies. While the potential benefits for insurers
are substantial, it is crucial to address key challenges like
data privacy, integration, and the models’ limitations, like bias
and hallucinations. As insurers continue to adopt and refine
these models, they can anticipate meaningful gains in cus-
tomer experience and overall business performance, as already
recognized by 40% of insurance executives. Responding to
growing customer expectations and operational demands
requires insurers to proactively adopt and integrate generative
AI technologies to stay competitive in the rapidly evolving
environment.
Munich Re Page 27/111
Tech Trend Radar 2025

Fine-tuning Large Language Models (LLMs) with their own

data allows insurers to adapt these models to their specific
needs. By calibrating them for e.g. industry-specific or com-
pany data, insurers can improve performance, reduce bias, and
increase accuracy. This customization enables more effective
policy analysis, claims processing and customer communica-
tion, ultimately driving business value and efficiency.

“Generative AI helps
insurers enhance risk
assessment, customize
policies, and optimize
claims handling.”
Dr. Matthias Kaper
Senior Data Scientist

Insurance companies have effectively integrated Generative AI Think harder: Reasoning models
to enhance operations. For example, Definity Insurance has
utilized Generative AI to streamline customer service at its The latest evolution in Generative AI has given rise to models
contact centers, reducing call durations and improving support that “think harder” and can engage in more complex reasoning.
quality. Canara HSBC Life Insurance has launched “OmniGen These models, called reasoning models, work by breaking
AI”, a tool that generates insights from various data points to down complex problems into smaller, more manageable
inform underwriting decisions. And Westfield Insurance has chunks. This allows them to analyze data in a more human-like
collaborated with IBM to streamline code explanation and way, identify relationships between different pieces of infor-
documentation, thereby boosting operational efficiency. These mation, and provide clear explanations for their conclusions.
initiatives showcase AI’s role in advancing efficiency and cus- Imagine a model that can think step-by-step, considering
tomer experiences in the insurance sector, despite often-chal- different approaches to a problem before coming up with
lenging regulatory environments. a solution. This process takes marginally longer but results
in more accurate and reliable outcomes. Its most notable
representative is OpenAI’s o1 model, which demonstrated
human Ph.D.-level accuracy for various scientific benchmarks.
The successor, o3, released in December 2024, surpasses o1’s
performance significantly but is also slower and much more
expensive.

The power of Small Language Models

Advances in training data have significantly improved the
quality of LLMs, enabling better generalization and perfor-
mance through increased diversity, size, and synthetic data
augmentation. The use of a substantially larger amount of
training data, which is now better curated and more thought-
fully selected, has also expanded models’ capabilities, allowing
them to learn and generalize more effectively. Furthermore,
the use of mixture of experts (MoE) models, which combine
multiple specialized models, has boosted performance.
Munich Re Page 28/111
Tech Trend Radar 2025

While LLMs of ever-increasing size have evolved in recent

years, the trend is also being influenced by the use of Small
Language Models (SLMs). Designed with a relatively small
number of parameters, SLMs perform impressively on
language tasks while consuming minimal computational
resources. They offer faster inference times and can run
smoothly on devices with limited processing power, including
smartphones. Their lower energy consumption not only
reduces costs but also minimizes environmental impacts.
While they may struggle with highly complex tasks and pro-
vide less nuanced understanding, they are sufficient for a vast
number of tasks. Notable examples include Microsoft’s Phi
series and Meta’s Llama series. SLMs are employed in chat-
bots, text summarization, sentiment analysis, language trans-
lation, and content generation – striking a balance between
performance and efficiency. Yet new architectures are appear-
ing that combine the advantages of both LLMs and SLMs,
where the strengths of both are used for the respective tasks.

“Strategically combining New closed- and open-source models

broad-scale intelligence of
Generative AI is rapidly advancing, with proprietary solutions
Large Language Models and open-source models competing for both the best perfor-
and precision of specialized mance and the highest efficiency. Open-source models reveal
models delivers accurate their details and can be deployed on their own infrastructures,
and cost-effective GenAI while closed-source models are not available to the public
and are commonly operated by the vendor itself. Examples of
solutions in insurance.” closed-source models are OpenAI’s GPT series, Anthropic’s
Claude and Google’s Gemini AI, Microsoft’s Copilot and
Andreas Schumacher
NVIDIA’s Fugatto.
Project Manager Artificial Intelligence

Open-source models include Meta’s Llama series, Databricks’

Specialized and multimodal models DBRX, and EleutherAI’s GPT-J or Mistral’s Mixtral and multi-
modal Pixtral models. DeepSeek’s open-source R1 model
Specialized models, such as coding and multimodal models, demonstrates competitive results with leading models at a
represent another significant advance in Artificial Intelligence. low training cost. Perplexity.ai has published a more neutral
Coding models are designed to assist with programming derivative: “R1 1776”.
tasks, including code generation, debugging and optimization.
Multimodal models can process and interpret multiple types of
data like text, images, audio, and even video – simultaneously,
Emerging trends and outlook
enabling more sophisticated and integrated applications. The convergence of Generative AI with computer vision and
In the insurance industry, these models offer transformative natural language processing (NLP) is transforming industries
use cases by combining and understanding the relationships across the board. By interpreting and generating multimodal
between various modalities. For automated underwriting, data, AI systems can now better understand and interact
multimodal models can analyze textual information alongside with the world.
images of properties or assets to assess risks more accurately.
In claims processing, they can expedite evaluations by inter- A prime example of this evolution is Google’s
preting photos of damage, generating reports, and even NotebookLM, an AI-powered research assistant that can sum-
detecting fraudulent claims through pattern recognition. This marize information and provide insights from user-supplied
not only reduces processing time but also improves accuracy documents, even in a podcast-like summary. Another promis-
and customer satisfaction. The rapid advances also address ing approach employs agentic systems where LLMs combine
growing integration demands and risks of obsolescence. various skills to complete abstract tasks.
Munich Re Page 29/111
Tech Trend Radar 2025

Data & AI

AI Agents
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Trial & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Efficiency gains in all areas, especially coding, customer engagement, − AI agents’ autonomy can pose security risks, like the unauthorized use
and data analytics of data or systems
− Ability to provide 24/7 customer support − AI agents can raise ethical questions, making human supervision
− Help in designing new insurance products essential
− Ensuring compliance with global data protection and AI regulations is
crucial

Overview AI agents – their potential

AI agents are advanced Artificial Intelligence systems AI agents will rapidly transform our world. According to
designed to autonomously complete tasks and make decisions Gartner, by 2028 one-third of all enterprise software applica-
based on user-defined goals. These agents can interact with tions will include agentic AI, up from less than 1% in 2024, and
external systems to take actions in the real world, making them at least 15% of day-to-day work decisions will be made auton-
more capable than traditional generative AI, which primarily omously through agentic AI, up from zero percent.
focuses on content creation. AI agent workflows have at least
three components: understanding the goal and potential Initially, Large Language Models (LLMs) were primarily used to
results (by applying LLMs), breaking the goal down into generate content, mostly in text form. Integrating additional
smaller tasks (planning, memory), and executing these tasks capabilities like planning and tool use has changed the focus
using their knowledge base and the tools available (tool use). from content generation to task completion. AI agents can
AI agents are a recent development with immense potential, now exhibit complex reasoning, remember past interactions,
and one that can be adopted quickly. and adapt their behavior in response to feedback (adaptive AI),
making them far more versatile and capable than their prede-
cessors.

AI agents can create significantly more economic value com-

pared to the initial LLMs because they are designed to autono-
mously perform tasks, make decisions, and interact with their
environment. Completing a task using LLMs is similar to
micro-managing, where each step is discussed and evaluated
by the user. AI agents perform these steps on their own, as
determined by the user, mimicking a competent co-worker.
This autonomy and versatility enables businesses to stream-
line operations, reduce costs, and ultimately create greater
economic value in many areas like complex workflows, code
documentation, and the modernization and automation of
repetitive tasks.

The advent of agentic AI will also create new risks, such as

targeted and advanced cyberattacks that give rise to “smart
malware”. This will require innovations to address the unique
risks and threats posed by systems that depend on LLMs and
GenAI.
Munich Re Page 30/111
Tech Trend Radar 2025

Emerging tech products utilizing

AI agents
In the tech ecosystem, GenAI model providers, cloud vendors,
and the open-source ecosystem are heavily investing in devel-
opments concerning AI agents. Platforms like Microsoft’s
Copilot, Amazon’s Bedrock and Google’s Vertex AI will support
the scalable implementation of these solutions, while open-
source frameworks like Autogen, LangGraph and CrewAI will
drive innovation and adaptability. More specifically, companies
are now also starting to offer embedded services in the finance
sector. The payment provider Stripe is one example, offering a
software development kit (SDK) for embedding Stripe services
in agentic systems.

Also in the insurance ecosystem, various technology providers AI agents at Munich Re and primary
are advancing the efficiency gains made possible by the imple-
mentation of AI agents, particularly in areas like customer insurers
support, automated underwriting, and claims processing. AI At Munich Re, we leverage AI agents to enhance our portfolio
agents have also sparked renewed interest in robotic process for clients. In our automated underwriting platform, REALYTIX
automation by replacing rigid, predefined rules in processes ZERO, AI agents power a Copilot feature that accelerates the
with a more adaptable approach that employs LLMs to plan design and adoption of new, digital insurance products. This
and automate workflows. In the field of agentic process auto- enables our clients to significantly reduce the time required to
mation, the first technology providers have now entered the bring new insurance products from concept to launch.
market.
Primary insurers are also starting to recognize the enormous
potential of GenAI for process automation in high-volume/
high-frequency operations scenarios. However, given the high
risk of errors or unintended outcomes, and the strict regula-
tions regarding AI, companies are currently shying away from
letting AI have full agency over how to approach and execute a
given process, especially if it involves transactional resources.

In contrast, agentic Retrieval-Augmented Generation (RAG)

has been identified as a low-risk application of AI agents, pro-
viding higher efficiency and superior accuracy than traditional
RAG, especially across large corpora of knowledge resources.
It is a complex and instructive, yet contained example of an
arbitrary orchestration of layered LLM resources autono-
mously supervised by a leading LLM acting as the agent.

Another area of conservative experimentation with agentic

automation approaches involves supporting complex scenar-
ios in underwriting workbenches. Non-transactional tasks
(e.g. data collection or scenario calculations) can even be pro-
actively and autonomously executed without explicit approval.
This human/machine collaboration will significantly enrich
underwriters’ working environment.

The future isn’t about having

more AI agents, but about making
them do more.
Andreas Schumacher
Project Manager, Artificial Intelligence
Munich Re Page 31/111
Tech Trend Radar 2025

“AI agents are quickly emerging as

a path to the large-scale adoption
of LLMs in the insurance industry.”

Matthias Beuerle-Liegel
Data Analytics Spezialist

Outlook
AI agents will redefine the insurance industry, driving effi- Despite these advances, challenges remain. Ethical considera-
ciency, accuracy and personalization. This technology will tions, such as ensuring fairness in risk assessments and com-
transform underwriting, claims management, and customer pliance with regulations, are critical. Accordingly, investments
engagement. In underwriting, AI agents can autonomously in AI for risk management practices and policies, as well as
analyze diverse datasets – from customer profiles to market platforms for managing and monitoring agent-based systems,
trends and real-time risk indicators – to deliver accurate risk need to be carefully considered. The codification of relevant
assessments and personalized policies. Advanced multimodal knowledge, strategic tech planning, and human-in-the-loop
models enhance these capabilities, allowing AI systems to control mechanisms should also be considered. To capitalize
process and interpret a variety of data types, including images on AI agents, it is helpful to consider how to best train and
and geospatial information. Claims management may see manage a blended digital and human workforce.
end-to-end automation, with agentic AI-powered systems
being used to validate claims, detect fraud, and coordinate By embracing AI agents, insurers can achieve groundbreaking
document verification. Customer engagement will also be improvements in operational efficiency, customer satisfaction,
transformed, with AI agents anticipating individual needs and and competitive advantage. This technology promises to set a
offering proactive risk management advice. new benchmark for innovation and agility in the insurance
industry.

Agents are key to creating

economic value from GenAI.
Dr. Andreas Nawroth
Leading Expert, AI & Quantum
Munich Re Page 32/111
Tech Trend Radar 2025

Data & AI

AI-Augmented Software Engineering

Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Trial & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Increased productivity through automation of repetitive tasks − Over-reliance on AI, potentially reducing critical thinking skills
− Faster time-to-market for digital insurance products and features − Algorithmic bias, leading to unfair outcomes
− Cost savings due to reduced manual labor − Security concerns, such as inadvertently introduced vulnerabilities
− Enhanced risk modeling using AI-generated algorithms − Data privacy issues, including potential exposure of sensitive
information

By recognizing these challenges and opportunities, organizations

can unlock the full potential of AI-ASE while ensuring robust, secure,
and fair implementation.

Overview Innovation by AI-Augmented

AI-Augmented Software Engineering (AI-ASE) is transform- Software Engineering
ing the software development landscape by boosting effi-
ciency, accuracy, and creativity. According to Gartner, by 2027, AI-Augmented Software Engineering leverages Artificial Intel-
70% of all developers are expected to be using AI-driven cod- ligence to revolutionize development processes. From code
ing tools, up from less than 10% today. This shift will enable generation and debugging to automated testing and CI/CD
sectors like insurance to gain competitive advantages, such as optimization, AI tools are increasingly becoming indispensable.
improved risk modeling, streamlined data processing, and For example, AI can detect potential security vulnerabilities in
faster deployment of digital solutions. However, organizations code and recommend fixes, while advanced code completion
must balance innovation with ethical considerations in order tools like GitHub’s Copilot can intuitively anticipate develop-
to harness AI-ASE effectively. ers’ intentions. Moreover, natural language interfaces like chat-
based tools empower developers to iteratively refine code,
generate tests, and handle traditionally tedious tasks with
ease. By reducing manual effort and minimizing errors, AI ena-
bles companies to deliver better services while optimizing
operational costs.

This growing reliance on AI is evident in platforms like GitHub,

where AI is integrated seamlessly across the ecosystem, and
in Google’s Project IDX, Cursor, and Windsurf, which were
designed with AI support at their core. For sectors like insur-
ance, where technology is essential, AI-Augmented Software
Engineering offers a competitive advantage. Companies can
innovate by streamlining data processing, improving risk
modeling, and accelerating the deployment of digital solutions.
In addition, studies show that incorporating AI into the devel-
opment cycle increases developer satisfaction, helping retain
talent in competitive markets. In countries with unfavorable
demographics, it also reduces the pressure to compete for new
talent.

Keyfact
By 2027, 70% of all professional developers will use
AI-powered coding tools, up from less than 10% today.
Munich Re Page 33/111
Tech Trend Radar 2025

Data & AI

AI Search Engines
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Adopt & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Provide instant, clear answers to customer questions. − Reduced visibility, as customers rely less on traditional search engines.
− Use LLMO to improve visibility and enhance customer engagement. − Difficulties ensuring content is selected and presented accurately
− Deliver personalized product suggestions through optimized conver- by LLMs.
sational content. − Adapting content strategies to meet AI Search Engine requirements.

Overview Beyond keywords: The new era of AI

The emergence of AI Search Engines is transforming how Search Engines
information is researched and accessed, moving away from
traditional keyword-based systems to context-aware, conver- The way we search for information is undergoing a dramatic
sational responses. This shift has profound implications for transformation, driven by the rapid emergence of AI Search
businesses, requiring them to adapt through strategies like Engines. These models, with their ability to process and gener-
Large Language Model Optimization (LLMO) to ensure con- ate human-like text, are reshaping the search experience. For
tent visibility and relevance. These AI Search Engines excel at over two decades, Google has dominated this space, holding a
delivering personalized, precise answers by interpreting user 91% market share in the $50 billion search advertising market
intent and integrating external knowledge sources. To stay (Harvard Business Review, 2024). However, LLMs like Search
competitive, businesses must embrace conversational content, GPT, Perplexity, Gemini, and YOU are driving a shift away from
structured data tailored for this evolving search paradigm. traditional keyword-based search engines toward conversa-
tional, context-aware responses tailored to user queries.

Unlike traditional search engines that rely on keyword match-

ing, AI Search Engines interpret user intent and deliver per-
sonalized, contextually relevant answers in a conversational
format. This evolution eliminates the need for users to scroll
through countless links, making it easier to find precise infor-
mation. By delivering responses adapted to individual prefer-
ences and behaviors, AI-powered search engines enhance
both relevance and user satisfaction.

A pivotal feature of LLM-based systems is Retrieval-Aug-

mented Generation (RAG), which integrates external knowl-
edge sources, such as databases and specialized documents,
into AI responses. While this enhances contextual relevance
and mitigates certain limitations of model-training data, the
quality and trustworthiness of the retrieved information
directly impact the reliability of answers.
Munich Re Page 34/111
Tech Trend Radar 2025

LLMO: Mastering LLM visibility in the

Key strategies for LLMO
new search landscape
− Aligning content with Natural Language Queries by
This shift has also given rise to a new field: LLM Optimization using conversational tone and incorporating content
(LLMO). Just as SEO revolutionized visibility in traditional − Using well-structured and hierarchical (meta-) data
search engines, LLMO is becoming essential in the LLM- to aid retrieval
driven landscape. Businesses must strategically adapt to this − Ensuring information is accurate and frequently
paradigm by ensuring their content aligns with how LLMs updated
retrieve and present information. Mastering LLMO will deter-
mine which brands thrive in this new environment.

Disrupting the search giants

The rise of AI Search Engines is challenging the dominance of
link-based search engines like Google. This disruption has
profound implications for the search advertising industry:

− Reduced reliance on links: AI Search Engines generate direct

answers, decreasing website traffic and visibility for busi-
nesses reliant on organic search.
− New advertising paradigms: Once available, companies must
integrate promotional content directly into LLMs as opposed
to relying on traditional search ads.
− Content adaptation challenges: Businesses must restructure
their digital presence, ensuring it’s optimized for conversa-
tional search to remain discoverable.

For businesses, adapting to this shift isn’t optional – it’s essen-

tial for maintaining relevance and visibility in a landscape
Insurance searches: How AI search
where traditional search models may soon become obsolete. personalizes and simplifies the process
Navigating AI Search: How to stay competitive and visible 1. Streamlined insurance discovery
AI search isn’t just a technological shift – it’s a strategic Consumers often begin their insurance research online. LLMs
imperative for businesses. To thrive in this new paradigm, revolutionize this process by offering instant, conversational
it is vital to: answers. For instance, a user asking, “What’s the best health
insurance for a family of four in Germany?” would receive a
− Adapt your content for LLM Optimization (LLMO): Ensure clear, personalized response, eliminating the hassle of inter-
your website content aligns with how LLMs retrieve and preting dense documents or exploring multiple websites.
present information, using natural, conversational language
and well-structured data. 2. Personalized recommendations
− Focus on personalization: Leverage LLM capabilities to LLMs analyze user behavior to deliver highly personalized
meet user needs with tailored, context-aware responses. suggestions. Tailored responses are provided for specific
− Monitor AI Search Advertising options: Stay ahead by scan- questions.
ning the market for new advertising options such as spon-
sored queries and product ads within AI Search Engines and 3. Simplified comparison shopping
continuously refine your content to remain visible and com- Customers often struggle to compare policies across provid-
petitive. ers. LLMs simplify this by aggregating data and presenting
− User-generated content is given more weight in the output. side-by-side comparisons of key factors – premiums, deducti-
Consequently, LLMO also means optimizing content on bles, and benefits – in an easily digestible format.
third-party or social media platforms.
4. Addressing common questions
By embracing AI Search Engines, businesses can adapt to LLMs can anticipate and answer frequently asked questions
the disruption of traditional models and seize opportunities instantly. For example, a query like “What’s the difference
to enhance user experiences and drive innovation. The future between term and whole life insurance?” is met with concise,
of search belongs to those who adapt – and the time to act clear explanations, improving consumer understanding.
is now.
Munich Re Page 35/111
Tech Trend Radar 2025

Data & AI

Legal Tech
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Trial & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Lawyers, firms and law departments: Greater efficiency and accuracy − Professional ethics: Lawyer regulation may conflict with some uses of
drive higher job satisfaction and higher profitability technology. For instance, several lawyers have been disciplined for not
− Consumers: Automated processes improve accuracy of legal advice properly supervising and confirming the accuracy of results when
and lower costs, resulting in greater access to, and satisfaction with, using generative AI for drafting legal documents submitted to the court
legal services
− Loss of skills over time: Over-reliance on technology may lead to lack
of creativity and forward thinking by lawyers

Overview Legal technology is being used

For lawyers, generative AI is increasingly being incorporated increasingly by lawyers and law firms
into legal drafting and research. Legal support technologies, Legal support technologies, including remote depositions
including remote depositions and court appearances, AI-as- and court appearances, AI-assisted transcription and trans-
sisted transcription and translation, e-discovery, data mining, lation, e-discovery, data mining, calendaring, conflict checking
calendaring, conflict checking and matter management are and matter management are evolving rapidly due to innova-
also evolving rapidly due to innovations in AI. For consumers, tions in AI. For consumers, online legal services that provide
online legal services that provide legal advice and document legal advice and document templates make access to justice
templates make access to justice cheaper and more widely cheaper and more widely available. Insurance is being
available. impacted by new legal and regulatory technologies, particu-
larly in those lines that cover third-party and regulatory
sanctions.

Commercial lawyers and law departments

Generative AI is being increasingly incorporated into legal
drafting and research. New technology allows more open-
ended queries and focused results. Contract review and even
contract language generation are also experiencing increased
use and popularity. Generative AI is even employed for plotting
litigation strategies. A survey conducted by US Legal Support
found that, in 2023, “21.25% of respondents used AI for legal
research, document management and predictive analysis. In
2024, that share has risen to 25.92%, an increase of 21.98%
year-over-year (YoY). This was reflected in our results, as pri-
vacy and security concerns are the biggest factors considered
when law firms vet legal technology vendors and litigation
support providers.” Law firms and law departments process
and store a great deal of confidential information: personal-
ly-identifying client information, client communications, confi-
dential agreements, intellectual property and the like. Due to
the nature of the data they hold, these entities are prime tar-
gets for cyberattacks, and the consequences of data breach
incidents are particularly severe for the firms’ reputations and
for their clients.
Munich Re Page 36/111
Tech Trend Radar 2025

Consumers
In the consumer space, “robot lawyers,” online legal services
that may or may not be supervised by lawyers, are becoming
increasingly common. These services provide templates
or automated solutions for everything from divorce, wills and
real estate purchases to setting up a small business and
resolving traffic infractions. In the United States, DoNotPay
recently reached a settlement with the Federal Trade Commis-
sion prohibiting the company from advertising its services
as equivalent to an actual lawyer and ordering a fine of nearly
$ 200,000. There are similar AI-powered online legal services
in India and the UK, and they will undoubtedly spread to other
jurisdictions.

These services have suffered setbacks and regulatory chal- For insurers that adopt new legal and regulatory technology,
lenges. In particular, scaling is difficult due to the many, often their legal departments can become much more efficient,
subtle, differences from one jurisdiction to another, even in making document drafting, processing, matter management
the same country. Due to the legal profession being tightly and legal research less time-consuming and human
regulated, many legal innovators using this model have faced resource-intensive. By leveraging advanced e-discovery and
lawsuits from lawyers’ organizations, as the Hague Institute data mining technologies, the overheads associated with
for Innovation of Law has found. To take just one example, litigation can reduce costs substantially. These efficiencies can
the online legal service Legal Zoom was sued in June 2024 increase overall profitability, while at the same time improving
by New Jersey residents for engaging in the unauthorized job satisfaction by freeing up legal staff to perform more
practice of law. The unauthorized practice of law is illegal in challenging and personally rewarding tasks.
most jurisdictions.

Implications of legal and regulatory tech Conclusion

for insurers In 2025, moving forward, we can expect the trends in legal and
Insurers are transforming their operations by using new legal regulatory technology to continue to build momentum. Com-
and regulatory technology. Just about every aspect of the mercial lawyers will avail themselves of increasingly sophisti-
insurance value chain is impacted by new tools. Insureds oper- cated tools that automate many legal tasks or make them
ating in highly regulated industries that adopt new technology more efficient. In the consumer market, legal services will
become better risks due to increased accuracy and lower continue to proliferate, albeit facing regulatory challenges from
odds of incurring regulatory expenses. Likewise, insureds that individuals, classes and lawyer organizations. For insurers,
employ more advanced legal and regulatory technology may increasing adoption of the latest advancements in legal and
experience fewer claims and, when they do, have lower associ- regulatory technologies will increase efficiencies and accuracy
ated costs. in many areas and improve job satisfaction for legal staff.
Munich Re Page 37/111
Tech Trend Radar 2025
→ Data & AI
→ Healthy Human
→ Connected Experience
→ Cyber & Crypto
→ Redefining Industries

2 Healthy Human
Munich Re Page 38/111
Tech Trend Radar 2025

Healthy Human

Unlocking opportunities
through AI, data analytics,
and personalized care

The healthcare landscape is rapidly AI medicine, for instance, is improving diagnosis and treat-
ment by identifying high-risk patients and personalizing treat-
evolving, driven by advances in Artificial ment plans. Personalized medicine is another key trend, with
Intelligence (AI), personalized medicine, tailored treatments based on individual genetic profiles lead-
ing to better patient outcomes and more effective prevention.
digital healthcare and behavioral
analytics. These trends are transforming Further, digital healthcare platforms are being combined with
behavioral analytics to facilitate seamless coordination
the way healthcare is provided and between healthcare providers and patients. This convergence
redefining how life and health insurers of technologies enables the real-time monitoring of health sta-
tus, facilitates preventative care, and promotes better disease
can integrate these technologies into management.
their business models.
In this chapter we have identified key technology trends that
present significant opportunities for life and health insurers to
enhance underwriting and risk management, drive business
growth, and deliver more tailored products and services.
Munich Re Page 39/111
Tech Trend Radar 2025

Healthy Human

Trend evolution and impact on the value chain

high medium low

Trends Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

& Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Healthy
Human
Personalized
trial
Medicine
AI Medicine
trial

Digital
adopt
Healthcare
Behavioral
adopt
Analytics
Munich Re Page 40/111
Tech Trend Radar 2025

Healthy Human

Personalized Medicine
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Trial & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Improved patient outcomes through tailored treatments − High costs limiting widespread adoption
− Development of new treatments for individual needs − Privacy concerns, as sensitive personal data could potentially lead to
− Early disease detection and prevention discrimination and bias
− Reduced healthcare costs through targeted treatments, reducing the − Lack of standardization and regulation
need for trial-and-error treatments and minimizing the risk of adverse
reactions
− Improved risk assessment for insurers and tailoring of products
through improved prognostication

Overview Potential and challenges

Personalized medicine is a rapidly evolving field that holds Personalized medicine has significant implications for the
great promise for improving patient outcomes and transform- insurance value chain, from risk assessment and underwriting
ing the way healthcare is provided. By leveraging advances in to claims management and policyholder engagement.
genomics, epigenetics, and related technologies, personalized It can be provided through the use of epigenetic analysis and
medicine enables healthcare providers to tailor treatments to other technologies that provide insights into an individual’s
individual patients’ unique needs and characteristics. In the health risks and characteristics. Although it may take time for
insurance industry, personalized medicine has significant personalized medicine to become integrated into mainstream
implications for risk assessment, underwriting, and claims medicine, it is important to consider that when it does, there
management. will be a substantial impact on the insurance value chain.
Insurers should proactively engage with industry stakeholders
and regulatory authorities to address potential challenges and
capitalize on opportunities. By doing so, they can ultimately
improve the risk assessment process and benefit their clients.

These various technologies also have the potential to trans-

form the way claims are managed by enabling insurers to tailor
claims handling to individual policyholders’ unique needs. This
can be achieved through the use of advanced analytics and
machine learning algorithms that analyze personalized bio-
markers against claims data to identify patterns and trends,
indicating where intervention could reduce claims severity or
frequency. Insurers can also leverage the integration of various
rehabilitation programs and interventions which can assist in
reducing second and subsequent claims or facilitate earlier
return to work through earlier and more precise interventions.
Munich Re Page 41/111
Tech Trend Radar 2025

Furthermore, personalized medicine can improve policyholder Precision medicine vs. personalized
engagement by enabling insurers to provide health and well-
ness programs that are tailored to individual policyholders’ medicine
unique characteristics. By leveraging advanced analytics and Precision medicine and personalized medicine are often used
insights, insurers can identify opportunities for engagement interchangeably, but are not the same. Precision medicine
and intervention, and provide more targeted support to policy- refers to the use of advanced technologies, such as genomics
holders. and epigenetics, to develop targeted treatments that are tai-
lored to individual patients’ unique needs and characteristics.
One of the biggest challenges lies in the costs associated with Personalized medicine, on the other hand, refers to the use of
unique therapies. Further challenges to adoption of the tech- these technologies to develop treatment plans that are tailored
nology relate to data protection and acceptable use of this sort to individual patients’ profiles and requirements. This means
of technology in insurance, which may be regulated in some that personalized medicine is a wider field that includes con-
markets. ventional therapies, too.

Key insights and recommendations for action

Insurers should develop a strategy for incorporating data that
may become available from personalized tests and therapies
into their risk assessment and underwriting processes, bearing
in mind that regulation may be lagging in this field.

Technologies enabling personalized medicine:

Data Analysis

Genomics: The study of genes and how they function is Machine Learning: Machine learning algorithms are
a key enabler of personalized medicine. Genomic analy- increasingly being used to analyze claims data and
sis can provide insights into an individual’s health risks identify patterns and trends. This can improve claims
and characteristics, and enable healthcare providers to handling and reduce costs.
tailor treatments accordingly.
Artificial Intelligence: AI is being used to analyze large
datasets and identify patterns and trends in health
Epigenetics: The study of gene expression and its data. This can be used to develop personalized treat-
impact on health is another key enabler of personalized ment plans and improve health outcomes.
medicine. Epigenetic analysis can reveal important
health information and allow for more targeted and
effective treatment strategies. Liquid Biopsy and Cir-
culating Biomarkers: Non-invasive liquid biopsies and
circulating biomarkers facilitate the detection and Therapies
monitoring of diseases.
Pharmacogenomics: The term is used primarily to
Single-Cell Analysis: This technology enables the describe customized pharmacotherapy that takes into
analysis of individual cells, providing insights into cellu- account not only the specific clinical picture, but also
lar heterogeneity and helping to identify specific cell the individual’s physiological composition and gender-
populations that may be driving disease progression. specific effects of medication. Personalized medicine
seeks to improve the stratification and timing of health-
care by utilizing biological information and biomarkers
Imaging and Radiomics: High-resolution imaging and
on the level of molecular disease pathways, genetics,
radiomics support the analysis of medical images to
proteomics and metabolomics.
identify biomarkers and predict treatment outcomes,
facilitating more accurate diagnoses and personalized Synthetic Biology and Gene Editing: Advances in
treatment plans. gene editing technologies like CRISPR/Cas9 are
enabling the development of novel therapeutics and
personalized treatments for genetic diseases.
Electronic Health Records (EHRs) and Health Infor-
matics: The integration of EHRs and health informatics 3D Printing and Bioprinting: These technologies are
is providing a comprehensive view of patient data, being used to create customized implants, prosthetics,
enabling healthcare providers to make more informed and tissue models, enabling the development of per-
decisions and develop personalized treatment plans. sonalized treatments and therapies.
Munich Re Page 42/111
Tech Trend Radar 2025

Early cancer detection is crucial

Cancer is a leading cause of death among individuals under 65
in many insurance markets, making it a pressing concern for
leading insurers and Munich Re. As a result, leveraging per-
sonalized medicine technologies to enhance cancer prognosis
is a key focus area. From a clinical perspective, early cancer
detection is crucial for improving treatment outcomes. How-
ever, universal cancer screening is currently not feasible due to
cost constraints and the potential for false positives. Develop-
ments in this field have also demonstrated to researchers that
the focus should be on the molecular profiles of specific types
of cancer rather than the organ of origin. This will assist in bet-
ter understanding the underlying biology of the disease in
order to develop more effective treatment strategies. For more
information about advances in cancer outcomes, please refer
to our Life Science Report.

Recent advances in liquid biopsy technology have enabled the Munich Re North America is currently collaborating with 11
analysis of tumor DNA, RNA, and other biomarkers in blood, carriers and two distribution partners to offer the Galleri® test
urine, and saliva. The Galleri® test, developed by GRAIL, is a to eligible policyholders on a post-issue basis. As of November
multi-cancer screening tool that can detect around 50 types of 2024, more than 7,000 tests had been completed by policy-
cancer. This test has demonstrated higher sensitivity rates for holders on a voluntary basis. Notably, 73% of positive results, at
more severe cancers and specificity rates of 99.5%, delivering an average patient age of 60, were for cancers with no current
highly accurate identification of cancer markers. screening available, highlighting the potential of this technol-
ogy to improve cancer detection and treatment outcomes. We
are likely to see a short-term impact on targeted and effective
treatment, which may lead to lower mortality claims in the
long term. For more information, see here.

Key insights and recommendations for action

Insurers should start investing in advanced analytics and
machine learning algorithms that can analyze claims data and
identify patterns and trends by combining the claims data with
data linked to personalized medicine technologies.
Munich Re Page 43/111
Tech Trend Radar 2025

Healthy Human

AI Medicine

Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Trial & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− High-precision patient data analysis can be used to recommend the − Risk of distorted results and inadvertent discrimination due to uncon-
most appropriate treatment options based on individual characteristics scious bias in AI system data
− Earlier detection, earlier intervention, and better triage, leading to − Errors and omissions in AI-driven diagnoses and treatment may raise
improved health outcomes and reduced healthcare utilization concerns about accountability and liability
− Automated tasks and data analysis can improve time management of − Over-reliance on AI systems can lead to reduced clinical skills and
HC providers critical thinking skills
− Risk of data privacy breaches and system errors, which can compro-
− NLP can improve outcomes for at-risks individuals
mise patient confidentiality
− More efficient underwriting by applying predictive analytics
− Risk of anti-selection risk if insured individuals with rarer conditions
can use this information for their benefit

Overview The impact of AI Medicine

Artificial Intelligence is transforming the healthcare landscape AI medicine is a field that uses Machine Learning and Artificial
at an unprecedented pace, with significant potential to impact Intelligence techniques to analyze medical data and make rec-
the insurance value chain. For insurers, healthcare profession- ommendations or predictions about patient care or drug devel-
als, and patients alike, AI holds the promise of revolutionizing opment. This trend is related to personalized medicine, which
every single aspect of care – from earlier diagnoses to person- focuses on tailoring medical treatment to an individual’s
alized treatment options. By utilizing machine learning and unique characteristics, such as their genetic profile, medical
Artificial Intelligence techniques to analyze medical data, history, and lifestyle.
AI medicine can improve diagnostic accuracy, reduce the risk
of missed or incorrect diagnoses, reduce adverse effects in One area where AI medicine is having a significant impact is in
clinical trials, and enhance clinical outcomes. In the insurance the development of new drugs. AI can be used to analyze large
industry, AI medicine can lead to improved mortality and amounts of data, including genomic data, data on chemical
morbidity, and to increased efficiency in data analysis and compounds, and clinical trial results, to identify potential drug
decision-making. To fully capitalize on these benefits, insurers candidates and predict their effectiveness. This subfield of
need to start leveraging AI medicine insights, while also AI medicine is referred to as Pharma AI.
addressing potential data-asymmetry risks.
Another area where AI medicine is being used is in the analysis
of medical images, such as retinal scans, X-rays and magnetic
resonance imaging. This can result in faster diagnoses, more
efficient use of human medical resources, and the generation
of large data volumes. AI can also be used to identify abnor-
malities and diseases, and to recommend the most appropri-
ate treatment options. In addition, AI algorithms can be used
to detect abnormalities not visible to the human eye, including
rare conditions.

Key insights and recommendations for action

Invest in applying AI-powered solutions in insurance
processes: Insurance companies will benefit from using
affordable AI-powered technologies to reduce the costs
of underwriting and claims adjudication.
Munich Re Page 44/111
Tech Trend Radar 2025

Figure: The application fields of AI Medicine

AI-based
drug discovery
or Pharma AI

Diagnostic AI-assisted
AI
imaging and diagnosis and
analysis Medicine prognosis

Predictive
analytics in
healthcare

Source: Munich Re

In health insurance, detailed healthcare data has been used for The application of healthcare data to predict healthcare utili-
some time to predict healthcare utilization through predictive zation has a strong parallel in how healthcare practitioners are
analytics. Insurers collect data from claims, electronic health leveraging AI to diagnose patients earlier. While the stakes
records, pharmacies, lab results, and demographic information. are higher in healthcare, the insurance industry can also learn
However, since the “explosion” of AI technology, advanced sta- from this trend. As AI medicine continues to evolve, insurance
tistical models and Machine Learning algorithms have been companies can tap into its expanding capabilities to enhance
deployed to analyze this data to identify patterns and trends, their value chain.
which can help to predict an individual’s likelihood of utilizing
healthcare services. One area where this convergence is especially evident is in
AI-augmented life insurance underwriting. Traditionally,
Predictive factors include comorbidities, disease severity, underwriters relied on health disclosures to assess the risk
treatment adherence, lifestyle factors, and socioeconomic fac- profiles of policyholders, determining standard rates or loading
tors. The resulting risk score informs premium pricing, care for additional mortality or morbidity risks. By combining health
management, and resource allocation. By leveraging predictive data with other data points and applying predictive analytics,
analytics, insurers can more effectively manage risks, reduce the underwriting process has become significantly faster and
costs, and improve health outcomes for their policyholders. more efficient.

“AI in medicine: A paradigm

shift in preventive medicine,
diagnostics and treatment.
This convergence of technol-
ogy and medicine promises
even more significant
advances in the years to
come.”
Dr. Pamela Chetty
Chief Medical Officer, Munich Re
Africa Branch
 Munich Re Page 45/111
Tech Trend Radar 2025

Munich Re is a pioneer in this field, having deployed AI-pow-

ered underwriting solutions in e.g. Asia, the Middle East, and
Africa. These solutions have been successfully implemented
and generating value for years, yielding tangible improvements
in customer experience. The life and health insurance industry
is now at a turning point in terms of monetizing AI, as
explained in detail in our white paper on Next Generation
AI-Augmented Underwriting.

Looking ahead, we can expect to see the customer journey

transformed through earlier and more comprehensive data
collection based on customer risk profiles. This will reduce
information asymmetry, minimize the need for lengthy under-
writing questions, and enable better risk classification. Munich
Re’s AI-augmented underwriting solution, which integrates
multiple data sources, AI models and technologies, is a prime
example of this transformation.

“AI retinal screening is becoming standard

practice in China, and it’s making the screen-
ing of eye diseases very affordable and acces-
sible. It’s not a routine requirement in insurance
medical exams yet, but we have been seeing
more AI retinal exams in our life insurance
applications.”
Dr. Hao Liu
Proposition Medical Doctor at Munich Re
South-East Asia, Middle East and Africa markets

Figure: Munich Re’s AI-augmented underwriting solution

uses multiple data sources, AI models and technologies to
transform the underwriting journey.

Multiple AI model predictions

National health, eHR data
Multiple risk scores
Wellness / activity data
Data collected on customer

Facial image, video

Back office data Medical
Application data evidence
ey
ng journ
n derwriti
Future u
ey
urn
g jo
it in
e rw r
und
t ing The future customer journey:
E xis
Medical reduced information asymmetry
evidence through earlier and wider data
collection based on customer risk
Back office data
profiles – fewer UW questions
Application data
required and better risk classification

Time since point of sale

Munich Re Page 46/111
Tech Trend Radar 2025

Healthy Human

Digital Healthcare
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Adopt & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Increased quality, quantity and types of data available on customers to − Regulatory limitations on using personal health data
be insured, which can drive improved customer propositions − Data security and safeguarding concerns
− Improved data analysis and decision-making in life & health insurance − Proving clinical acceptance of digital tools, which drives adoption in
− Simplified coordination between healthcare providers and parties in insurance
the insurance value chain
− Automated and improved processes such as underwriting

Overview Categorizing Digital Healthcare

The healthcare industry is undergoing a significant transfor- technologies
mation with the integration of digital technologies, which
are revolutionizing the way patients receive care and interact Digital Healthcare is a multifaceted trend that encompasses
with healthcare professionals. This trend, known as Digital various components, which impact the insurance value chain
Healthcare, is not only improving patient outcomes but also to varying degrees. Dividing the trend into two categories can
generating vast amounts of data that can be leveraged to facilitate a better understanding of its current and future
enhance the insurance value chain. From electronic health impact on insurance.
records (EHRs) to telehealth and remote monitoring, Digital
Healthcare solutions are streamlining coordination between Healthcare data management refers to digital solutions for
healthcare providers, insurance companies, and patients. collecting, storing, and analyzing healthcare data to improve
patient outcomes and operational efficiency. Examples from
this category include electronic health records (EHRs) and
interoperability technologies such as real-time health system
(RTHS) platforms. These technological fields have a signifi-
cant impact on the insurance industry, as they enable secure
data sharing and provide valuable insights for healthcare
providers, insurance companies, and researchers alike.

Post-issue
review
Light-touch
underwriting

Automated
underwriting
Munich Re Page 47/111
Tech Trend Radar 2025

Electronic data
Data
management
Interoperability The transformation of underwriting via technology
tech will accelerate this year. In the United States, elec-
tronic health records (EHRs) have reached critical
data flow
NLP thresholds of coverage and speed of return, mak-
Digital
Healthcare
ing them essential for life insurance underwriting.
The full spectrum of AI tools – machine learning,
AI health
Digital health coaching Natural Language Processing and Gen AI – will
technology enable the industry to move beyond manual
Digital review of records to full automation of decisions.
therapeutics
Patrick Sullivan
SVP, Integrated Analytics, Munich Re Life North America
Ambient Digital
Health scribes

Smart devices

Smarter devices

Telehealth

Source: Munich Re

In turn, digital health technology encompasses a wide range of The Automated EHR Summarizer extracts relevant informa-
digital solutions that enable healthcare providers and patients tion from EHRs to provide a profile of an individual’s health
to play a more active role in healthcare in various stages of the history including physical measurements, medical diagnoses,
value chain. This category includes many outstanding exam- procedures, treatments, lab results, prescriptions, and tobacco
ples, such as Natural Language Processing (NLP), ambient use, which are essential factors for assessing risks.
digital scribes, smart devices, smarter devices, telehealth,
remote photoplethysmography tools, and AI health coaching. We apply a range of AI solutions, Natural Language Process-
ing (NLP) and machine learning algorithms, paired with our
While contributing to the insurance value chain to varying risk assessment expertise, to process complex and messy data
degrees, they all add to the growing pool of health data which quickly and accurately. Further, the EHR Summarizer supports
may at some stage prove to be useful in the chain, as they are automated underwriting by providing streamlined data for
all aimed at improving health outcomes. When considering digital consumption in rules engines or scoring models, and
using any of these technologies, insurers need to be aware of improves the speed and efficiency of manual reviews in the
evolving regulations, such as getting consent to share smart light-touch underwriting and post-issue review contexts.
device information under standard personal information pro-
tection legislation, which can vary considerably from region to Munich Re’s Automated EHR Summarizer is just one example
region and pose challenges to progress. of how Digital Healthcare solutions can be used to improve the
efficiency and accuracy of insurance processes, while also
The impact of Digital Healthcare is significant, and it is enhancing the insurance applicant experience by minimizing
increasingly being adopted in the insurance value chain. To friction and reducing the time to decision.
show how the myriad of technologies can work together, it’s
worth highlighting an example from the United States that
combines electronic health records (EHRs) and NLP techno-
logies.

Munich Re has made it easier to use electronic health records Key insights and recommendations for action
for risk assessment with its Automated EHR Summarizer, a Invest in Digital Healthcare solutions: Insurance companies
tool that analyzes and summarizes large volumes of patient should invest in Digital Healthcare solutions that enable the
data and is designed to improve the efficiency and accuracy of collection, storage, and analysis of healthcare data to improve
underwriting. key processes in the insurance value chain like underwriting.
Munich Re Page 48/111
Tech Trend Radar 2025

Healthy Human

Behavioral Analytics
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Adopt & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Offers a more accurate understanding of individuals’ health and life- − Can perpetuate existing biases if the data or algorithm design is
style, enabling enhanced risk assessment and more informed deci- flawed, leading to unfair treatment of certain customer groups.
sion-making. − Compliance with relevant laws and regulations is important, so as to
− Facilitates customer segmentation, allowing insurers to craft personal- protect customer privacy and personal data.
ized products and services that meet individual needs, driving − Incorrect or falsified behavioral data can lead to incorrect risk or pricing
increased sales and reduced acquisition costs. assumptions, highlighting the need for robust data validation and
− Helps identify undisclosed health risks, mitigating the likelihood of verification processes.
fraudulent claims. Reveals new opportunities for customer engage- − Clients may not accept the use of certain variables, such as geo-loca-
ment, enabling targeted marketing campaigns and improved customer tion tags.
retention. − Justifying specific variables and outcomes can be difficult.

Overview Applications of behavioral analytics in

The insurance industry is on the cusp of significant change, the insurance value chain
driven by the increasing availability and sophistication of The integration of behavioral analytics across the insurance
behavioral analytics. This technology enables the collection, value chain is transforming the way insurers operate, from
monitoring, and analysis of data on individual actions and product design to claims processing. By leveraging behavioral
intentions from digital products, providing unprecedented data, insurers can create personalized sales and upselling
insights into human behavior. As the amount of data generated opportunities that meet their customers’ unique needs. For
across multiple data sources continues to grow, the relevance example, data-driven insights can be used to tailor complex
of behavioral analytics in the insurance value chain increases life insurance products to individual customers’ needs, high-
accordingly. From risk assessment and customer segmenta- lighting relevant features and benefits at the most relevant
tion to fraud prevention and personalized product offerings, time.
the potential applications of behavioral analytics in insurance
are vast and varied. This personalized approach can foster trust and confidence in
the product, increasing the likelihood that customers will feel it
meets their changing needs. Furthermore, behavioral analytics
can enhance underwriting processes, enabling insurers to
make more informed decisions about risk assessment and
pricing.

In the area of risk management, behavioral data can help

detect patterns and anomalies and detect instances of poten-
tial fraud, thereby improving claims processing efficiency. One
example is analyzing clients’ digital footprints as they com-
plete online risk questionnaires. Additionally, behavioral ana-
lytics can facilitate more effective customer segmentation,
allowing insurers to develop targeted marketing campaigns to
improve customer retention.

The health and insurance sector has witnessed significant

advances in risk assessment, driven by the increasing adoption
Key insights and recommendations for action of behavioral analytics. Munich Re is leveraging behavioral
Invest in customer segmentation: Ensure that behavioral data analytics to strike a balance between risk assessment and cus-
is being collected to identify which customers are more likely tomer experience. This approach has yielded valuable insights,
to take up a product based on their needs, thereby reducing shedding light on the potential and limitations of behavioral
distribution costs. data in risk assessment.
Munich Re Page 49/111
Tech Trend Radar 2025

One of the primary challenges in using behavioral data for risk

assessment is the lack of transparency and explainability.
Behavioral scores can identify high-risk individuals, but often
lack concrete evidence, making it difficult to justify underwrit-
ing decisions to clients or brokers. This issue is further compli-
cated by potential model biases. This has been observed in
many markets that Munich Re operates in, e.g. in China, where
large datasets are available.

As a result, there is a growing trend towards leveraging medi-

cal data at the point of sale for risk screening, as it provides
more tangible evidence and is more widely accepted by the
insurance industry.

Further, data protection and privacy concerns in many coun-

tries have restricted the use of certain data points, such as
phone numbers and geo-location tags for risk assessment. As
a result, we have had to adapt and focus on more targeted and
market-acceptable variables.

A successful example of this approach can be seen in Munich The DRC’s behavioral analytics capabilities offer a range of
Re’s development of the Dynamic Risk Calculator (DRC) model benefits, including policyholder behavior analysis and
in India. By combining applicants’ profile information and informed decision-making. By applying clustering algorithms,
health declaration data, the DRC model provides a more anomaly detection techniques, and classification algorithms,
cost-effective and market-acceptable alternative to traditional insurers can identify higher-risk client profiles and take proac-
behavioral scores. This model classifies policyholders into five tive measures to mitigate them by applying different onboard-
risk categories, enabling insurers to tailor their underwriting ing requirements. For instance, the DRC can predict the likeli-
processes and onboarding requirements to each category. hood of policyholders filing early claims or lapsing their
policies, enabling insurers to create dynamic underwriting
rules and onboarding requirements.

The success of the DRC model highlights the potential of

behavioral analytics in terms of enhancing risk assessment in
the life and health insurance sector. By leveraging targeted
and market-acceptable variables, insurers can optimize their
risk assessment processes, improve their overall profitability,
and provide a more seamless customer experience. As the
insurance industry continues to evolve, behavioral analytics
will likely play an increasingly important role in shaping the
future of risk assessment.

Leveraging our joint data resources and analytical expertise, we

will continue to expand our data inventories and refine the model
algorithms. We look to apply what we have learned through our
behavioral research and analysis to practical use for the benefit
of our clients and insurance consumers in the China market.

Eric Zhao
General Manager Life & Health, Munich Re China

Key insights and recommendations for action

Develop fraud prevention strategies: Fraudulent claims can be
reduced by identifying high-risk clients and focusing the avail-
able budget on them.
Munich Re Page 50/111
Tech Trend Radar 2025
→ Data & AI
→ Healthy Human
→ Connected Experience
→ Cyber & Crypto
→ Redefining Industries

3 Connected Experience
Munich Re Page 51/111
Tech Trend Radar 2025

Connected Experience

Exploring the convergence

of immersive technologies, AI,
and infrastructure innovations

The field of connected experience In 2025, many trends are expected to continue to mature.
Extended Reality for instance will remain a significant trend,
encompasses trends associated with blending virtual and real environments to create more immer-
or facilitated by networking and infra- sive experiences. The number of insurance use cases in virtual
spaces will grow, although it will still take time until e.g. the
structure technologies. This concept Metaverse becomes a part of everyday life.
refers to the interactions between
AI of course has and will continue to have a major impact on
various systems, devices, and entities, shaping connected experience technologies. Trends such
as well as those between users or as “Edge AI” and “Autonomous Interactions” highlight the
possibilities that Artificial Intelligence offers. The new trend,
between users and machines, all of “Autonomous Interactions” focuses on the capabilities and
which are often enhanced by immersive uses of “AI agents”, which can e.g. autonomously complete
decision-making or problem-solving tasks on behalf of a user
technologies. or another system.

As users, we are increasingly accustomed to continuous con-

nectivity, interacting with digital devices and consuming digi-
tal services. In Ambient Computing, interactions are seamless,
without the user even noticing them. In this regard, e.g. IoT
infrastructure plays a crucial role in reducing data processing
costs and increasing energy efficiency. Newly arising stand-
ards such as 6G are set to provide even better connectivity
around the globe, further boosting speed and reliability.
Munich Re Page 52/111
Tech Trend Radar 2025

Connected Experience

Trend evolution and impact on the value chain

high medium low

Trends Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

& Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Connected
Experience

Metaverse hold

Ubiquitous
assess
Connectivity
Ambient
trial
Computing
Extended
trial
Reality

Edge AI trial

Autonomous
trial
Interactions
Munich Re Page 53/111
Tech Trend Radar 2025

Connected Experience

Metaverse
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Hold & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− New products: Insurance for virtual assets, identities, and cyber risks − (Lack of) user-friendliness: Easy-to-use, affordable, and accessible
− Customer engagement: Immersive consultation and claims handling hardware is still in development
− Recruiting & onboarding: Virtual hiring, training, and onboarding − Interoperability: Lack of standards to ensure smooth transitions
between platforms
− Infrastructure: Reliance on 5G and cloud tech for speed and power may
pose challenges
− Security: Robust data protection and cybersecurity are essential

Overview The Metaverse from the perspective

The term “Metaverse” has gained considerable attention in of the insurance industry
recent years. From tech giants to startups and media compa- The Metaverse refers to an expansive digital space created
nies, billions have been invested to make the vision of an through the convergence of Augmented Reality (AR), Virtual
immersive digital space a reality. And it’s no passing fad; Reality (VR), Blockchain, and Artificial Intelligence (AI). In this
rather, it’s a clear direction in which technology and society will space, the boundaries between physical and digital realities
move in the coming years. The Metaverse has the potential to blur. It enables users to engage in interactive and immersive
fundamentally change the way we live, work, and learn. environments as avatars, whether for social encounters,
business collaborations, or creative projects. These inter-
operable platforms aim to facilitate interaction across various
digital ecosystems and provide a seamless user experience.
In recent years, companies and organizations have begun
exploring the versatile possibilities of the Metaverse for e.g.
medical training, social interactions, shopping, and in the
manufacturing industry.

But a number of challenges still need to be addressed, e.g.

user-friendliness and interoperability. Easy-to-use, affordable,
and accessible hardware is still in development. Also, the lack
of standards can hinder seamless transitions between differ-
ent platforms, making it difficult for users to navigate the
Metaverse efficiently. Infrastructure also poses challenges, as
the reliance on 5G and cloud technology for speed and power
can differ and create disparities in user experiences depending
on their geographic location.
Also, security and data protection are paramount to offer users
a trusted environment.

Although the Metaverse is still under development, its poten-

tial is undeniable. Companies, governments and institutions
are already investing heavily in this future technology, and we
are on the cusp of a new era where digital and physical realities
will merge, creating countless opportunities for innovation and
growth. In the insurance industry in particular, it holds enor-
mous potential.
Munich Re Page 54/111
Tech Trend Radar 2025

1. New insurance models 3. Prevention and claims management

The Metaverse opens up numerous new opportunities for the Further, the Metaverse offers insurers significant opportunities
insurance industry, both in terms of product development and with regard to prevention and claims management. With the
customer interaction. Digital assets like NFTs, virtual real help of virtual reality, for example, training courses can be held
estate, and in-game goods are gaining in importance, increas- to optimize behavior in risk situations. For companies, this
ing the demand for specialized insurance for digital risks. could mean preparing their employees for workplace accidents
Insurance companies can offer policies that protect virtual or cyber-attacks through immersive training. Augmented
goods against hacking, data loss or manipulation. However, Reality technologies could also help loss assessors to evaluate
traditional underwriting models must first be adapted to the physical damage more efficiently by providing additional data
specific risks of virtual environments. through virtual tools.

2. Customer experience and sales 4. AI, Blockchain, and data analysis

The possibility of creating immersive and interactive customer The technological infrastructure of the Metaverse is domi-
experiences is particularly lucrative for insurance companies, nated by AI and Blockchain. Insurers can use AI to assess risks
allowing them to set up virtual branches or consulting areas in real-time by analyzing data from the virtual environment,
where customers can receive personalized advice from ava- while Blockchain can be used to secure transactions and proof
tars. This could not only increase efficiency, but also appeal to of ownership in the metaverse, making the insurance process
younger, tech-savvy target groups. For example, insurers could more efficient and transparent. For example, insurance con-
use simulations in the Metaverse to visualize risks, such as the tracts can be integrated directly into smart contracts that
effects of natural disasters, to better explain their policies. automatically settle claims as soon as certain conditions are
met.

The Metaverse offers us as an insurer new

possibilities for customer communication.
Matthias Nawrocki
Head of Metaverse, ERGO Group AG
Munich Re Page 55/111
Tech Trend Radar 2025

Connected Experience

Ubiquitous Connectivity
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Assess & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Enhanced service delivery: Improved connectivity empowers insurance − Security vulnerabilities: The growing use of connected devices
companies to deliver innovative and enhanced user experiences increases the risk of cyber-attacks, data breaches and disruptions,
through smarter, interconnected systems and AI-powered platforms posing significant challenges for individuals, businesses and govern-
− Streamlined operations: Insurance organizations can optimize risk ments alike
modelling, decision-making and processes by leveraging real-time − Privacy concerns: With more extensive data collection, concerns about
data exchanges across networks of interconnected devices protecting personal and corporate information throughout insurance
processes have become even more important, requiring advanced
cybersecurity measures

Overview Our ability to go online from almost

The ability to stay connected online from virtually anywhere anywhere is crucial for many tasks
has become an essential aspect of modern life. Ubiquitous
Connectivity ensures that devices can continuously generate, The concept of Ubiquitous Connectivity integrates devices
share and process data without interruptions, even under seamlessly across diverse locations and service providers.
adverse conditions. This approach not only accelerates the adoption of advanced
technologies in underserved and remote regions – leveraging
Ultimately, Ubiquitous Connectivity is transforming the way advances like Low Earth Orbit (LEO) satellites – but also
societies work, while also promoting new ways of living, work- incorporates capabilities for edge computing and hybrid cloud
ing and engaging on both the technological and social level. systems. Ubiquitous Connectivity enables real-time data
processing like instant insurance claim validation and analyt-
ics closer to the source, which can reduce latency, improve
reliability, and enhance overall system efficiency. Businesses
can leverage this constant data flow to optimize operations,
enhance user experiences, and deliver smarter services.
Thanks to connected device infrastructures, insurers can mon-
itor homes and assets proactively and offer alerting services
that prevent losses before they occur.

Broadband technologies such as fiber-optic One of the perennial challenges for IoT networks concerns lim-
itations in wireless data transfer speeds and network coverage
and 5G, which help ‘future-proof’ networks,
gaps. However, recent advances, including the rollout of 5G,
continue to grow rapidly to meet the increasing Wi-Fi 6E, Low-Power Wide-Area Networks (LPWAN), and
demand for high-quality, affordable and ubiqui- next-generation satellite networks, are addressing these con-
tous connectivity (…). straints. These technologies, combined with developments in
edge computing, AI and advanced IoT sensors, are making
Source: OECD, 2024 solutions like autonomous systems, wearable devices, and
smart homes increasingly viable and efficient.

Keyfact
A new world record in wireless transmission has been set by
transferring data at a speed of 938 Gigabits per second.
Source: University College London, 2024
Munich Re Page 56/111
Tech Trend Radar 2025

Connected Experience

Ambient Computing
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Trial & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Enhanced risk management through improved access to wider- − Lack of standardization limits device interoperability and has slowed
ranging real-time data points, e.g. real-time underwriting models adoption
− Smart resource management and predictive maintenance help − Widespread data collection raises ethical and legal questions
conserve resources and reduce the costs of losses about user privacy and data ownership, also in insurance use cases
− Operational efficiency: IoT-powered ambient computing minimizes − Connectivity challenges and limited bandwidth can disrupt real-time
the need for manual inspections data exchanges

Overview The benefits of computers and the

Expanding on the concepts of smart homes and the Internet of internet without consciously using them
Things (IoT), Ambient Computing defines a world where com-
puters and the internet seamlessly interact in our daily lives – The use of smart devices in homes, businesses, and public
without the need to consciously use them. To make this vision spaces has surged over the past decade, but when it comes to
a reality, a strong network of smart, connected devices is called seamlessly working together in our daily lives, the reality has
for. Sensors and devices must be ubiquitous, interoperable, often fallen short. Though individual smart technologies are
and capable of processing large and complex data feeds. very good at specific tasks, creating a seamless and user-
friendly experience has proven extremely difficult.

Ambient computing seeks to bridge the gap. This approach

uses technologies that can predict needs in advance – when a
person’s gaze, movements, and behaviors are enough to pre-
dict their needs and deliver highly personalized services with-
out any conscious input on their part.

The ultimate goal is to create an environment where technol-

ogy remains in the background but still operates correctly and
intelligently to enhance an individual’s experience.

With regard to e.g. Usage-Based Insurance (UBI), Ambient

Computing data can update highly personalized insurance
As enabling technologies mature, Ambient products like pay-as-you-go home, health, or property insur-
ance on the basis of the client’s actual usage and risk expo-
Computing has the potential to revolutionize our
sure.
interactions with the digital world – bringing us
closer to a future where technology effortlessly Recent advances give hope that this ambient future will be
aligns with human behavior. here soon, driven by innovations in Artificial Intelligence (AI),
specifically Natural Language Processing (NLP), Large Lan-
Norman Thoms guage Models (LLMs), and communication standards like 5G
VP IoT Technologies, HSB and Wi-Fi 6E.
Munich Re Page 57/111
Tech Trend Radar 2025

Connected Experience

Extended Reality (XR)

Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Trial & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Extended Reality (XR) enables virtual risk demonstrations for custom- − Malicious applications could result in data leakage, such as providing
ers, explaining complex insurance coverages unintentional insights into a user’s communication, field of view or
− AR-enabled devices can provide data overlays in asset inspections location
and claims surveys, e.g. risk categories − XR devices may create a public perception of surveillance, leading to
− XR can be used for remote recruitment processes and immersive reputational risk for early adopters
training, e.g. for sales agents

Overview Extended Reality is reaching

Extended Reality (XR) encompasses Virtual Reality (VR) – the mainstream
computer-generated artificial 3D environments, Augmented Meta was the surprise winner in the market for Extended Real-
Reality (AR) – the real environment enhanced with digital ity devices in 2024. While Apple’s Vision Pro delivered the
overlays, and Mixed Reality (MR) – where the virtual and real highest quality at a premium price point in a somewhat clunky
environment are merged. Currently used in niche use cases, package, Meta produced a stylish and more affordable product
XR is gaining wider adoption as end-user devices become in cooperation with popular glasses brand Ray-Ban. These
increasingly available and affordable. market dynamics underline that the technology is mature and
ready for various private and commercial use cases.

XR will be combined with increasing AI capabilities in recog-

nizing and classifying real-world objects, providing even richer
overlays in AR and more realistic simulations in VR.

Virtual customer interactions in the sales process can become

immersive, e.g. enabling insurance agents to give tangible
demonstrations of risks. In the context of global workforces,
the next generation of virtual meetings in XR will be more
engaging, allowing a wider range of expression than flat video
streams. The Meta glasses already show how XR may find its
way into daily life. With increased adoption, the social awk-
wardness of constant video recording may subside.

Keyfact
More than 1 million Meta Ray-Ban glasses were sold in 2024.
Source: Counterpoint Research
Munich Re Page 58/111
Tech Trend Radar 2025

Keyfact
The B2C market revenue of AR/VR hardware is projected
to grow to US$ 62bn by 2029.
Source: Statista

AR technology is starting to be used in underwriting processes sensors can endanger privacy and confidentiality. It may
in insurance. Digital overlays on smartphone camera feeds are extend insurers’ exposure to fraud risks, as social profiling,
being used to guide SME customers to provide details of their impersonation and fishing become more accurate and
property. believable with more data points.

Underwriters may soon get digital overlays during inspections, Socially, VR/MR poses a risk of isolation if virtual alternative
in the same way as claims adjusters during surveys, to speed realities allow an all-too-easy escape from real-world
up the risk categorization and impact assessment processes. challenges. See Tech Trend Radar 2024 for more details on
Ever-increasing data collection through cameras and other Extended Reality (XR).
Munich Re Page 59/111
Tech Trend Radar 2025

Connected Experience

Edge AI
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Trial & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Fast and intelligent decision-making that supports, for example, − Local computing power is more limited than in a cloud approach,
claims processes or preventive measures so deployable AI models on the edge are usually less complex
− Reduced costs due to less data traffic − Maintenance and updates might be challenging on spread edge
− Local processing enables a better response to local security require- devices
ments or regulations

Overview Insurers can benefit from emerging

Edge computing technologies have been around for several Edge AI solutions
years, allowing data to be collected and processed as close to The rise of Edge AI is driven by the increasing need for faster
its source as possible, at the edge of networks. decision-making and processing in various applications. With
the proliferation of IoT devices, such as wearables in healthcare
With the rise and growing maturity of AI, edge architecture is or industrial edge devices in general, along with the overall
often enhanced by local AI applications on edge devices such increase in data volume generated, Edge AI provides a solution
as IoT devices, smartphones, and other hardware, minimizing to the challenges associated with traditional cloud computing
the need for extensive cloud resources. This enables improved models.
intelligent decision-making capabilities in real time, along
with reduced latency, enhanced data privacy and lower band- By leveraging local processing power, organizations and users
width usage. can achieve quicker insights and responses, which is espe-
cially crucial in industries that require immediate actions
based on real-time data analysis, for example where IoT is
used for damage detection or damage prevention.

In the same way, insurers can benefit from emerging Edge AI

solutions in B2B and B2C business. Use cases in the insurance
industry exist, for example, in motor insurance, where built-in
AI-based smart systems can not only offer driving assistance
but also deliver analytic data, e.g. in the event of an accident.
Wearables with AI applications support customers in a health-
ier lifestyle, and IoT devices with intelligent decision capabili-
ties can be used in an industrial or private home context to
reduce the risk of damage, e.g. through leakage surveillance or
early warnings of device failures based on analysed data that
deviates from normal patterns. See Tech Trend Radar 2024
for more details on Edge computing.

Keyfact
The Edge AI software market is projected to grow at
a CAGR of 24.7% starting from US$ 1.92bn in 2024 to
US$ 7.19bn by 2030.
Source: marketsandmarkets.com
Munich Re Page 60/111
Tech Trend Radar 2025

Connected Experience

Autonomous Interactions
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Trial & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Autonomous Interactions and Machine Customers can automate − AI relies on complex systems and data, which can sometimes lead
tasks, optimize processes and reduce the need for manual labor, to errors or inaccurate outputs, potentially causing harm, creating
leading to cost savings and higher operational efficiency liabilities, or leading to other kinds of financial loss
− They enable highly tailored experiences for customers. For example, − AI systems can perpetuate biases and discrimination, leading to unfair
autonomous systems can provide customers with customized outcomes or treatment of certain groups
insurance solutions − AI systems can increase the risk of business interruption (BI) due to
− Machine Customers can facilitate real-time transactions and auto- cyber-attacks or technical errors, which can lead to significant financial
mate decision-making, making financial interactions faster and more losses and reputational damage
efficient

Overview AI adoption grew significantly in 2024

Autonomous Interactions refer to AI systems that engage AI is being used by more and more companies to automate pro-
directly with customers. AI-driven chatbots are one example. cesses and decision-making. Autonomous Interactions, for
Machine Customers are AI agents that interact autonomously example, leverage Artificial Intelligence to perform actions, make
with businesses to complete tasks traditionally managed by decisions, or respond dynamically to changing conditions.
people. These systems make decisions based on data analysis Machine Customers and Autonomous Interactions represent
and real-time insights, performing functions such as buying specific use cases for AI Agents.
products.
McKinsey & Company’s 2024 Global Survey on AI, for example,
The use of Machine Customers comes hand-in-hand with shows 72% of respondents using AI for at least one function.
new risks. These AI systems can be used to contract with other Generative AI is being used by 65% of respondents.
parties or perform tasks autonomously. If the AI system makes
mistakes, it can give rise to liabilities – which is why some The use of AI comes with increased awareness about AI risks.
insurers have already incorporated AI exclusions in their prod- For example, AI and Generative AI were ranked as the #1 concern
ucts. This could leave companies unprotected as the “age of in the “Emerging Technologies” category across five distinct tar-
AI agents” advances. Understanding the capabilities and limi- get segments in Munich Re’s RiskScan 2024 report. The grow-
tations of Machine Customers, Autonomous Interactions and ing adoption of, and investment in, AI has led to a breakthrough
AI agents is crucial for harnessing their full potential, under- in “AI Agents.” AI Agents can be understood as systems or pro-
standing the risks and, finally, insuring them. grams capable of performing tasks autonomously on behalf of a
user or another system. These agentic systems are the founda-
tion of Machine Customers and Autonomous Interactions.

How do AI agents work?

The decision-making abilities of AI agents vary widely. Rule-
based systems rely on fixed logic and are effective for certain
Human-in-the-Loop (HITL) processes are still tasks but lack the flexibility to adapt to new circumstances.
seen as a critical component in various scenarios In contrast, learning agents leverage AI algorithms to process
involving AI agents, although the effectiveness data, identify patterns and refine their behavior over time. These
agents are well-suited for dynamic environments where
of human involvement to mitigate these agents'
adaptability is critical. Probabilistic and utility-based agents
errors is questionable. which optimize actions based on calculated likelihoods or goals
further enhance decision-making. Even advanced agents,
Michael von Gablenz however, are constrained by the quality of their training data
Head of AI Insurance and the boundaries of their programming.
Munich Re Page 61/111
Tech Trend Radar 2025

Human-in-the-Loop (HITL) processes are still viewed as a crit-

ical component in scenarios involving AI agents, although the
impact of human involvement on mitigating their errors is
uncertain. For example, in their paper “The Allocation of Deci-
sion Authority to Human and Artificial Intelligence,” Athey et
al. find that in cases involving well-functioning AI systems,
humans in the loop will pay less attention and miss errors pro-
duced by AI systems, rendering HITL processes ineffective. As
such, it might be preferable for organizations to build less
powerful AI Agents if humans are to pay attention and HITL
processes are to be used as important risk-mitigating tools.

Risks and opportunities

While Machine Customers and Autonomous Interactions can Human-in-the-Loop Human-out-of-Loop
reduce costs and increase revenues while making economic
Assisted Intelligence Automated Intelligence
interactions faster, they also introduce new risks, including
Fixed
Systems assist humans in tasks Systems automatically execute
errors in AI model outputs, bias and discrimination. To name and/or decision making. They simple or routine tasks. They do
but a few implications, BI exposure increases under cyber cov- do not learn or adapt. not learn or adapt.
erage and negligent supervision exposures increase, alongside
Augmented Intelligence Autonomous Intelligence
the risks that are most commonly considered, namely data
Systems assist humans in tasks Systems act on their own to
Adaptive

privacy and governance, reliability, security, transparency and and/or decision making. They complete complex tasks and
fairness (AI Index Report 2024 – Artificial Intelligence Index). make precise recommendations make decisions. They learn to
and learn to improve accuracy improve accuracy over time.
For instance, an AI agent system could automate the entire over time.
process of forecasting supply and demand, managing inven-
Source: Based on Rao, Dr. A.S., Verweij, G. (2017). Sizing the prize - What’s
tory, and placing orders as a Machine Customer. Once the the real value of AI for your business and how can you capitalise? [online].
system has analyzed vast amounts of data – such as historical Available at https://www.pwc.com/gx/en/issues/analytics/assets/pwc-ai-
sales trends, weather forecasts and market conditions – it analysis-sizing-the-prize-report.pdf

can predict the future demand for products like winter coats.
Based on this prediction and the available supply stock, the
Machine Customer triggers an order with suppliers autono- This process gives rise to several risks. If the AI relies on biased
mously through a connected e-commerce platform, such as an or outdated data, it might place an excessive order, resulting in
automated procurement system. The system might place the overstocking and higher costs. The lack of transparency in AI
order directly with manufacturers, negotiating terms and decision-making processes could also make it challenging to
quantities based on predefined rules and real-time pricing assess why certain decisions were made which then result in
data. It then makes the necessary logistics arrangements to financial loss. aiSure™, the flagship product from Munich Re’s
ensure the inventory is replenished on time, interacting with Insure AI team, offers a mechanism for risk transfer to the
warehouse management systems to maintain appropriate model provider in the form of a model performance guarantee,
stock levels. enabling confidence and certainty regarding the financial
benefits of AI adoption.

Conclusion
Machine Customers and Autonomous Interactions through
AI agents offer transformative opportunities across diverse
sectors, but require thoughtful implementation to balance
their autonomy with human oversight. Addressing risks
through transparency, robust testing, ethical frameworks, and
risk transfer mechanisms is essential for their responsible
deployment.
Munich Re Page 62/111
Tech Trend Radar 2025
→ Data & AI
→ Healthy Human
→ Connected Experience
→ Cyber & Crypto
→ Redefining Industries

4 Cyber & Crypto

Munich Re Page 63/111
Tech Trend Radar 2025

Cyber & Crypto

From decentralized data

economies to deepfake defense,
navigating the challenges of a
rapidly evolving landscape

Of all technological trends, Crypto and The Decentralized Data Economy, which relies on cryptocur-
rencies using blockchain and Ethereum technologies, experi-
Cyber consistently surprise us with new enced tremendous growth and remains on a trajectory toward
opportunities and risks. Every day brings greater adoption and utility in business. Facilitating cheaper,
faster transactions that provide an immutable audit trail
a new development in this area, attracts consumers and investors alike. But the Decentralized
challenging us all to identify and weigh Data Economy brings with it a number of risks, including lack
of regulatory oversight and a history of facilitating nefarious
up the opportunities and risks, even if activity. Automated Compliance that incorporates artificial
they are not readily discernible when intelligence is revolutionizing what was recently a labor-inten-
sive, largely manual task. As in other areas, automation allows
evaluating the nascent technologies. practitioners to concentrate efforts on identifying previously
This year, several trends have captured unnoticed risks and mapping more effective strategies. The
rising threat of mis- and disinformation, one of the big global
our attention. risks, has heightened the need for effective Deepfake Defense
mechanisms. Deepfake tech can be used to create convincing
but false content that can have serious consequences. The
evolution of AI is driving Non-Human Identity (NHI), which
connects millions of systems and devices. As NHI is more
ubiquitous, the need for security and privacy standards is
critical. AI is also central to developing the Digital Immune
System – a concept derived from human immunity, whereby
entire systems are engineered with a holistic approach to
maintaining the confidentiality, integrity and availability of
information. The rapid convergence of emerging technologies
in the cyber landscape has the potential to increase exposures
and, as a result, widen the gap between the potential total
economic loss from cyber risks and the covered risks by insur-
ance. Insurers are offering solutions to close this gap continu-
ously. The overall ambition is to strengthen resilience to cyber
risks, because they are the flip side of digitalised relationships,
processes and products. Munich Re shares analyses of the
cyber risk landscape annually in its report “Cyber Insurance:
Risks and Trends 2025”. Munich Re is committed to investing
in collaborations to improve underlying data, further develop
sophisticated cyber modelling and innovate solutions.
Munich Re Page 64/111
Tech Trend Radar 2025

Cyber & Crypto

Trend evolution and impact on the value chain

high medium low

Trends Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

& Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Cyber &
Crypto
Decentral-
ized Data assess
Economy
Automated
assess
Compliance
Digital
trial
Assets
Deepfake
trial
Defense
Non-Human
adopt
Identity
Digital
Immune assess
System
Munich Re Page 65/111
Tech Trend Radar 2025

Cyber & Crypto

Decentralized Data Economy

Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Assess & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Monetization of standardized data-sharing interfaces − Data protection as a challenge

− Integration of insurance products into broader (financial, health) − Security and robustness of APIs in data transfers
services; enhanced risk assessment; personalized insurance offers − Data strategies need to be redesigned
based on individual customer profiles

Overview Open data ecosystems: unlocking

Insurance companies are increasingly being encouraged to value through secure data sharing
enhance their data storage, quality, and interfaces. By doing so,
they can ensure they remain more than just providers of risk The concept of a Decentralized Data Economy is fundamen-
capital and maintain strong customer interactions. Opening tally reshaping the way organizations and individuals interact
up to third parties through standardized application program- with data. The aim is to enable organizations to share data
ming interfaces (APIs), with customer consent, can provide securely without fear of losing it, while giving individuals the
access to relevant insurance data and create more collabora- opportunity to distribute their data according to their prefer-
tive opportunities. Potential examples include tax-relevant ences and exchange it for benefits. At the core of this is the
insurance information for providers of tax return software. As data ecosystem – an evolution from digital ecosystems – ,
Blockchain approaches are still in the experimental stage and which forms the backbone of innovative products and services
have not yet had any groundbreaking successes in the insur- driven by seamless data transfers and interoperability.
ance industry, this year’s focus is on innovative forms of data
sharing that go beyond B2B for more customer-centricity. Data ecosystems allow previously isolated insurance products
to evolve into holistic, customer-centric risk management
solutions. For instance, by leveraging property geodata, insur-
ers could automatically incorporate building components like
solar panels into coverage plans. Large Asian insurers have
demonstrated the potential of these ecosystems, especially in
health insurance.

Advances in technologies like AI, Blockchain, and cloud com-

puting support real-time data integration, which allows health
insurers to integrate data from wearable devices to track phys-
ical activity, paving the way for dynamic premium adjustments
or wellness incentives.

This reflects the growing importance of data ecosystems in

achieving unprecedented levels of data sharing and interoper-
ability between stakeholders across the entire insurance value
chain.
Munich Re Page 66/111
Tech Trend Radar 2025

The drive for interoperability is already underway. Initial

approaches to the interoperable use of data can be seen in
institutional (B2B) data transfers between insurers and their
partners. While these interfaces initially served insurers,
their true potential lies in unlocking value for customers. With
the customer’s consent, their insurance-related data can be
linked to third-party systems via an API – while the provision
and use of the API can also be monetized. Standardized data
interfaces facilitate smoother data transfers, paving the way
for insurers to monetize their data assets while delivering
enhanced outcomes to their customers.

In Singapore, customers can already get a holistic view of their

overall financial situation by participating in the Singapore
Financial Data Exchange. The regulatory landscape is evolving
to support this shift toward data ecosystems. One major mile-
stone is the Framework for Financial Data Access (FIDA).
Introduced by the European Commission in June 2023, this
initiative establishes a regulated environment for custom-
er-centric data sharing. Supported by regulatory initiatives,
insurers have a unique opportunity to lead the way in creating
a Decentralized Data Economy where trust, innovation, and
customer sovereignty form the basis of future growth.
Munich Re Page 67/111
Tech Trend Radar 2025

Cyber & Crypto

Automated Compliance
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Assess & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Improved efficiency, accuracy, and client experience − Requires supervision and maintenance to ensure accuracy and
− Reduced costs of compliance reliability
− Improved risk profile − Mistakes can be very costly and incur greater cost to remediate
− Increased employee satisfaction − New laws challenge developers to roll out updates continuously
− Improved customer satisfaction and loyalty

Overview Potentials and benefits of

Automated Compliance systems take the old processes of Automated Compliance
tracking frameworks to spreadsheets, and spreadsheets to
manual confirmation, to the next level. They write reports and For the financial sector, the cost of compliance is soaring due
generate logs. They identify anomalies and prioritize remedia- to a continuously evolving regulatory landscape and greater
tions. While regulatory compliance was once tedious and globalization, demanding compliance with what are often
labor-intensive, automated systems streamline the process. complex and conflicting rules. For insurers, using AI systems
Automated regulatory compliance has experienced rapid to continuously monitor transactions for suspicious activity
growth with the advancements in artificial intelligence. Adop- and flag potential compliance issues in real time could support
tion of automation can greatly reduce the cost of compliance. insurance fraud investigations and lower claims costs.

Automating the management of proliferating state, national

and global data privacy and cyber security mandates, and gen-
erating automated audit trails to document compliance activi-
ties, can result in unprecedented efficiencies. This can free up
human resources to focus on more complex and rewarding
tasks. Eliminating, or at least minimizing, mundane activities
can improve employee satisfaction and increase retention of
valued personnel.

Automated compliance has a vast range of applications.

Financial institutions can use automated systems to monitor
transactions, detect fraud, and ensure compliance with regula-
tions such as Know Your Customer (KYC) and Anti-Money
Laundering (AML).

Several countries have adopted expansive privacy laws. Within

countries, states and municipalities have adopted their own
privacy laws. All of these laws have their own, often unique,
sometimes complementary or overlapping requirements.
Compliance automation can include tools that automate data
subject requests, data mapping, and consent management,
ensuring that organizations meet data privacy requirements.
Munich Re Page 68/111
Tech Trend Radar 2025

“Automating legal compli-

ance can simplify the under-
writing process. Software
helps enhance consistency
and reduce errors.”
Tim Zeilman
Vice President, Global Cyber
Product Owner, HSB

Automated systems can help organizations assess security less regulated industries can accept higher rates of error, the
vulnerabilities, implement security controls, and monitor financial services, healthcare, critical infrastructure and other
network activity to comply with cybersecurity regulations. highly regulated industries require much lower error rates.
Compliance automation can help organizations track and
manage supplier compliance with regulations related to envi- At this time, while error rates of automated compliance soft-
ronmental, social, and governance (ESG) standards. ware have improved since they were initially introduced, they
still require evaluation and continued supervision by humans
While automated compliance holds great promise, it is not prior to, during and following implementation.
the kind of technology that we can set and forget. Automated
systems require human supervision and maintenance. While
Munich Re Page 69/111
Tech Trend Radar 2025

Cyber & Crypto

Digital Assets
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Trial & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Enhanced liquidity: Fractional ownership and ETFs boost market − Cybersecurity: Digital assets vulnerable to hacks and smart contract
access and depth. flaws.
− Efficiency: Smart contracts cut costs by automating key processes. − Legal uncertainty: Despite regulatory efforts such as the EU’s MiCAR,
− Faster settlement: Atomic settlement ensures quick, error-free questions surrounding ownership and governance persist.
transactions. − Operational risk: Failures in digital asset platforms, custody services,
− Fewer errors: Single-transaction settlements minimize counterparty or technology providers can disrupt services - strong resilience is key.
risks and failures. − Regulatory complexity: Global regulatory fragmentation complicates
− Rising insurance demand: Growth in digital assets drives demand compliance.
for tailored coverage. − Lack of standards: Missing global standards and interoperability hinder
adoption

Overview Adoption is accelerating

Digital assets, such as cryptocurrencies, tokenized assets, and The growing importance of digital assets Bitcoin ETFs have
NFTs, have further cemented their existence in the financial been paving the way for broader institutional adoption, helping
industry. The record-breaking launch, in terms of growth, of to establish digital assets as a potential alternative to tradi-
Bitcoin and Ether ETFs in the course of 2024 demonstrated tional investments. In 2024, the approval of Ether ETFs further
the resilience and maturity of this asset class. Compared to cemented Ether’s position as the second-largest cryptocur-
traditional investments such as gold, digital assets offer more rency. According to on-chain data, U.S.-listed Bitcoin ETFs
flexibility, transparency, and accessibility, leading the way for amassed $109 billion in assets within just eleven months –
innovation in the financial industry. Within the digital asset a figure nearing the $121 billion that U.S.-listed gold ETFs took
space, two emerging trends in particular, Decentralized Physi- nearly two decades to achieve.
cal Infrastructure Networks (“DePIN”) and Real-World Asset
(“RWA”) Tokenization, are currently showing promising addi- Adoption is accelerating globally, with financial institutions,
tional use cases for blockchain technology. In order to generate banks, and asset managers actively entering the market. Major
new opportunities in this expanding asset class, it is important players such as Fidelity, JPMorgan, and Deutsche Bank are
for insurers to keep up with developments in the digital asset developing digital asset products, adding custody solutions
ecosystem. and staking services into their offering. This wave of institu-
tional participation underscores the mainstream attention that
digital assets are attracting, reinforcing their maturity and
scalability.

At the same time, tokenized assets are transforming access to

traditionally illiquid markets, including, but not limited to, real
estate and fine art. By enabling fractional ownership, tokeniza-
tion democratizes high-value investments, allowing everyday
investors to participate in markets that were once out of reach.
Munich Re Page 70/111
Tech Trend Radar 2025

Key trends in the Digital Asset

ecosystem
Decentralized Physical Infrastructure Networks (DePIN) inte- Tokenization is transforming how assets – both tangible, like
grate blockchain technology with physical infrastructure, such property and commodities, and intangible, such as intellectual
as telecommunications and decentralized storage, fostering property and digital content – are bought, sold, and owned.
community-driven business models. By leveraging blockchain It allows for fractional ownership, making these assets more
for secure, transparent coordination and incentive mecha- liquid and accessible to a broader range of investors. For
nisms, DePIN ensures trustless participation and verifiable example, BlackRock’s tokenized fund “BUIDL” provides access
transactions. For example, Helium, a company Munich Re to infrastructure investments through blockchain, enabling
Ventures invested in in 2016, uses blockchain-based token greater transparency and flexibility for investors. Similarly, in
rewards to incentivize users to build and maintain decentral- 2024, Siemens issued a € 300 million blockchain-based bond,
ized wireless networks. This strategy not only lowers opera- leveraging tokenization to streamline issuance processes,
tional costs but also redistributes ownership to individuals, bypass traditional intermediaries, and reduce transaction
promoting a more inclusive system. Industry experts highlight costs. In total, tokenization of real-world assets (excl. stable-
DePIN’s transformative potential in industries like IoT and coins, which are a cryptocurrency pegged to fiat currency to
telecommunications, enabling decentralized coordination, maintain a stable value) have grown to more than US$ 12bn
security, and monetization at scale. (as of October 2024, source).

Increase in network
usage leads to better
token rewards
Other Commodities Tokenized Treasuries Private Credit

12.5

10.0
Total RWA Value (US$B)

Supply-side growth
DePin Token rewards
attracts users and incentivize infrastructure
entrepreneurs Flywheel deployment 7.5

5.0

2.5

User-generated fees
boost service offerings 0.0
Jan-20 Jul-20 Jan-21 Jul-21 Jan-22 Jul-22 Jan-23 Jul-23 Jan-24 Jul-24

Source: Datsenko, A. (2023). Decentralized Physical Infrastructure Source: Sharma, S., Tan, C. (2024). RWAs: A Safe Haven for On-Chain
(DePIN) Explained [online]. Available at https://ideasoft.io/blog/ Yields? [online]. Binance. Available at: https://www.binance.com/en/
what-are-decentralized-physical-infrastructure-networks-depin/ research/analysis/rwas-a-safe-haven-for-onchain-yields

Tokenization represents a fundamental shift in

how we understand asset ownership, creating
immense opportunities for liquidity and accessi-
bility. However, it also introduces unique risks –
particularly around smart contract vulnerabilities
and challenges posed by regulatory developments.
As insurers, our role is to bridge this gap, providing
confidence and protection to stakeholders by
addressing insurable risks in this evolving land-
scape.
Dr. Andre Knoerchen
Head of New Tech Underwriting
Munich Re Page 71/111
Tech Trend Radar 2025

Central to tokenization are smart contracts, which automate Beyond these challenges, tokenization offers transformative
the issuance, transfer, and enforcement of tokenized assets. benefits for industries beyond finance. By way of example, real
These programs or algorithms on the blockchain reduce the estate investments become more accessible through frac-
need for intermediaries, ensure real-time execution of transac- tional ownership, while supply chain management can utilize
tions, and minimize the risk of human error. For example, a tokenized commodities for enhanced traceability. Even intel-
smart contract could automatically distribute revenue from a lectual property, such as music rights or patents, can be
tokenized property to its fractional owners, ensuring transpar- tokenized, unlocking new revenue streams and improving
ency and accuracy. By providing tamper-proof execution, ownership transparency.
smart contracts enhance trust and efficiency, making
tokenized assets an attractive option for investors and busi-
nesses alike.
Conclusion: The future of Digital Assets
The leading Digital Assets are no longer speculative novelties
However, reliance on smart contracts also introduces risks. – they have become essential to the financial ecosystem. The
Vulnerabilities in their code can be exploited, potentially lead- adoption of Bitcoin and Ether ETFs highlights their growing
ing to financial loss or unauthorized access to assets. Recent legitimacy and potential. At the same time, trends like DePIN
research conducted by SolidityScan revealed that around 10% and tokenization are expanding blockchain's applications
of audited smart contracts contained critical flaws, emphasiz- beyond finance, touching industries like infrastructure and real
ing the need for rigorous testing and continuous monitoring. estate. For insurers and other stakeholders, staying ahead of
Additionally, legal disputes may arise if the terms of a smart these developments is important to capturing new business
contract conflict with traditional regulatory frameworks or are opportunities, managing emerging risks, and driving innova-
ambiguously coded. As tokenization becomes more wide- tion. Digital assets, such as Bitcoin, Ether, stablecoins and
spread, mitigating these risks will be essential to ensure its tokenized RWAs are not just a passing trend – they represent a
long-term success. profound shift in how value is created, transferred, and insured.
Munich Re Page 72/111
Tech Trend Radar 2025

Cyber & Crypto

Deepfake Defense
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Trial & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Development of AI tools for detecting and mitigating synthetic media. − Detection must be nearly flawless to maintain trust and credibility.
− Educational campaigns to raise awareness about deepfake risks and − Rapid advances in deepfake technology challenge current defenses.
recognition. − Missteps or lack of transparency could erode public confidence.
− Cross-industry partnerships to establish standards and strengthen
defenses against misinformation.

Overview Combating digital deception

Deepfake and disinformation defense involves identifying, mit- In an era dominated by digital transformation, deepfakes and
igating, and preventing the spread of manipulated media and disinformation have emerged as formidable threats to infor-
false information. By leveraging advanced AI tools, organiza- mation integrity. Deepfakes, a product of Generative AI, can
tions can detect artificial content, authenticate media, and convincingly fabricate audio, video, and images. While these
safeguard against the misuse of deepfakes in scams, misinfor- technologies demonstrate impressive advances, their misuse
mation campaigns, and reputational attacks. This defense enables identity theft, fraud, and erodes trust in digital content.
strategy is critical for maintaining trust, protecting sensitive Similarly, disinformation – the deliberate spread of false infor-
information, and ensuring the integrity of communication mation – undermines societal cohesion and democratic pro-
channels. As deepfake technology evolves, proactive measures, cesses, making it one of the most pressing challenges of the
including robust detection systems and public awareness, are digital age.
essential to counter these sophisticated threats effectively.
Deepfake technology is evolving rapidly, making it increasingly
difficult to distinguish authentic media from fabricated con-
tent. Cybercriminals exploit these tools for personal gain, from
impersonating executives in corporate scams to disseminating
fake political statements aimed at manipulating public opin-
ion. The consequences extend beyond the immediate targets,
undermining trust in media and institutions globally.

In a quest to undermine the truth and trust, disinformation

campaigns leverage false or misleading content to disrupt
societies and influence decision-making. Social media
platforms amplify the reach of such campaigns, spreading
disinformation at unprecedented speeds. This creates echo
chambers that reinforce false narratives, making it harder to
combat their influence. Governments, organizations, and
individuals alike are grappling with the far-reaching implica-
tions of this digital deception.
Munich Re Page 73/111
Tech Trend Radar 2025

Despite the risks, opportunities exist to address deepfake and weakened public trust in digital media. Without transparent
disinformation threats effectively: and effective solutions, this erosion of trust could intensify,
impacting businesses, institutions, and individuals who rely on
− Advancing detection technologies the credibility of digital information.
AI-powered detection tools can analyze content authenticity
and identify deepfake media. These solutions can integrate
with social platforms and news outlets to flag suspicious
Building a proactive defense
material, restoring public trust in digital ecosystems. To counter these challenges, organizations must take pro-
active measures. Investing in robust detection systems,
− Strengthening public-private collaboration establishing transparent policies, and fostering cross-industry
Partnerships between governments, tech companies, and collaboration are critical steps toward mitigating digital
researchers can lead to unified strategies for combating deception. Additionally, public awareness initiatives play a vital
digital deception. By pooling resources and expertise, stake- role in educating individuals on how to identify fake content
holders can create global standards and share best practices and comprehend the broader implications of disinformation.
to enhance defense mechanisms. By addressing these challenges with urgency and precision,
we can work toward to create a safer and more trustworthy
− Enhancing media literacy digital ecosystem where the integrity of information remains
Public education campaigns focused on media literacy can uncompromised.
empower individuals to critically evaluate digital content,
reducing the impact of misinformation. These programs play
a vital role in creating a digitally resilient society.
Deepfakes and insurance
The proliferation of deepfakes has vast and uncertain implica-
− Increasing regulatory scrutiny tions for insurance, from underwriting through to claims, all
A number of countries and regions have laws that regulate implicating fraud. For example, deepfakes can be used in
deepfakes, including the European Union, China, South applications for insurance to misrepresent the security of a
Korea and individual states within the United States. In the computer system or to enhance the appearance of a property.
US, a number of jurisdictions have laws that govern the In claims, deepfakes can be used to make a fraudulent claim
development and use of deepfakes. These laws range from or could be the basis of a claim. For example, deepfakes are
criminalizing the unauthorized mimicking of a voice or increasingly used in social engineering to impersonate clients,
depiction of an identifiable individual or the use of a minor’s vendors or internal executives in attempts to gain unauthor-
image, to requiring disclosure when deepfakes are used. ized access to systems or to persuade someone with authority
to make a fraudulent funds transfer or misdirect shipments of
inventory. This means that detection of deepfakes is, and will
Risks and challenges remain, critically important in order to mitigate insurance risks.
Detection systems face the critical challenge of achieving
near-perfect accuracy, as even minor errors can significantly
erode confidence in their effectiveness. However, this risk can
Defending reality
be mitigated by combining automated detection tools with The rise of deepfakes and disinformation presents one of
human oversight to enhance reliability and trust. Recent the most pressing challenges of the digital age, demanding
advances in machine learning and media-modality fusion immediate and sustained action. While the technology behind
techniques have shown promise in improving detection rates these threats is impressive, its misuse has far-reaching impli-
by analyzing multiple features simultaneously, such as facial cations for trust, security, and societal cohesion. Combating
expressions and audio signals. digital deception is not just a technological endeavor; it’s a
collective responsibility involving governments, industries, and
Another pressing issue is the rapid evolution of deepfake individuals.
technology. As detection tools improve, so do the methods
used to create more sophisticated and convincing deepfakes. As we develop advanced detection tools, foster cross-industry
Techniques like adversarial training and multi-modal data collaboration, and enhance public awareness, the question
integration are emerging as vital tools in this ongoing techno- remains: Are these efforts living up to the hype? The answer
logical arms race, which demands continuous innovation to depends on our ability to stay ahead of evolving threats and
stay ahead of malicious actors (Gupta et al.). The widespread turn defensive strategies into proactive solutions.
proliferation of deepfakes and disinformation has already

Key Findings
According to a McAfee survey, 70% of respondents stated they
are not confident in distinguishing between real and cloned
voices. This highlights the growing sophistication of AI voice
cloning technology and its potential to exploit vulnerabilities in
human trust.
Munich Re Page 74/111
Tech Trend Radar 2025

Cyber & Crypto

Non-Human Identity
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Adopt & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− NHI enables secured systems to communicate and share information − NHIs increasingly targeted in cyber attacks.
without human interaction. This promotes efficiency and enhances − Many NHIs are not monitored and are poorly secured.
security. − The current market for NHI is highly fragmented.
− When properly configured and monitored, NHI reduces unauthorized
access.
− Facilitates compliance with industry standards.

Overview NHIs can make systems more efficient,

Identity Access Management (IAM) is a core component of secure and compliant but suffer from
cybersecurity. IAM refers to how users can access digital vulnerabilities
resources and what they can do with them. Non-Human Iden-
tity (NHI) – also sometimes referred to as Machine Identity – NHI is being used increasingly in many businesses. Re/insur-
is a component of IAM and is increasingly important within ance is an ideal application. For example, Application Pro-
information technology and cybersecurity. NHI includes all gramming Interfaces (APIs) enable reinsurers and cedents
methods of managing access to digital resources that do not to share data and services, creating a seamless experience
involve humans. For example, a password is a conventional for insureds. But NHI is about much more than one system
IAM method whereas an automated method enabling one connecting to another. In many cases, multiple systems
system to access another is an NHI method of access. NHIs are connected. NHIs enable IoT, OT equipment, AI Large
include service accounts, APIs, IoT devices, and bots (see Language Models, Databases, Robotic Process Automation
figure below). and a large number of other tools.

While many businesses and industries have deployed NHI, the

technologies suffer from several weaknesses. NHIs tend to be
insecure, often using security credentials that are static, and
have excessive privileges and weak encryption. They are
increasingly being targeted in cyberattacks. According to the
NHI Management Group, 80% of identity breaches involved
NHIs. Compromised NHIs frequently lead to successful cyber-
attacks. The lack of NHI security has led to recent government
initiatives to provide guidance and to publish best practices.

Key Findings
NHI use is proliferating rapidly but suffers from security vul-
nerabilities. Government guidance and standards are pushing
industry to improve security so that the technology can
advance safely.
Munich Re Page 75/111
Tech Trend Radar 2025

With the exponential growth of Non-Human Identities (NHI),

enterprises must adopt comprehensive platforms that move
beyond traditional username-and-password models to ensure
robust IAM programs.
Sidra Ahmed Lefort
Munich Re Ventures

The Australian Signals Directorate’s Australian Cyber Security Also in October of 2024, CISA issued “Internet of Things
Centre (ASD’s ACSC) in collaboration with national cyber Security Acquisition Guidance” and the United States
security authorities of the United States, United Kingdom, National Institute of Standards and Technology (NIST) has
Canada, New Zealand, Germany, Netherlands, Japan, and the been developing and maintaining its “Cybersecurity for IOT
Republic of Korea, has published “Principles of Operational Program” since at least 2017. While the security of NHIs is a
Technology (OT) Cyber Security”. The document provides best challenge, the proliferation of guidance and standards, as well
practices to help organizations design, implement, and man- as the extensive coordination among nations to develop and
age secure OT environments, ensuring business continuity for maintain them, is encouraging for the future security and
critical services. utilization of NHIs.

Non-Human Identities (NHIs)

Types of NHIs

Service Technical Admin System

API Keys Tokens Secrets Certificates
Accounts Accounts Accounts Accounts

Where Are They Found

Machines / Devices Software Workloads

– Servers – Containers, Microservices

– Mobiles – Virtual Machines (VMs), Databases
– Desktops – Applications, Services, Scripts
– Internet of Things (IOT) – API Processes/Services
– Operational Tech (OT) Equipment – Robotic Process Automation (RPA)
– Tech Equipment used in Industry – AI LLMs, Chatbots

Source: Non-Human Identity Management Group (2025). The Ultimate Guide To Non-Human Identities [online].
Available at: https://nhimg.org/the-ultimate-guide-to-non-human-identities
Munich Re Page 76/111
Tech Trend Radar 2025

Cyber & Crypto

Digital Immune System

Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Assess & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Improved cybersecurity and reduced risks. − Potential over reliance.

− Lower costs due to fewer security incidents. − Data privacy risks.
− More accurate risk assessments. − Requires sophisticated governance by qualified personnel,
− Better customer experience. who are in short supply in the market.
− More robust underwriting data.

Overview A holistic approach

Improvements in cybersecurity at all levels, from software Originally conceived by Gartner, the digital immune system
development to active risk management, influence the rapid is a collection of practices and technologies used in software
development and deployment of digital immune system (DIS) design, development, operations and analytics aimed at
capabilities. A DIS is a holistic approach to an entity’s entire mitigating risks, and improving cyber security and resilience.
IT infrastructure aimed at optimizing the user experience and Key elements of a digital immune system include continuous
reducing system failures. As this technology is expanded a validation, autonomous testing, chaos engineering, observa-
nd improved, all components of the insurance value chain can bility and auto-remediation.
benefit from greater security, resilience and enhanced user
experience. According to Precedence Research, the global digital immune
system market was worth an estimated USD 32.50bn in 2024,
For re/insurers, entities employing proven DIS are likely to be will grow to USD 36.70bn in 2025, and is projected to reach
easier to underwrite. These systems offer increased security, around USD 109.36bn by 2034.
transparency, and an improved user experience overall, leading
to fewer claims and greater profitability. Most of the major advances in the DIS can be attributed to
advances in generative AI. Working in tandem with existing
technologies and practices, generative AI allows for more
efficient and accurate autonomous testing, simplifies chaos
engineering, documents observations, and provides automated
remediation and continuous validation. As these technologies
and processes improve, it is essential for humans to supervise,
maintain, document and contribute to the continuous develop-
ment of the DIS. Due to the challenges presented by the
increasing level of sophistication of ever evolving threats, the
systems also have to evolve to reflect and outpace the threat
landscape.

According to The Business Research Company, cyber threats

are driving the digital immune system market. In particular, the
increased frequency, severity and sophistication of phishing
and malware attacks, interconnectivity and technology vulner-
abilities are resulting in greater interest in, and the need for, the
sort of coordinated approach that the digital immune system
can provide.
Munich Re Page 77/111
Tech Trend Radar 2025

Insurance implications
As adoption of the DIS increases and technology harmoniza-
tion is improved, insurance will benefit in several ways. Cyber
insurance, in particular, will see fewer and less severe claims
related to cyber attacks. Data breaches, particularly due to
cyber attacks, will also become less frequent. Business inter-
ruption, perhaps the most costly component of cyber remedia-
tion, will also likely be reduced, as the DIS provides a docu-
mented system roadmap that can reduce recovery time.

Figure: Six Elements of a Digital Immune System

Software Supply
Observability
Chain Security

Site Reliability AI-Augmented

Digital Immune
Engineering Testing
System

Chaos
Autoremediation
Engineering

Source: Gartner®, Improve Software Quality by Building Digital Immu-

nity, 17 October 2022, figure 2. GARTNER is a registered trademark and
service mark of Gartner, Inc. and/or its affiliates in the U.S. and interna-
tionally and is used herein with permission. All rights reserved.

Conclusion
The digital immune system will make further advancements
over the coming years. The advent of generative artificial
intelligence, together with rapid developments in quantum
computing, will undoubtedly assist us in gaining an advantage
over bad actors and facilitate new gains in risk mitigation
and management.

Key Findings
The digital immune system market is poised to grow at a
CAGR of 12.9% between 2024 and 2034.
Munich Re Page 78/111
Tech Trend Radar 2025
→ Data & AI
→ Healthy Human
→ Connected Experience
→ Cyber & Crypto
→ Redefining Industries

5 Redefining Industries
Munich Re Page 79/111
Tech Trend Radar 2025

Redefining Industries

Remote sensing, renewable

energy, EVs, and climate
resilience: navigating the
changing risk landscape

The convergence of technological The use of remote sensing technologies, such as satellite
imagery and drones, is revolutionizing agricultural insurance
advancements and changing market through improved crop monitoring and risk assessment. The
conditions is redefining traditional transition to renewable energy sources, including solar, wind,
and hydro, is creating new opportunities and risks for the
industries, with far-reaching insurance industry, from investment in transmission and distri-
implications for the insurance sector. bution systems to the development of insurance solutions.
The rise of Electric Vehicles (EVs) is also driving change, with
insurers needing to adapt to new risks and opportunities, from
higher repair costs to cyber threats. Climate Resilience is
becoming increasingly important, with businesses seeking to
improve their understanding and management of climate-re-
lated risks. Meanwhile, the protection of trade secrets is taking
on newfound importance in a globalized, digitalized economy,
with robust risk solutions and collaboration between insurers
and businesses crucial for safeguarding valuable information
assets. Other trends, such as the Hydrogen Economy, the
Humanoid Robotics, and Autonomous Mobility, are also poised
to have a profound impact on the insurance sector, from
enhancing operational efficiency to creating new products and
services. Ultimately, these trends will collectively reshape the
insurance industry, requiring insurers to innovate, adapt, and
respond to emerging risks and opportunities.
Munich Re Page 80/111
Tech Trend Radar 2025

Redefining Industries

Trend evolution and impact on the value chain

high medium low

Trends Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

& Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Redefinding
Industries
Crop
adopt
Intelligence
New Energy
adopt
at Scale
Trade Secret
trial
Protection
Electric
adopt
Vehicles
Hydrogen
trial
Economy
Humanoid
assess
Robots
Autonomous
assess
Mobility
Climate
adopt
Resilience
Munich Re Page 81/111
Tech Trend Radar 2025

Redefining Industries

Crop Intelligence
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Adopt & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Expansion of crop insurance in the food production chain − Technical and accuracy-related limitations
− Continuous crop monitoring − Legal and privacy risks
− Faster and more accurate claims settlement

Overview Remote sensing and crop insurance:

Remote sensing, one of the main trends in the crop insurance Transforming risk management
market, refers to the use of technologies to capture, process
and interpret information about crops and rural areas without Agriculture is inherently vulnerable to climatic risks, biological
the need to be physically present. It makes particular use of agents and environmental variations that can compromise
satellite image processing to assess risks, monitor crops and production and mean significant losses for farmers. In this
help speed up the claims adjustment process. In this context, it context, agricultural insurance plays a fundamental role in mit-
plays a leading role in helping insurance companies offer cus- igating these risks, guaranteeing financial protection for farm-
tomized products for their clients, promoting greater adher- ers. However, the accuracy and efficiency of risk analysis pro-
ence to insurance and a better experience for their policyhold- cesses, crop monitoring, and the settlement of potential claims
ers. Despite the progress made in recent years, it is important are constant challenges for insurance companies.
to note that there is still a significant protection gap, leaving
around 60% of global agricultural production vulnerable, In recent years, remote sensing has emerged as a revolutionary
which represents an impressive US$ 113bn. Remote sensing is tool in this sector, enabling precise and automated crop moni-
key to helping the market tackle this problem. toring. The use of satellite images, drones and IoT (Internet of
Things) sensors has transformed the way insurance compa-
nies assess agricultural risks and determine insurance premi-
ums, while also allowing them to process claims more quickly
and reliably.

Remote sensing and the underwriting

process
Policy underwriting, the stage of risk assessment and pricing,
requires a thorough analysis of the risks associated with a par-
ticular crop and region. Traditionally, this process has relied on
face-to-face visits by experts and limited historical data. With
remote sensing, insurers can access detailed information on
vegetation health, climatic conditions and historical produc-
tion performance through indices like the NDVI (Normalized
Difference Vegetation Index) and the EVI (Enhanced Vegeta-
tion Index).
Munich Re Page 82/111
Tech Trend Radar 2025

Because data is captured cycle after cycle, insurers now have

access to a vast pool. By gaining a more in-depth understand-
ing of the reality and performance pattern of the area in ques-
tion, they can offer customized products adapted to the spe-
cific needs of each farmer. This customization not only offers
insurers more security when accepting/pricing risks. It also
helps to boost sales, reducing anti-selection and contributing
to the territorial dispersion of portfolios, which is essential for
volatile lines such as Agro.

Continuous crop monitoring and Faster and more accurate claims

risk mitigation settlement
Once a policy has been issued, remote sensing allows insurers The moment when a claim occurs is the key point in the farm-
to monitor crop conditions on an ongoing basis, generating er’s experience with their insurer, as this is when they realize
data that makes it possible to identify water stress patterns, the true value of their policy. Although this phase still requires
possible management errors, and damage caused by adverse field inspections, the use of remote sensing has dramatically
weather events. This sort of monitoring is crucial for taking accelerated this process. In practical cases in the United
swift and assertive action – be it by excluding risks associated States, remote sensing is used to assess damage after hail-
with management from the policy, meaning that these are not storms, for example, significantly reducing the time it takes to
covered, or by anticipating the occurrence and impact of analyze claims and sometimes eliminating the need for an
claims on crops. In addition to managing their portfolios more in-person inspection. Once crops are monitored remotely, it is
effectively, insurers become part of the production process possible to assess and prove the occurrence of events (in addi-
with farmers, providing timely service with the quality and agil- tion to hail, these include drought, frost, excessive rainfall, etc.)
ity they need. Data can also be shared with farmers, allowing and determine the extent and severity of the damage caused.
them to contribute to decision-making and improve the effi- This data generates reports that help to settle claims more
ciency of agricultural operations in the field. quickly and decisively.

Figure: Example of image processing using remote sensing capabilities. Turning images into information.

Images processed – two types of NDVI

2020-03-31 2021-02-24 2022-01-15

2023-03-11 2024-01-30 2024-12-25

Real representation of a farm Pure NDVI

Good quality
(healthy)
Poor quality

2020-03-31 2021-02-24 2022-01-15

Polygon (geographical file – image to processed)

2023-03-11 2024-01-30 2024-12-25 False color

composition

Good quality
(healthy)
Poor quality

Source: Munich Re Agro Digital Suite

Munich Re Page 83/111
Tech Trend Radar 2025

“Reducing anti-selection, exploiting the poten-

tial of each crop and region, excluding areas
not suitable for planting, and pricing those that
are suitable fairly, based on the reality of each
risk, is one of the main applications of remote
sensing in crop insurance.”
Luiz Leonardo
Agro UnderwriterI

Challenges and final considerations

Despite numerous advances, the use of remote sensing in agri- The future of agricultural insurance is increasingly digital, and
cultural insurance still faces challenges. Image accuracy can remote sensing will play a key role in building a more resilient
be compromised by cloud cover, and data interpretation and technological sector.
requires investments in training and technology. What is more,
regulatory and privacy issues must be considered when using With this in mind, and with the aim of helping our clients to
geospatial information. tackle the main challenges related to agricultural insurance in
Brazil, Munich Re currently offers a set of tools known as the
As technologies evolve and become more accessible, however, Agro Digital Suite (ADS). Broadly speaking, it includes (i) a
it is undeniable that remote sensing is becoming an essential platform that enables the insurer to distribute insurance, (ii) a
tool for innovation in agricultural insurance. Its use not only mechanism for analyzing risk at plot level based on remote
improves the accuracy of risk and claims analysis, but also sensing, and (iii) an active portfolio management module.
enables more efficient and transparent management for insur-
ers and farmers.
Munich Re Page 84/111
Tech Trend Radar 2025

Redefining Industries

New Energy at Scale

Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Adopt & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Considerable growth potential throughout the renewable/ − Rapid technological development with unproven technologies and
low-emission energy sector resulting new risks
− New concepts of performance guarantees for producing assets and − Increasing risk of natural perils to renewable energy facilities
power market participants − Increasing complexity throughout the value chain
− Changes in power market design require new risk transfer solutions − Increasing weather dependency of the entire power system

Overview Low-emission energy sources

Ongoing electrification requires renewable energy sources like In 2024, the global average temperature exceeded the pre-in-
solar, wind (on-/offshore), hydro power and other technologies dustrial level by 1.5°C for the first time, making it the warmest
to produce electricity with low or zero CO2 emissions. Further- year on record (Source: Copernicus Climate Change Service).
more, to match energy supply and demand and to effectively To limit the temperature increase to at least well below 2°C,
integrate large quantities of volatile energy generation into the net-zero emission commitments have been implemented by a
power system, significant investments in the transmission and majority of nations. The installed capacity of renewable ener-
distribution system, storage facilities and the digitalization of gies, however, still has to increase tremendously between now
the power system are required. and 2030.

Major growth is expected primarily in solar photovoltaics (PV)

and wind (on- and offshore) installations over the next few
years. Significant cost reductions for PV modules are acceler-
ating this trend. Specifically, PV or solar thermal assets are
highly vulnerable to natural perils like hail, storms or flash
floods. To a lesser extent, this applies to onshore wind farms as
well. However, with increasing turbine ratings and heights, the
risk of lightning strikes increases. Advanced lightning protec-
tion systems are being developed to mitigate this risk. Since a
significant share of new offshore wind farm development
zones are located in areas at a high risk of tropical cyclones,
design standards for offshore turbines must be adapted to
withstand local hazards.

Compared to the growth in solar PV and wind power, new

hydro and nuclear power installations look comparatively
small. In absolute terms, however, these technologies remain
relevant for providing base load power during times with low
production from renewable sources. Nuclear power might see
a renaissance in the more distant future with small modular
reactors and new technologies like fusion reactors.
Munich Re Page 85/111
Tech Trend Radar 2025

Power grid Storage

Since the majority of large-scale renewable production facili- The underlying production volatility of wind turbines and solar
ties is distant to load centers, the expansion and reinforcement PV remains a challenge for the stability of the power system.
of transmission systems is a key prerequisite for the success Currently, grid stability is secured through backup power
of the energy transition. plants and demand-side management, providing balancing
power or reducing system demand as the case may be.
Furthermore, a higher number of interconnectors is required to
facilitate the exchange of electricity between regions and Large-scale Battery Energy Storage Systems (BESS) are
countries. This infrastructure is crucial for utilizing geographi- becoming increasingly important, as surplus energy can be
cal smoothing effects, which can help balance out fluctuations stored and fed back into the grid during times of scarce pro-
in power supply and demand. Additionally, a more extensive duction from renewable energy sources. The rapid decline in
network of interconnectors can advance the integration of battery costs make the technology more competitive for short-
power markets, allowing for more efficient and coordinated term storage than conventional backup power plants. As a
energy distribution. By increasing the number of interconnec- result, BESS will play a pivotal role in the energy transition.
tors, we can also create opportunities for further flexibilization
of the power system, enabling it to adapt more easily to chang- However, this technology is not without risk: recent fires in
ing energy needs and conditions. BESS projects around the world demonstrate the associated
risk and the need for protection systems. Improved monitoring
A significant fraction of these transmission systems will be and early warning systems need to be in place to prevent fires
offshore, connecting offshore wind farms or linking island from starting in these installations. Also, stringent rules for
systems to the onshore grid. These projects pose significant separation between neighboring units and surrounding
challenges as they need to respond (partly) to new needs to property need to be adhered to in order to limit fires to single
withstand harsh offshore conditions while increasing trans- compartments.
mission capacity.

The distribution system also requires large-scale investment

to enable the transition from a purely consumer towards a
Power markets
prosumer (producer-consumer) driven system. Digitalized The increasing share of renewable energy sources along with
grids and measuring devices (e.g. smart meters) need to be the associated production volatility require new approaches
introduced to enable prosumers to participate in power regarding system stability and production security. Despite the
markets and to allow the Distribution System Operator (DSO) typical “energy only” markets in which power producers
to steer prosumer production and consumption. According receive compensation per generated MWh, so called “capacity
to the World Energy Outlook 2024 of the International Energy markets” are becoming increasingly important in providing
Agency, total annual investment in transmission and distri- back-up power in times of low production from solar and wind.
bution grids will have risen to US$ 690bn in 2030. In these markets, producers are paid for their available produc-
tion capacity. This results in increasingly volatile power prices,
which pose a relevant risk to industrial offtakers and power
producers. Back-up power providers are heavily penalized in
the event that a given reliability level is not met. Munich Re has
developed new insurance solutions to transfer some of these
risks.
Figure: Estimated Global asset value (bnUSD)
Munich Re provides risk transfer solutions for projects in con-
2030 2023
struction and operation for all of the aforementioned technolo-
0 1000 2000 3000 4000 gies. Beyond that, we offer tailored solutions, e.g. performance
guarantees as well as power market risks.
Solar PV

Wind Onshore

Wind Offshore

Hydro

Nuclear

BESS

Source: Munich RE (2025). Estimated Global asset value [bnUSD].

Based on 1) Data by International Energy Agency (2024). World Energy
Outlook 2024. [online]. Available at https://www.iea.org/reports/
world-energy-outlook-2024, [original data]. Licence: CC BY 4.0
All rights reserved IEA. And based on 2) Data by Global Wind Energy
Council (2024). Global Wind Report 2024. [online] Available at https://
www.gwec.net/reports/globalwindreport/2024
Munich Re Page 86/111
Tech Trend Radar 2025

Redefining Industries

Trade Secret Protection

Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Trial & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Despite the large insurance capacity required, the first product − Pricing requires interdisciplinary expertise (IT, Cyber, Legal)
providers will be able to capture a large market share − Software solutions can compromise security and privacy due to
− Advantageous contractual terms, as the protection of trade secrets comprehensive monitoring
applies for as long as secrecy is maintained − Valuation of trade secrets is complex
− As there is no specific insurance cover for trade secrets, a first mover
approach in this field could be lucrative. The insurance industry
could develop a risk solution jointly

Overview Trade Secret Protection:

A trade secret is an information asset (piece of intellectual A strategic imperative
property) with an inherent economic value. Trade secrets
encompass proprietary formulas, algorithms, strategies, and In a globalized, digitalized economy, the risk of losing trade
processes that give businesses their unique edge. Unlike secrets has never been greater. Software developed in-house,
patents or trademarks, trade secrets derive their value from code, formulas, and the algorithms that form the core of AI
confidentiality. However, this very secrecy makes them par- systems are trade secrets that call for appropriate risk solution
ticularly susceptible to risks such as cyber attacks, insider concepts.
threats, and corporate espionage. Hence, protection against
the loss, manipulation or damage of trade secrets is crucial. Large industry clients, CIOs and Heads of Intellectual Property
are increasingly interested in protection concepts for trade
secrets. The need for robust trade secret protection is not just
a legal or operational concern – it is a strategic imperative.
According to recent statistics, the annual global loss targeting
intellectual property amounts to about $60 billion. With
roughly 80% of today’s S&P 500 companies chiefly relying on
intangible assets, developing risk solutions for competi-
tion-relevant trade secrets is vital.

While international frameworks like the Trade-Related

Aspects of Intellectual Property Rights (TRIPS) agreement,
EU Directive 2016/943, and national laws advise trade secret
owners to take appropriate technical, organizational and/or
legal measures, the actual implementation of these measures
is inconsistent. This is mainly because the legislator only
defines a minimum level of protection and refrains from explic-
itly protecting trade secrets, instead placing that responsibility
with the trade secret holder.
Munich Re Page 87/111
Tech Trend Radar 2025

For insurance companies, this constitutes a market potential. Businessclients should comply with certain verifiable stand-
According to an internal evaluation, there is currently a signifi- ards such as ISO 27001 certification (ISMS) before insurance
cant protection gap, with only ~15% of information asset losses cover can take effect. Potential liability ceilings will likely be
insured, compared to ~60% of property assets. The develop- based on the amount of the inherent economic value of a trade
ment of risk solutions is still in its infancy, with hybrid risk secret and take into account existing safeguards. In general,
solutions being tested. An example would be a business inter- insurers should agree on a minimum set of measures that are
ruption insurance with compensation for financial losses from used to protect valuable information in digital form. These
interruptions caused by cyber theft of trade secrets. For hybrid could include data encryption, security protocols for system
solutions to be effective, collaboration between insurers and access, employee nondisclosure agreements, and various
business is essential. Before calculating premiums, insurers physical security methods. Software solutions that offer data
will have to insist that appropriate precautions are taken to classification, monitoring, and alerting functionas should also
preserve the confidentiality of the trade secret. be considered when designing insurance cover.

The market has recognized the danger of losses

due to the disclosure of trade secrets without any
specific insurance coverage.
Tim Klein
Executive Assistant, ERGO Group
Munich Re Page 88/111
Tech Trend Radar 2025

Redefining Industries

Electric Vehicles
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Adopt & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− New customers' needs and fears − Shift in underlying risk drivers impacting
· Common fears, e.g. range anxiety, residual value, fire risks · frequency through change in driving behaviour, driving characteris-
· Motor insurance and affinity business address EV fears tics, car usage, etc.
− Enhanced data collection to support pricing and underwriting · average claims cost due to increased repair costs or specific work-
− EV claims handling processes as general lever to improve overall force training
process − Less frequent but more complicated to extinguish EV fires
− Partnerships across the entire insurance value chain – from new − New risks related to connectivity and software-predominance of EV,
players to bigger players such as cyber-related risks

Overview Introduction
The ongoing shift from Internal Combustion Engine (ICE) The adoption of EV technology is altering customer insurance
Vehicles to Electric Vehicles (EV) is directly impacting the (re) needs and changing risk profiles. Additionally, the transition
insurance industry and the performance of motor books. from mechanically defined to software-defined vehicles is
Customer needs are changing, as is the risk typology. underway. To stay ahead, we recommend that insurers adapt
their products, pricing, and claims management proactively to
Insurers are well advised to apply a proactive approach when reflect the evolving EV landscape, exploring new partnerships
it comes to EV adoption in their respective markets to stay and services to remain competitive.
ahead of the curve across the entire insurance value chain.
This starts with defining the right products and services,
followed by new pricing techniques, and ends with adapted
Opportunities
EV claims management and new partnerships. EVs sales uptake also brings a variety of opportunities within
the insurance space. One of the most important ones is taking
customer fears and demand seriously and providing the right
insurance products based on these fears. The more fact-based
the concerns, the better (e.g. see fire risk below). The insurance
industry as a whole might need to rethink its offering within
the service landscape. For example, roadside assistance ser-
vices for EVs require dedicated services when vehicles are
having battery problems or are even catching fire. This calls for
much more intense customer interactions as before – people
don’t automatically know what to do.

The residual value of EVs is now another big topic. It is cur-

rently seen as a threat for private individuals when buying new
or used EVs. Insurance products aimed at protecting the resid-
ual value based on the right estimates and calculations can be
seen as an opportunity within the market, provided the right
knowledge is readily available. Customer needs related to EV
also differ from one market to the other. Flooding and marten
bite risks or battery disposal regulations are different. But in
most markets, these customer needs are currently being taken
into account in product development.
 Munich Re Page 89/111
Tech Trend Radar 2025

EV tailored services EV product add-ons Cover driven by local context

High

Potential current target Monitor and plan in long term

Battery damage
Liability
(via external factors)
(cable and wallbox)
Extended warranty Partnerships
EV residual value
protection
Customer demand

Battery disposal

Roadside assistance
EV fleet Cyber attacks
management

Animal bites

Autonomous driving

Complexity/time & cost to implement High

Source: Munich Re

Affinity business is also creating new opportunities for insur- New EV claims handling processes are currently being devel-
ers. Demand for Extended Warranty and GAP insurance for oped at dealerships (e.g. special local regulations to improve
electric vehicles is rising rapidly. OEMs are exploring ways to safety when working on damaged EVs). Processes need to be
offer these products to their EV customers through digital understood at insurance companies and a more digitalized
channels or at dealerships. This can help maintain customer approach to claims handling is required – especially when
relationships and support dealership operations. As electric working with new market entrants from Asia. This can present
vehicles require less maintenance, resulting in fewer customer an opportunity to improve the overall claims handling process
interactions, offering these products can be mutually benefi- and also make it future-proof – not only for EV business.
cial.
Lastly, EV business is no longer a pure OEM business. Lots of
Enhanced data collection and new types of pricing and under- new partnerships spanning the whole EV ecosystem are
writing can also bring new opportunities. EV data is currently already visible and growing. This also includes new potential
still a rare commodity, as most insurers are trying to collect for insurers – be it with new market entrants (manufacturer) or
their own data and learn from it. But EVs also provide plenty of within the entire service offering to EV customers, which can
new data points which can be used to make underwriting and be offered within the world of insurance as well. The graphic
pricing work more accurate and tailored to EV drivers. In provides an overview of the main current opportunities within
response to this need, Munich Re has been working to develop a special ranking system.
a deeper understanding of the risks associated with EVs, and
to help insurers assess the level of risk associated with newly
launched EV models, even in the absence of historical data.
Munich Re Page 90/111
Tech Trend Radar 2025

Risks
While the underlying risks associated with motor insurance “As Electric Vehicles
remain the same for electric vehicles as for internal combus-
transform the automotive
tion engines, there are differences in terms of risk levels and
the underlying drivers associated with them. industry, insurers must
accelerate their own trans-
Munich Re studies have, for example, shown that, all things formation to stay relevant
being equal, accident frequency tends to be higher for EV than and deliver value in a rapidly
for ICE. A few key drivers have been identified to explain this
increase: changing world.”

− Most buyers of EV are purchasing this type of vehicle for the Falk Albers
first time. Several studies show that it takes some time to Senior Executive Partner,
adapt to a new type of vehicle. This applies in particular to Global Leader Mobility,
EV, whose acceleration capabilities are significantly higher Insurance Consulting
− The geographical spread of EV is skewed towards urban
areas, more prone to frequent traffic collisions
− The breakdown of EV drivers by age is also skewed towards
younger, less experienced drivers compared to ICE drivers,
i.e. drivers that are more prone to accidents
− Several cases of fraudulent commercial usage (ride-hailing,
food delivery, etc.) have been identified in various geogra-
phies, as the lower cost of usage of EV can incentivize
insured persons to use their vehicles commercially.

The global consensus is that the average cost of claims for EV We have seen a lot of media coverage about EV fires, creating
is higher than for ICE. Studies have shown that, besides the impression that EV are more prone to fire than ICE. How-
EV-specific parts (battery, high-voltage cabinet, etc.), non-EV ever, a global consensus is emerging that EV do not pose an
specific parts (such as a bumper or a light) are usually more increased fire risk compared to ICE, and statistics actually
expensive for EV than for ICE. This, combined with the need tend to show a lower fire risk for EV compared to ICE.
for specialized training for the workforce and the additional
complexity of some repairs, contributes to an overall increase Still, EV fires are much more complicated to extinguish due
in the cost of claims. to their intrinsic nature and the huge amount of energy stored
in the battery. Specific procedures and additional resources
are required to properly extinguish an EV fire and prevent its
reignition.

The rise of EV is also directly connected with an increase in the

level of vehicle connectiveness and the importance of onboard
software. As a result, vehicles are exposed to new cyber risks
and are vulnerable to attacks.

This aspect is still being investigated across the (re)insurance

industry given the limited current number of observations.
Nevertheless, the exponential rise in what are now being called
software-defined vehicles will push further studies in this
direction and drive the development of adequate risk assess-
ment.
Munich Re Page 91/111
Tech Trend Radar 2025

Redefining Industries

Hydrogen Economy
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Trial & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Hydrogen has the potential to redefine and decarbonize the energy − Hydrogen infrastructure and storage capacities are essential for wide-
system. spread adoption.
− Rapid growth in production capacities is required to meet the projected − Diverse risks and uncertainties, coupled with political and regulatory
demand. challenges, are hindering hydrogen bankability, the growth of clean
− Innovative risk transfer solutions can help facilitate the transition. hydrogen production, and systematic deployment.

Overview Hydrogen and its potential

The energy sector is seen as a major contributor to global for the energy transition
emissions, with ever-increasing efforts to transition toward
lower emission intensity. Alongside renewable energies and The need to adopt measures to limit the severity of climate
direct electrification approaches, sustainable energy carriers change has reached an unprecedented level. At the same time,
are expected to play a vital role. Among them, hydrogen is a political developments are strengthening the desire for local
versatile option that can be produced with high efficiency and energy security and less reliance on fossil fuels. In this context,
power for a wide range of applications, combining different it is commonly accepted that hydrogen represents an impor-
sectors as part of an envisioned Hydrogen Economy. However, tant pillar for a successful energy transition.
leveraging the Hydrogen Economy’s decarbonization potential
will require overcoming significant hurdles to the deployment The Hydrogen Economy refers to an industrial and economic
of clean hydrogen projects. system design in which hydrogen serves as one of the main
energy carriers for various sectors. With its multiple applica-
tions – direct use, combustion, re-electrification in fuel cells
or long-term storage – hydrogen has the potential to power
carbon-intensive industries, buildings and heavy-duty trans-
portation, as well as to balance out seasonal fluctuations in
renewable electricity production.

Furthermore, hydrogen is essential to the production of various

chemicals and fuels, and can even be employed to decarbonize
steel and cement production. Importantly, however, in order
for a Hydrogen Economy to comply with the 2050 emission
targets, hydrogen must be produced via a very low emission-
intensity pathway.

Clean hydrogen is primarily produced from renewable elec-

tricity such as solar or wind power using an electrochemical
process called electrolysis. The resulting “green hydrogen”
originates from water, thus contributing to a decreased
reliance on fossil fuels.
Munich Re Page 92/111
Tech Trend Radar 2025

“Sustainable energy carriers associated with low emis-

sion intensity are vital for a successful energy transition.
In this context, hydrogen produced from renewables
will play a key role, but its deployment is currently ham-
pered by a variety of uncertainty factors. Well-thought-
out risk mitigation and transfer solutions can pave the
way for investment security and a rapid upscaling of
green hydrogen technologies.”

Dr. Kathrin Ebner

Technology Lead Hydrogen & E-Mobility

Another pathway to low-emission-intensity hydrogen produc- release of the harmful by-products of combusting carbon-
tion involves integrating carbon capture utilization and storage based energy carriers (carbon dioxide, soot, etc.). To leverage
(CCUS) technologies into the Steam Methane Reforming the resulting overall emission reduction potential of clean
(SMR) process for extracting hydrogen from natural gas (“blue hydrogen, its production would have to be rapidly upscaled
hydrogen”). This aims at preventing the release of the CO2 alongside corresponding infrastructure. To put that into
emissions produced when the methane molecule is split into numbers, in 2023 global hydrogen demand reached 97 Mt,
the atmosphere. So, while still of fossil origin, blue hydrogen’s mostly derived from methane without any carbon capture
carbon footprint is favorable when compared to the direct use (“gray hydrogen”). By 2050, hydrogen production is expected
of fossil methane as a fuel. With that in mind, it is often to reach 500 Mt, with clean hydrogen production accounting
referred to as a “bridging technology.” As a product of hydro- for the vast majority.
gen use, only water and heat are formed, eliminating the

Key Findings
For hydrogen to fulfill its planned role as a major contributor
to global emission reduction, rapid deployment of production
capacities and upscaling of infrastructure are required.
Munich Re Page 93/111
Tech Trend Radar 2025

Currently, however, clean hydrogen deployment is facing many investment in the clean energy industry in the period leading
challenges. On the one hand, market volatility, regulatory up to 2030 to achieve net-zero targets by 2050 will need to be
changes and geopolitical uncertainties affect the feasibility of at around US$ 4 trillion.
hydrogen projects. On the other, there is a limited understand-
ing of the technical risks associated with the technologies due A combination of financial, institutional and regulatory instru-
to limited long-term experience and data availability. ments will be needed to de-risk hydrogen projects, including
suitable policy frameworks to reduce uncertainty and encour-
Uncertainties regarding component failure and underper- age investment. In the context of de-risking, the insurance
formance, unexpected maintenance and repair costs or system industry is a key stakeholder offering a variety of risk transfer
downtime, contribute to hesitant capital investments and lim- solutions. This allows even less fully understood technical
ited bankability of electrolysis plants. Accelerating the energy risks to be addressed, helping to overcome the reluctance
transition will require enormous amounts of capital – the Inter- towards less mature technological approaches like electrolytic
national Energy Agency (IEA) estimates that annual global production.

How will the Hydrogen Economy change over the next year?

The Hydrogen Economy is set for significant growth over the next 12 months. Key drivers
2 Questions for include increasing low-carbon hydrogen projects, technological advances in electrolyzers,
and increased global investment. Europe is leading the way with strong initiatives, while
Patrick Hinze US policy uncertainty could slow progress. Production costs are expected to drop, making
hydrogen more competitive. Infrastructure development, such as the Hamburg Green
Hydrogen Hub, is crucial. Despite challenges like cost and policy hurdles, hydrogen
remains central to global energy transition efforts and corporate strategies.

Why is insurance so important to boosting hydrogen production?

Adequate risk management is vital for hydrogen projects, as it mitigates the financial vol-
atility associated with technology underperformance. It insures investors and stakehold-
ers against potential losses, enhancing project bankability. By providing confidence in the
reliability of hydrogen technologies and operations, performance insurance attracts capi-
tal, accelerates development, and supports the scaling of this emerging energy sector.

Patrick Hinze
Head of Emerging Green Technologies
Munich Re Page 94/111
Tech Trend Radar 2025

Redefining Industries

Humanoid Robots
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Assess & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Increase in operational efficiencies and quality − Modelling risks

− Improved underwriting and casualty risks − Product and service offering uncertainty
− New market for robot insurance products − Cyber security and privacy concerns
− Reduction in claims costs − Safety implications
− Autonomous handling of repetitive tasks or tasks that are dangerous − Ethical dilemmas
for humans

Overview A new era of human and machine

The robotics revolution is here, reshaping industries at an In the not-so-distant past, the term “robotics” conjured up
unprecedented pace, and the insurance industry is no excep- images of manufacturing floors and assembly lines. Today, the
tion. Robots, automation, and artificial intelligence are not field of robotics has transcended its industrial roots, trans-
only redefining traditional processes but also creating entirely forming into a ubiquitous presence that touches nearly every
new paradigms of risk, coverage, and customer interaction. aspect of our daily lives. While robots have become an integral
Equipped with advanced AI and sensors, humanoid robots will part of our lives, their capabilities and applications have
enhance efficiency, reduce costs, and provide real-time data expanded exponentially in recent years, promising to revolu-
for better decision-making, leading to higher process effi- tionize the way we live, work, and interact with each other.
ciency and more personalized services. Yet insurers will face
new challenges, namely identifying what kind of robotic tech-
nologies will make a difference with regard to risk and the
Driving forces of the robotic revolution
required coverage offerings. The robotic revolution is being driven by the convergence of
several key technologies, including advancements in increas-
ingly sophisticated sensors, actuators, power systems and
artificial intelligence, amplifying their potential for multifunc-
tional capabilities and collaboration with humans. The inter-
section of humans and robotics is a hallmark, with distinctions
between the physical and digital realms rapidly disappearing
across industrial domains from healthcare to services and
agriculture.

The emergence of household robots and

disruptive breakthroughs in humanoids
Once considered as purpose-oriented niche luxury items,
household robots have transitioned to become household
staples, fundamentally reshaping people’s everyday lives.
From AI-enabled personal virtual assistants to automated
laundry-folding robots, they are operating in hyper-connected,
automated ecosystems, for example to optimize energy use,
monitor properties and automate repetitive tasks.
Munich Re Page 95/111
Tech Trend Radar 2025

While the “ChatGPT“ moment for robots was expected years

ago, tectonic shifts may appear in 2025, transitioning the
purpose-bound robotics we are familiar with to general-
purpose robots, with recent breakthroughs in Generative AI
to emerge in Humanoids.

A combination of factors has made Humanoid robots smarter,

cheaper and more human. The cost of components and
production has decreased, with Unitree, for example, readying
its Humanoid “G1” for mass production with a tag of $16.000.
Developments in neural networks and machine learning algo-
rithms have empowered Humanoids to learn and adapt to new
situations, and achieve heightened performance and efficiency
as they interact with their environments in a manner that
simulates human-like intelligence. Breakthroughs in computer
vision have expanded robotic capabilities in areas such as
object recognition, tracking, and manipulation.

They mirror the human form and functionality and can inte- Surge in collaborative robots (“Cobots”)
grate seamlessly into our physical and social environments.
The evolution of humanoid robots has traversed a significant “Cobots” are increasingly becoming an integral part of the
spectrum, from rudimentary mechanical prototypes to modern workplace. Unlike traditional industrial robots, which
sophisticated, autonomous entities. Moreover, these AI-driven operate in isolated zones, cobots are designed specifically
systems enable humanoid robots to decipher and respond to work alongside humans, augmenting productivity while
to human gestures and emotions with precision, effectively ensuring enhanced safety. Industries like manufacturing, retail,
bridging the divide between humans and machines. Pioneer- and healthcare are integrating cobots to streamline operations.
ing companies including Tesla (Optimus), Boston Dynamics For example in logistics, cobots expedite the sorting of goods
(Atlas), NVIDIA (GR00T), 1X (Neo), and Figure (Figure 02), while minimizing errors. In healthcare, robotic systems enable
as well as Chinese manufacturers like Unitree (G1), are at surgeons to perform high-precision procedures, and support
the forefront of harnessing the transformative potential of eldercare by enhancing mobility and monitoring capabilities.
humanoid robots, which promises to redefine the very fabric
of industry ecosystems and imbue machines with new levels
of autonomy and human-like intelligence. Expansion of autonomous
vehicles and drones
Autonomous technologies powered by advanced neural net-
works and machine learning algorithms continue to evolve,
with autonomous vehicles, autonomous mobile robots (AMRs)
and drones achieving greater efficiency and reliability. Drones
and AMRs are being deployed for last-mile and warehouse
logistics, while autonomous trucks enable cost reductions and
address labor shortages. These technologies bring unique
risks that insurers need to consider. For example, cybersecurity
threats loom large, with the potential for hackers to disrupt or
control autonomous systems. Cyber and product liabilities,
as well as vehicle coverage can become a focus.

In today’s robotics era, the most critical factor for

organizational success is not the adoption of new
technologies, but the development of new capa-
bilities, skills, and mindsets.

Andreas Schumacher
Project Manager Artificial Intelligence
Munich Re Page 96/111
Tech Trend Radar 2025

Robotics in healthcare Impacts of the robotic revolution

Robots are playing an increasingly crucial role in healthcare,
for insurers
encompassing services such as diagnostics, surgical assis- Robots blur traditional definitions of liability. If a robot injures
tance, and patient care. This can influence underwriting, as someone or causes damage, determining fault is increasingly
medical malpractice must adapt to include potential failures of complex. Insurers must design products that address shared
robotic systems. If robotic healthcare solutions are proven to liability among manufacturers, software developers, and oper-
reduce human errors and improve outcomes, long-term ators. Connected robots in homes, healthcare, and industrial
healthcare costs may decrease. workplaces are vulnerable to hacking. Cyberattacks could lead
to physical harm, financial losses, or data breaches. Insurers
need to expand physical cyber coverage to protect against
Technological shifts in agriculture these risks. Robots generate vast amounts of real-time opera-
The agricultural sector is experiencing a seismic shift with the tional data. Insurers can leverage this data to refine risk mod-
emergence of robotics, as autonomous tractors, drone crop els, enabling dynamic pricing. They can leverage data analytics
monitors, and robotic harvesters reshape the farming land- and AI to predict risks and offer personalized insurance solu-
scape, making it more efficient and environmentally friendly. tions.
Farmers can optimize resources, reduce waste, and increase
yields. As these innovations amplify productivity, they also
usher in a new wave of risks, including mechanical break-
Conclusion
downs, programming glitches, and dependence on GPS and By 2025, humanoid physical robots will be revolutionizing
connectivity, which agricultural insurers must adapt to by the insurance industry. Their use in claims assessments, risk
developing novel products and policy bundles. management, customer service, and fraud detection will
significantly increase operational efficiency, reduce costs, and
enhance the customer experience. Insurance companies that
embrace these technologies early on will gain a competitive
edge by offering faster, more personalized services and
improving their data-driven decision-making processes.

Key Findings
“More than 250,000 humanoid robots are anticipated to be
shipped worldwide by the end of 2030.”
Source: Goldman Sachs
Munich Re Page 97/111
Tech Trend Radar 2025

Redefining Industries

Autonomous Mobility
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Assess & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− AVs have the potential to significantly reduce accidents by eliminating − Liability may shift from drivers to manufacturers, potentially disrupting
human error. the traditional personal auto insurance model.
− The emergence of AVs creates opportunities for insurers to develop − The timeline for mass adoption of Level 5 AVs remains uncertain, mak-
new products and services tailored to the unique risks associated with ing it challenging for insurers to accurately predict the impact on their
this technology. business.

Overview Levels of autonomous driving

The automotive industry is undergoing a profound transforma- To understand the impact of Level 5 AVs, it is essential to
tion with the emergence of Level 5 Autonomous Vehicles grasp the different levels of autonomous driving defined by the
(AVs), often called driverless cars. AVs promise numerous ben- Society of Automotive Engineers (SAE). These levels range
efits, such as reduced traffic congestion, increased safety, from 0 (no automation) to 5 (full automation), with each level
energy conservation, and independent mobility for those una- representing an increasing degree of vehicle autonomy.
ble to drive themselves. In short, AVs have the potential to rev- Level 5 represents the ultimate goal, where the vehicle can
olutionize not only how we travel but also various industries. perform all driving tasks under all conditions without any
As we explore the potential impact of Level 5 AVs on the insur- human intervention.
ance landscape, we examine projected market penetration,
develop hypotheses about their effects, and analyze how insur-
ance models might adapt to this disruptive technology. Projected market penetration of Level 5
Autonomous Vehicles
Considering the emerging landscape of Level 5 Autonomous
Vehicles (AVs), the precise timeline for mass adoption remains
uncertain. However, projections suggest that Level 5 auton-
omy is unlikely to be broadly available before 2035, according
to S&P Global Mobility, a leading automotive research firm.
However, the transition to Level 5 autonomy is expected to be
gradual, with increasing levels of automation becoming more
prevalent in the coming years.
Munich Re Page 98/111
Tech Trend Radar 2025

Hypotheses on the impact of Level 5

AVs on the insurance Industry
There are certain commonly held beliefs when it comes to the
impact of AVs on the insurance industry. We propose three
hypotheses to explore:

Hypothesis 1: Decreased accident frequency

This hypothesis is based upon the assumption that wide-
spread adoption of AVs has the potential to significantly
reduce accidents by eliminating human error, a major contrib-
uting factor to crashes. Some calculations suggest that acci-
dent rates for AVs could be 50 times lower than for conven-
tional cars. However, it is important to acknowledge that the
data is still limited and is based primarily on controlled testing
environments. As the technology continues to evolve and
more self-driving cars hit the road, there is still a need for con-
tinuous monitoring and improvements to maintain this low
accident rate.

Hypothesis 2: Shift in insurance liability

This hypothesis is supported by the notion that liability will
shift from drivers to manufacturers as Level 5 AVs emerge.
Widespread AV adoption will likely result in car insurance
being replaced by product liability insurance, ultimately borne
by auto manufacturers. This is because Level 5 AVs are
designed to operate without human intervention, making the
manufacturer responsible for the vehicle’s actions.

Hypothesis 3: Emergence of new insurance products

This hypothesis is supported by observations of the evolving
insurance landscape and anticipation of new risks associated
with AVs. There will likely be a need for new product develop-
ment to address emerging autonomy risks which may include
coverage for cybersecurity threats, software malfunctions,
and product liability.

These hypotheses were chosen because they address key Conclusion

areas of potential disruption in the insurance industry related
to the introduction of Level 5 AVs. Understanding the potential Level 5 Autonomous Vehicles are poised to revolutionize the
changes in accident frequency, liability, and the types of insur- automotive industry and bring about significant changes in the
ance products needed will be crucial for insurers to adapt insurance landscape. While mass adoption may still be some
and thrive in this evolving landscape. years away, the transition to higher levels of automation is
already underway. The insurance industry is well advised to
proactively adapt to this technological shift by developing new
risk assessment models, tailoring insurance products, and col-
laborating with stakeholders in the AV ecosystem. By embrac-
ing innovation and addressing the unique challenges and
opportunities presented by Level 5 AVs, the insurance industry
can contribute to the safe and responsible integration of this
transformative technology into society.
Munich Re Page 99/111
Tech Trend Radar 2025

Redefining Industries

Climate Resilience
Impact high medium low

Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

Adopt & Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Opportunities Risks

− Prepare for climate impacts, reduce potential losses and safeguard − Complexity of climate models, data gaps and scope of analysis can be
long-term resilience challenging
− Comply with emerging regulations such as EIOPA and PRA guidelines − Identifying climate hazards without estimating financial materiality
− Drive innovation by identifying new products, services, or business can be ineffectual
models for climate change − Failing to take appropriate action to mitigate or adapt to identified risks
could result in long-term harm

Overview Assessing physical climate risks

In 2024, natural disasters resulted in US$ 320 billion in losses, for a resilient future
underscoring the growing impact of climate risk on lives, econ-
omies and assets. To address these challenges, organizations, In 2024, natural disasters caused overall losses of US$ 320
governments and communities must assess vulnerabilities billion, with a clear upward trend in recent years, in particular
and prepare for rising threats, including extreme weather for insured losses. As climate threats intensify, organizations,
events and chronic stressors like sea-level rise. Climate risk governments and communities must address mounting risks
assessments involve three steps: understanding exposure, to lives, assets and economies. The growing frequency and
measuring financial impacts, and managing material risks. intensity of extreme weather and chronic stressors like rising
Sectors like finance, real estate, energy, and government are sea levels highlight the need to assess vulnerabilities and
increasingly adopting these assessments. Conducting holistic prepare effectively. Tools like Munich RE’s Location Risk Intel-
and consistent climate risk assessments is crucial to main- ligence platform offer vital insights into physical risks and
taining long-term resilience. future scenarios, enabling proactive measures to build resil-
ience.

Climate risk refers to the potential adverse effects of climate

change on natural and human systems. Physical risks arise
from direct climate phenomena, such as hurricanes, floods,
droughts, heatwaves, and rising sea levels. These hazards can
damage production sites and infrastructure, disrupt supply
chains, and threaten ecosystems and communities. Physical
risks are often location-specific, requiring detailed analysis to
identify and manage exposures.

Conducting a climate risk assessment involves three steps:

understanding, measuring, and managing. The “understand”
step consists of analyzing the exposure of different areas or
assets to various climate-related risks and how the risk
changes over time in different climate scenarios. This can
Key Findings already give an indication as to which risks, geographies, or
The increasing financial impacts of climate risks underscore assets are at low, medium or high risk today and in the future –
the urgent need for climate risk assessments and resilience and provide guidance on how to refine the next steps in the
strategies. assessment process.
Munich Re Page 100/111
Tech Trend Radar 2025

“Climate risk and resilience Restoring ecosystems also provides natural buffers against
climate impacts. In addition to physical adaptations, opera-
will be a key topic for the
tional changes, such as modifying supply chain logistics or
insurance industry going adjusting agricultural practices, can help organizations align
forward – not only for with emerging climate realities. Risk transfer strategies can
adequate pricing and risk also help manage the financial impacts of climate events
through products like parametric insurance and catastrophe
management but also to
bonds.
develop new products and
superior data & analytics.” Certain sectors are particularly proactive in conducting
climate risk assessments due to their operational needs and
David Fischer exposure to climate impacts. These include:
Senior Product Manager,
Climate Risk − Financial sector: Banks, insurers, and asset managers use
risk assessments to evaluate investment and underwriting
vulnerabilities and opportunities, and to comply with disclo-
sure mandates.
− Real estate: Developers and property managers analyze
The second step (“measure”) then involves estimating the location risks to ensure the durability and long-term value of
potential financial impact of relevant risks in terms of a loss their properties.
metric by taking into account the vulnerability of exposed − Agriculture: Farmers and agribusinesses assess climate
systems. This helps to determine whether a risk affecting a patterns to protect crop yields and adapt farming practices
given asset or concentration of assets poses a financially to changing weather conditions.
material risk to a business or community. Munich Re’s Loca- − Energy: Power and industrial companies evaluate risks to
tion Risk Intelligence platform can provide comprehensive critical infrastructure, such as power grids and renewable
support throughout both of these steps. energy systems, to maintain reliable energy supplies.
− Government: Authorities and planners use risk assessments
The third step (“manage”) in a comprehensive approach to to inform disaster preparedness, infrastructure development,
climate risk involves the definition of, and decision on, the and resilient urban design.
management of financially material climate risks. The results − Logistics and supply chain: Companies analyze transporta-
of the first two steps provide a foundation for a range of tion networks and supplier locations to mitigate disruptions.
actions aimed at reducing exposure and building resilience.
Developing a risk mitigation strategy and adaptation measures Across all these sectors, tools like Location Risk Intelligence
is among the most critical responses. These measures can provide accurate physical risk assessments, scenario model-
include investments in resilient infrastructure, such as flood ling, and data-driven insights, enabling stakeholders to make
barriers and heat-resistant buildings. informed decisions and bolster their resilience.

Understand Measure Manage

− Analyze the exposure of a − Quantify the severity of different − Define and decide on risk
portfolio or a single location hazards (e.g. flood depths [m], mitigation strategies (accept –
to key natural disasters under Max 1-Day precipitation [mm], avoid – adapt – transfer)
the present climate Drought Lengths [days] etc.) − Engage in risk transfer solutions
− Understand the growth in risk − Quantify the financial impact of to respond to climate and nature
expected with climate change climate risks for a single location risks (e.g. parametric insurance,
− Identify risk concentrations or whole portfolios catastrophe bonds)
Munich Re Page 101/111
Tech Trend Radar 2025

Impact on value chain (1/2)

How technology impacts the insurance business

Trends Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

& Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Data & AI
Spatial
hold
Intelligence
Artificial
General hold
Intelligence
Insurance
API Stand- adopt
ardization
Quantum
assess
Computing
Synthetic
hold
Data
AI Democra-
adopt
tization
AI
adopt
Governance
Generative
adopt
AI

AI Agents trial

AI-Aug-
mented
trial
Software
Engineering
AI Search
adopt
Engines

Legal Tech trial

Healthy
Human
Personalized
trial
Medicine

AI Medicine trial

Digital
adopt
Healthcare
Behavioral
adopt
Analytics

high medium low

Munich Re Page 102/111
Tech Trend Radar 2025

Impact on value chain (2/2)

How technology impacts the insurance business

Trends Maturity Insurance Value Chain

Product Design Sales & Risk Customer Claims Data &

& Pricing Distribution Underwriting Policy Admin Management Engagement (incl. Reserving) Analytics

Connected
Experience

Metaverse hold

Ubiquitous
assess
Connectivity
Ambient
trial
Computing
Extended
trial
Reality

Edge AI trial

Autonomous
trial
Interactions

Cyber &
Crypto
Decentral-
ized Data assess
Economy
Automated
assess
Compliance
Digital
trial
Assets
Deepfake
trial
Defense
Non-Human
adopt
Identity
Digital
Immune assess
System

Redefinding
Industries
Crop
adopt
Intelligence
New Energy
adopt
at Scale
Trade Secret
trial
Protection
Electric
adopt
Vehicles
Hydrogen
trial
Economy
Humanoid
assess
Robots
Autonomous
assess
Mobility
Climate
adopt
Resilience

high medium low

Munich Re Page 103/111
Tech Trend Radar 2025

Our solutions

Data & AI

aiSure
aiSure. provides insurance coverage for losses resulting from
errors made by AI systems, including incorrect outputs or
flawed predictions.

Insure AI

Location Risk Intelligence

Location Risk Intelligence Platform is a modular SaaS plat-
form that enables our clients to understand, measure, and
manage risks from natural hazards and climate change on a
global scale. In underwriting, risk management, and portfolio
steering, a detailed assessment of exposure is necessary
due to the increasing threat of natural disasters.

The solution for the insurance industry

UW Monitoring Tool
The Portfolio Monitoring Tool sets a new standard for informed
decision-making, providing our clients with a comprehensive
view of their portfolio. Offering truly customizable, in-depth
analyses that go beyond conventional financial reporting, the
tool delivers actionable insights through features like cohort
retention and rate change analyses. By seamlessly integrating
Munich Re's data assets and GenAI capabilities, clients can
enhance the quality and business value of their own data.

Reinsurance Property/Casualty
Munich Re Page 104/111
Tech Trend Radar 2025

Our solutions

Healthy Human

alitheia iRisk
The transformative risk assessment engine and instant deci- Our next-generation risk assessor and selector for primary
sioning system is designed to help life insurance carriers and insurers and reinsurance clients integrates Munich Re under-
distributors efficiently scale, enhance the customer experi- writing rules, algorithms, and risk scoring based on external
ence, and manage risk – enabling them to grow with confi- data sources to provide more accurate risk selection at the
dence. application stage.

alitheia Munich Re Digital Solutions

Automated EHR Summarizer CLARA plus

The automated EHR Summarizer simplifies access to the EHR Our solution for disability calculation allows insurers to auto-
data needed for life insurance underwriting. By normalizing matically calculate the degree of disability, leading to faster
EHR data directly from the source, it highlights core data claims decisions and higher reactivation rates with lower
points identified as essential by underwriters and provides administration costs.
triage guidance.
CLARA plus
Automated EHR Summarizer

Clareto MIRA PoS

Clareto offers life insurance carriers and distributors access Our automated underwriting solution empowers life insurers
to the largest interoperability network for risk assessment. to make underwriting decisions quickly. The solution is based
This network includes fully implemented, expertly managed, on Munich Re's underwriting guidelines, regularly updated to
and continuously optimized connections to electronic health reflect the latest risk research and market trends.
records (EHRs), health information exchanges (HIEs), and
other digital health data sources - covering over 70% of the MIRA PoS
U.S. population.

Clareto

ALLFINANZ SPARK/NOVA MIRApply Insured

Our automated underwriting platform provides life and health Our web-based applicant interview solution streamlines the
insurers with a flexible rule set that allows them to change application process for life insurers by providing underwriters
rules, reflecting their unique underwriting philosophy and with reliable risk assessment and a risk-rating proposal based
business strategy, as well as to integrate third-party data on our constantly updated underwriting guidelines, making
through data exchange. the process faster and more accurate while optimizing the
customer experience.
Automated underwriting solutions
MIRApply Insured
Munich Re Page 105/111
Tech Trend Radar 2025

Our solutions

Healthy Human

Next Generation AI-Augmented SHIELD

Underwriting Our claims system is designed to allow health insurers to
automate their medical claims adjudication, detect potential
Our latest underwriting solution for life and health insurers fraud, waste, and abuse, and justify denials with the right
combines deep domain knowledge, responsible AI, data, medical reasoning.
modern technology, and risk acceptance to assess risks at the
point of sale. SHIELD

Advance with confidence

Predictive Underwriting Solution SMAART

(Predictor)
SMAART is a comprehensive claims reserving, portfolio
Predictor is an additional module leveraging predictive tech- monitoring, and pricing system for health insurance busi-
nologies to enhance underwriting processes with an increased nesses, providing a 360-degree view of performance and ena-
processing rate, reduced evidence collection, and improved bling the management of portfolio profitability. It offers tools
customer experience. to manage health business reserves, monitor policy and prod-
uct performance, handle renewals, and inform pricing deci-
AI for insurance​ sions, all while considering factors like seasonality and medical
trends to make accurate projections. SMAART is leveraging
a user-friendly experience and robust, actuarially-sound
methodology.

SMAART
Munich Re Page 106/111
Tech Trend Radar 2025

Our solutions

Connected Experience

Automated Underwriting Platform Cert2Go

for facultative & corporate business Cert2go's fully digital and automated underwriting process
enables users to efficiently reinsure small to medium-sized
Munich Re F&C´s Automated Underwriting Platform bundles single risks, accelerating workflows and streamlining adminis-
all digital insurance solutions for business with facultative trative workload. Currently, cert2go is available for all property
risks as well as corporate insurance coverage. lines of business.


Automated Underwriting Platform cert2go

IoT Sensor Solutions REALYTIX ZERO

Insurance companies are looking for new ways to use tech- REALYTIX ZERO is a cloud-based, automated underwriting
nology, such as Internet of Things (IoT) sensors, to improve platform which helps launch, distribute, and scale new or
their operations, elevate their customer experience, and create existing insurance products. With REALYTIX ZERO, you can
new coverage opportunities. Munich Re's Sensor Solutions digitize, automate, and launch any insurance product in a
by HSB® is an IoT solution that helps insurers implement sen- matter of days and adjust it in real time using our no-code/
sor-based IoT programs to address common risks like water, low-code approach and API integration. Our new embedded
fire, and equipment failure for commercial and personal lines Python capabilities now enable us to seamlessly integrate
insurers, along with their policyholders. even complex pricing logic in a remarkably short time frame.

Internet of Things (IoT) REALYTIX ZERO

Munich Re Page 107/111
Tech Trend Radar 2025

Our solutions

Cyber & Crypto

In today's rapidly changing cyber landscape, effective risk Squalify

management is more crucial than ever for our clients. By
bringing together expert knowledge from insurance, reinsur- As a wholly-owned subsidiary of Munich Re, Squalify
ance, and external partners, we can provide robust risk assess- provides a software as a service platform that helps companies
ments and accumulation modelling that help make cyber risks quantify and manage cyber risk. It uses a battle-tested model,
insurable. This enables us to support our clients in mitigating licensed from Munich Re, to provide realistic financial loss
the consequences of cyberattacks and protecting their metrics and transparency on cyber exposure. Squalify has
businesses from potential harm. With a deep understanding access to industry-leading data from over 1 million cyber
of cyber risks and a disciplined approach to market growth, claims, combined with proprietary algorithms, allowing for
guided by a clearly defined risk appetite, we strive to deliver accuracy and scalability. By transforming complex cyber risk
tailored solutions that meet the unique needs of our clients. data into clear, financial insights, Squalify enables decision-
Please also refer to our newly published report Cyber Insur- makers to evaluate cyber risk, prioritize investments, and align
ance: Risk and Trends 2025, which provides an outlook on cybersecurity strategies with business objectives.
the cyber risk landscape and the surrounding dynamics affect-
ing cyber insurance and market demand. The report presents Cyber Risk Quantification for the Boardroom
insights, facts and figures which we want to share with our
clients.

Digital Asset Protection Home Cyber Protection HSB

As the blockchain ecosystem continues to evolve, new risks Homeowners are becoming increasingly vulnerable to cyber
and challenges are emerging for businesses and investors threats as their personal and professional lives become more
operating in the crypto economy. The rise of digital assets and connected. With over 15 years of expertise in cybersecurity,
decentralized applications has created a growing need for HSB offers products and services to help homeowners
specialized insurance solutions that can provide security, mitigate these risks, respond to breaches, and recover from
reliability, and stability. Our Digital Asset Protection Solutions cyber crimes, providing them with the necessary knowledge
aim to address these risks, leveraging Munich Re's expertise and support to protect themselves in a rapidly evolving
in risk management and insurance to help safeguard digital digital landscape.
assets and provide confidence for companies and investors
to operate effectively in this innovative space. HSB Home Cyber Protection


Digital Asset Protection
Munich Re Page 108/111
Tech Trend Radar 2025

Our solutions

Redefining Industries

Epidemic and Pandemic Risk Solutions FIVE

Epidemic and Pandemic Risk Solutions provide comprehen- FIVE provides investment indices for savings and retirement
sive risk transfer and investment solutions to protect against products, accessible via reinsurance with Munich Re Markets.
catastrophic disease outbreaks. This offering addresses the It offers a diversified, rules-based approach to manage
insurance gap for non-damage business interruptions caused risk and potential returns. FIVE supports insurers by bundling
by epidemics. It includes coverage for business interruption, investment and guarantee components, ensuring cost-
temporary site closure, and other specific epidemic-related efficiency and robust product expectations. It provides access
risks. The solutions are tailored for various industries, includ- to a broad investment spectrum and can be implemented
ing hospitality, manufacturing, retail, mining, and construction, efficiently in liquid markets. FIVE's strategies are designed to
leveraging proprietary models and strategic partnerships to be transparent, and aligned with client performance goals.
ensure robust risk management.
FIVE (Login required)
Epidemic and pandemic risk solutions

Electric Vehicles Analytics (EVA) Infrastructure Risk Profiler

Electric Vehicles Analytics (EVA) leverages advanced machine Infrastructure Risk Profiler is a comprehensive service
learning techniques to provide comprehensive risk assess- designed to enhance the performance of infrastructure invest-
ment and pricing solutions for electric vehicles. EVA enables ment portfolios. It provides a detailed analysis of six key risk
insurers to more accurately estimate risks, define risk appetite, dimensions: macroeconomics, technology, natural hazards,
and confidently underwrite emerging exposures. It includes execution, operations, and market conditions. This approach
detailed statistical analysis to identify and classify EVs based allows investors to make informed, reliable decisions by under-
on technical specifications, helping insurers to optimize profit- standing the overall risk profile of their projects. The service
ability and reduce anti-selection. EVA supports insurers in aims to save time and money while boosting investor confi-
adapting to market standards and local regulations, ensuring dence through transparent and thorough risk assessments.
robust risk management and competitive pricing.
Infrastructure Risk Profiler
Insurance Solutions for Electric Vehicles
Munich Re Page 109/111
Tech Trend Radar 2025

Our solutions

Redefining Industries

HySure
HySure is a risk transfer solution designed to support the
low-carbon hydrogen industry. It provides product and perfor-
mance guarantees for electrolyzer manufacturers, helping
to mitigate financial risks and enhance investor confidence.
HySure covers high repair costs, underperformance, and avail-
ability issues, helping to manage financial risks and increase
confidence in the market.

Securing the Power of Green Hydrogen

Insight
Insight is an analytics tool designed to optimize the automa-
tion process and underwriting performance. It provides
detailed insights into underwriter workloads, average under-
writing duration, trends, and rules impacting referrals. By
leveraging this data, insurers can enhance efficiency, reduce
costs, and improve decision-making processes.

Underwriting analytics

Talaria
Talaria is an embedded finance solution designed to stream-
line trade finance for small and medium-sized enterprises
(SMEs). It integrates seamlessly into B2B platforms, offering
invoice-to-cash solutions that enhance liquidity and cash flow
management. Talaria leverages Munich Re's expertise in data
analytics and machine learning to provide risk-intelligent
pricing and credit underwriting. This solution simplifies the
administrative burden of trade finance, making it more acces-
sible and efficient for SMEs.

Talaria Solutions
Munich Re Page 110/111
Tech Trend Radar 2025

Team
Martin Thormählen Andreas Hofheinz
CTO Business Technology – Editor, Munich Re Consultant Data Analytics, Munich Re

Daniel Grothues Thilo Horner

Chief Architect Primary Insurance, ITERGO Senior Product Manager, Insurance Solutions, Munich Re

Michal Nejthardt Thibault Imber

Head of Digital Solutions South Africa and SSA, Principal, Insurance Consulting @ Digital Solutions,
Munich Re Africa Branch Munich Re

Daniela Schuler Dr. Matthias Kaper

Enterprise Architect, ERGO Group Senior Data Scientist, Munich Re

Monique Ferraro Milan Kästner

Counsel Cyber Practice, HSB Risk Analyst, Munich Re

Michael Hild Tim Klein

Foresight Passionate, Munich Re Executive Assistant, ERGO Group

Christian Hobelsberger Dr. Andre Knoerchen

Integrated Student (Statistics and Data Science), Head of New Tech Underwriting, Munich Re
Munich Re
Divashan Moodley
Actuarial Analyst, Digital Solutions,
Munich Re Africa Branch

Bernhard Müller
Underwriting Manager Power & Utilities, Munich Re

Neil Munro

Contributors
Principal, Insurance Consulting @ Digital Solutions,
Munich Re

Matthias Nawrocki
Head of Metaverse, ERGO Group AG
Dr. Tobias Aigner
Senior Project Manager, New Energy, Munich Re Dr. Andreas Nawroth
Leading Expert AI and Quantum, Munich Re
Matthias Beuerle-Liegel
Specialist Data Analytics, Munich Re Dr. Alina Nizamutdinova
Project Manager Artificial Intelligence, Munich Re
Dr. Pamela Chetty
Chief Medical Officer for South Africa and SSA, Georg Opora
Munich Re Africa Branch Senior Enterprise Architect, ERGO

Florian Eberl Luisa-Marie Schmolke

Technical Architect, Munich Re Innovation Developer, ERGO Innovation Lab

Dr. Kathrin Ebner Andreas Schumacher

Risk Analyst, Munich Re Project Manager Artificial Intelligence, Munich Re

Luiz Leonardo Leite Filho Norman Thoms

Agro UW, Munich Re Vice President, IoT Technologies, HSB

David Fischer Zachary Toner

Senior Product Manager, Climate Risk, Munich Re Head of Section Workflow & RPA, Munich Re

Norbert Hackner Halyna Vasylevska

Senior Consultant, Insurance Consulting, Senior Manager, Portfolio Development,
Digital Solutions, Munich Re Munich Re Ventures

Patrick Hinze Michael von Gablenz

Global Head of Green Tech Solutions, Munich Re Head of Insure AI, Munich Re and HSB
Munich Re Page 111/111
Tech Trend Radar 2025

Contact Munich Re
Martin Thormählen
Chief Technology Officer
[email protected]

Contact ERGO
Daniel Grothues
Chief Architect Primary Insurance
[email protected]

Münchener Rückversicherungs-Gesellschaft (Munich

Reinsurance Company) is a reinsurance c ­ ompany organised
© 2025 under the laws of Germany. In some c ­ ountries, including in
Münchener Rückversicherungs-Gesellschaft the United States, Munich Reinsurance Company holds the
Königinstrasse 107, 80802 München, Germany status of an unauthorised reinsurer. Policies are underwritten
by Munich Reinsurance Company or its affiliated i­ nsurance
Picture credits: Iana Kunitsa / Getty Images; Mina De and reinsurance subsidiaries. Certain c ­ overages are not
La O / Getty Images; pixdeluxe / Getty Images; da-kuk / available in all juris­dictions.
Getty Images; Malambo C/peopleimages.com / Adobe
Stock; SDI Productions / Getty Images; Yaroslav Kushta / Any description in this document is for general information
Getty Images; Andriy Onufriyenko / Getty Images; Yana purposes only and does not constitute an offer to sell or a
Iskayeva / Getty Images; Andriy Onufriyenko / Getty solicitation of an offer to buy any product.
Images; Xuanyu Han / Getty Images; Just_Super / Getty
Images; Imagine Photographer / Getty Images; Nikada / The companies and products mentioned throughout this
Getty Images; Master / Getty Images; Emanuel M report are for exemplary purposes only. Use of them does not
Schwermer / Getty Images; AerialPerspective Works / necessarily imply any affiliation with or endorsement by
Getty Images Munich Re (Group).