1 Data-driven Banking

Contents

1 Introduction 2

2 The State of Data-driven Banking in Switzerland 4

3 Technological Drivers for a Data-driven Future in Banking 5

3.1 Cloud . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.2 Open Financial Ecosystems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.3 Artificial Intelligence (AI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

4 Regulatory Aspects 10
4.1 Data Governance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.2 Regulatory Compliance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

5 Use Cases for Data-driven Banking 13

5.1 Use Case 1 - Cost Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
5.2 Use Case 2 - Risk Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.3 Use Case 3 - Revenue Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

6 Pathway to Becoming a Data-driven Bank 17

7 Conclusion 19

Authors 19

Bibliography 20
 Introduction 2

1. Introduction
The pace of development and adoption of new tech- industry internationally is shown by a survey conducted
nologies has accelerated in recent years. Companies by the Cambridge Centre for Alternative Finance. Their
can no longer delay the introduction of modern tech- report shows that 85 percent of financial institutions
nologies but must react quicker in order to be able to already use some forms of AI in banking (Cambridge-
maintain their position in the market in the long term. Centre, 2020).
Two key future technological developments that busi-
nesses will need to respond to are the increase in avail- However, financial institutions are not among the pi-
1
able data and advances in artificial intelligence (AI) . oneers in the field of data analytics and AI but large
Both trends have already received a lot of attention in technology companies such as Amazon, Apple, Google,
the past years and according to several studies will con- Meta and Netflix. They have harnessed these concepts
tinue to have a significant impact on our daily lives as to be amongst the most valuable publicly traded com-
well as the global economy. panies in the world. The established financial institu-
tions, in contrast, are lagging behind, even though they
According to PricewaterhouseCoopers (PwC), AI could
have been operating in one of the most data-intensive
contribute up to USD 15.7 trillion to the global econ-
sectors of the economy for decades. Yet several drivers
omy in 2030, with USD 6.6 trillion resulting from in-
exist that drive widespread adoption of data-driven
creased productivity and USD 9.1 trillion coming from
banking. A visualization of these drivers is given in Fig-
consumption side effects (PwC, 2017). Data collected
ure 1.1. First, technological developments in particu-
by McKinsey & Company show a similar picture. The
lar influence the use of data and corresponding anal-
potential annual value of data analytics and AI across
yses to generate added value for banks. Three of the
various industries is estimated between USD 9.5 trillion
most relevant technological drivers in this context are
and USD 15.4 trillion, with the retail, consumer pack-
developments in cloud computing, the emergence of
aged goods and public and social sectors driving these
open financial ecosystems and the already mentioned
values. The results also show that the banking (po-
advances in artificial intelligence (magenta layer). In
tential annual value of USD 1 trillion) and insurance
addition to technological drivers, data governance and
(potential annual value of USD 1.1 trillion) industries
compliance with existing data- and technology-related
are expected to benefit from these technological trends
regulations are also considered important influencing
(McKinsey, 2020). A great future potential of AI is
factors in the application of data-driven banking (dark
also confirmed by Gartner. According to their assess-
green layer).2
ment, the plateau of certain AI technologies will only
be reached within five to ten years. Data analytics and
Only by taking into account technological develop-
AI are expected to further grow and be used to also
ments and the regulatory environment can the poten-
determine essential connections and insights, to help
tial of data-driven value creation by financial institu-
predict demand for services and to detect the chang-
tions be fully exploited. Depending on the application,
ing patterns of customer behavior by analyzing data in
the added value can be generated through concrete
near real-time (Gartner, 2021).
use cases in the levers for cost reduction, risk reduction
The fact that AI not only has great future potential but and revenue expansion (blue layer).
has already been able to gain a foothold in the financial
2 Note that social and economic drivers can also play an important
1 For the purpose of this white paper, Artificial Intelligence (AI) will be role in the implementation of data-driven banking, but are not
understood to include Machine Learning (ML). specifically discussed in this white paper.
3 Data-driven Banking

Open Financial
Cloud Artificial Intelligence (AI)
Ecosystems

Data Governance and Regulatory Compliance

Data-driven Bank

Cost Reduction Risk Reduction Revenue Expansion

Figure 1.1: Drivers and potential of data-driven banking. Source: Own illustration

The rest of the white paper follows the layer structure described in more detail. Subsequently, specific use
in Figure 1.1. After an overview of the current state cases for banks are derived and described for each of
of data-driven banking in Switzerland, the aforemen- the levers to improve business success, i.e., cost reduc-
tioned main drivers, technological and regulatory, are tion, risk reduction or revenue expansion.
 The State of Data-driven Banking in Switzerland 4

2. The State of Data-driven Banking in

Switzerland
The Swiss financial center is of great importance not tomers. Due to the higher volumes, banks were able
only for the domestic economy, but also internation- to absorb the decreasing margins’ negative impact on
ally. At the end of 2021, there were 239 banks in their revenues. Hence, taking a look at the figures
Switzerland, employing 90,577 full-time equivalents shows that Swiss banks have become more efficient
and managing assets worth CHF 8,830 billion, of which over time and that the effect of digitization has mate-
around 47 percent were attributable to foreign clients. rialized over the past years (Ankenbrand et al., 2022).
In the area of cross-border wealth management for A consistent focus on data-driven banking could fur-
private clients, Switzerland is the world leader (Swiss ther improve the position of the Swiss financial sec-
Bankers Association, 2022). In order to maintain this tor. For example, financial institutions could accelerate
position in the long term, the sector must adapt dy- their decision-making processes and understand their
namically to the constantly changing environment in customers better using data and adequate analytics.
which it operates. Digitization, enabled by technologi- This evolution could further contribute to improved ef-
cal progress, is one of these driving forces that is having ficiency and better risk management, for example, by
a significant impact on banking business and operating minimizing compliance risk and improving fraud de-
models. tection. Finally, banks have the opportunity to deliver
significant revenue growth by offering an enhanced
There are signs that the general trend towards digitiza- personalized experience for their customers by using
tion has led to efficiency gains at Swiss banks. In par- methodological concepts from the fields of data ana-
ticular, financial institutions were able to increase the lytics and AI.
balance sheet total and the assets under management
while keeping their costs relatively stable over the last However, most Swiss banks have not ventured deeply
years, as shown in Figure 2.1. into data-driven banking yet. This is despite the fact
that the potential of data analytics has generally
been recognized by bank representatives in Switzer-
180
land (Ankenbrand et al., 2019). Furthermore, Swiss
160 banks assess the importance of open financial ecosys-
Indexed at 100

140 tems, for example for data exchange, to increase sig-

nificantly in the future (Ankenbrand et al., 2022). With
120
the exception of hedonic models for real estate valua-
100
tion, such a broad-based exchange of data across dif-
80 ferent providers hardly takes place in the Swiss finan-
2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
Balance sheet total Assets under management
cial center today, although the potential of a standard-
Total income Total expenses ized, API-based data exchange between financial in-
stitutions would be beneficial, as it could alleviate one
Figure 2.1: Size, costs and income indicators for Swiss
banks indexed at 100. Source: SNB (2022a,b,c) of the biggest challenges in data analysis, namely the
availability of the required data in sufficient quantity
and quality. Hence, Swiss banks seem to be aware of
The positive development is also reflected on the rev- the important future role of data in advancing their
enue side, but not as clearly. This indicates that parts business and operating model. The specific technolog-
of the efficiency gains were passed on directly to cus- ical drivers of this are discussed in more detail in the
following chapter.
5 Data-driven Banking

3. Technological Drivers for a Data-driven Fu-

ture in Banking
Technological developments have influenced the bank- • Behavioral data: interactions across different
ing business in the past and will continue to do so in channels (digital & non-digital) such as web us-
the future. The drivers explained in more detail in the age data, CRM data
following, i.e., cloud, open financial ecosystems and AI
are considered to be particularly relevant for the Swiss Data typically needs to be pre-processed before it can
financial industry as they enable flexible scaling, a stan- be stored efficiently, shared with third parties or used
dardized and therefore efficient form of interaction be- for data analytics. This includes steps such as the merg-
tween different providers, as well as the latest method- ing of different data sources, data quality assurance,
ological approaches to finding new insights in differ- outlier detection and applying data transformations.
ent areas of the financial domain. The fundamental Often, this is a labor-intensive activity, which uses a
resource of all three technology drivers is data, on the large share of the available resources in an analytics
basis of which financial institutions can create added project.
value for themselves and their customers.
The term big data is often used in connection with
The data available at banks can be divided into the fol- data at banks, as the institutions themselves have large
lowing three main types: amounts of internal data. According to Schroeck et al.
(2012), big data can be broken down into the four V-
• Master data: customer data, current and past dimensions volume (scale of data), velocity (speed of
positions, products and services, demographic, arrival of new data), variety (different forms of data),
socio-economic, geographic data and veracity (different data quality). Volume is often
• Transaction data: cards, payment transactions, seen as the most important dimension. The challenges
account movements of big data often lie in setting up an appropriate IT

4
Data size

0
0 1 2 Today
3 4 5 6 7 8 9The10
future
11 12 13 14 15
Figure 3.1: Big data and information overload. Source: Bloem et al. (2013)
 Technological Drivers for a Data-driven Future in Banking 6

infrastructure and data management system that col- • Transparency and collaboration between institu-
lects and stores data from various (internal) sources tions and cloud providers with regard to mea-
and uses computational power to process the data. A sures ordered by the authorities and the courts
key difficulty when dealing with big data is to separate • Choosing a cloud provider and its subcontractors
noise from signal. Figure 3.1 illustrates that the propor-
• Auditing the cloud services and the cloud infras-
tion of irrelevant data in relation to the relevant data is
tructure used to deliver them
expected to grow in the future as more and more data
is gathered.
The Swiss Bankers Association (SBA), the banks indus-
The three main technological drivers of data-driven try group, has worked with member institutions, audit
banking, described in more detail in the following, can firms and providers to draw up a set of guidelines for
help financial institutions generate maximum value the use of cloud services and to help banks migrate
from data. their data to the cloud more easily and securely (Swiss
Bankers Association, 2020a). The main reason for this
3.1. Cloud was that there were legal and regulatory uncertainties
that needed to be clarified in order for banks to use the
To benefit from data-driven strategies, the data pro- cloud in a compliant manner. The working group has
cessing infrastructure must become more open and focused on developing a legally non-binding guide as
accessible to allow flexibility to integrate and scale an interpretive tool for practitioners. Uncertainties are
new banking product and services. New requirements rated as high or hindering migration to the cloud in the
can no longer be met with the traditional on-premise above four areas (i.e., governance, data processing, au-
model. Migrating large parts of IT services to the cloud thorities and proceedings, audit). The clarification of
is an integral part of any digitization strategy. The the regulatory issues in the guide allows banks to act
undisputed advantage of a cloud strategy also requires quickly and flexibly and provides pragmatic and secure
a comprehensive cloud governance as well as new skills, solutions. Even with this guide, the assessment of the
and offers the opportunity to strategically reposition risks of migration to the cloud remains with the indi-
the organization (Burger, 2019). Leading financial in- vidual banking institutions. Therefore, each bank must
stitutions such as DBS (Hire Digital, 2021) are already decide for itself to what extent and with which provider
transforming into data-driven organizations, largely op- it wants to use cloud solutions.
erating in the cloud.
A representative survey of the Lucerne University of Ap-
Swiss Banks are following the trend to the cloud with plied Sciences and Arts, which was conducted as part of
caution. So far, they have been reluctant to migrate the IFZ Sourcing Study 2022 (Blattmann et al., 2022),
their data, especially sensitive customer data, to the shows reasons for and against the use of a cloud infras-
cloud. However, a development can be observed re- tructure. The main arguments in favor are the available
garding the shift of the operation of the less criti- software and functionality (71%) and the low prices
cal environments (e.g., the development environment) for computing and storage services due to easy scal-
without customer data to the cloud, while the actual ing (43%). However, the data storage abroad (71%)
production environment continues to be hosted on- and perceived insufficient data security (46%) speak
premise. against it. In addition, the survey shows that for more
than three quarters of banks, IT (77%) plays the cen-
Four areas are of critical importance when it comes to
tral role when it comes to moving banking services to
migrating banking solutions to the cloud (Swiss Bankers
the cloud. The management (43%) also plays a signif-
Association, 2020a):
icant role when it comes to bringing the topic of cloud
• Processing data on bank clients in compliance into the bank. This may have something to do with the
with banking secrecy fact that cloud computing is now no longer seen merely
7 Data-driven Banking

as a way of optimizing costs. Rather, cloud computing end customer, who receives a financial product or ser-
is becoming increasingly important as a technical com- vice, as well as the providers who operate the front-
ponent of future business and operating models. The end to this customer. The third layer describes those
study concludes that the time has also come for smaller providers that are finally responsible for the execution
and medium-sized retail banks to look into cloud com- or settlement of the financial products or services, as
puting, and that quite a few banks seem to have more well as the custody of the corresponding assets, while
or less concrete ideas or plans for shifting workload layers four and five relate to the technological infras-
to the cloud on the table. One reason for this is that tructure and the handling of data by these providers.1
cloud computing is also becoming increasingly impor- In open financial ecosystems, all five horizontal layers
tant in connection with open financial ecosystems (e.g., are to some extent interconnected via an API environ-
open banking), another technological trend with impli- ment (lines shown in magenta), which in turn defines
cations for the Swiss banking landscape. the security of the system, its compliance with legal and
regulatory requirements, and the certification and iden-
tification of approved parties.
3.2. Open Financial Ecosystems
Open financial ecosystems can take various forms, with
Open financial ecosystems are a trending topic in bank-
the opening up of banks often being subsumed under
ing and describe the leveraging of seamless API-based
the term “open banking”. In general, open banking rep-
interactions between financial services providers, as
resents a development in the financial services industry
well as non-financial firms and individuals for value cre-
that allows players to share financial information elec-
ation. In general, open ecosystems can be understood
tronically, securely and only under conditions that cus-
as systems between interacting organizations and are
tomers approve of. With regard to the architecture in
enabled by modularity and complementarity proper-
Figure 3.2, application programming interfaces (APIs)
ties (Hakanen, 2021), with data as the main resource.
allow third-party providers (TPPs) like FinTech compa-
An architecture for structuring different manifestations
nies to access financial information efficiently, which
of financial ecosystems is given in Figure 3.2.
promotes the development of new apps and services.
The architecture distinguishes between five different 1 Formore information on the architecture, see chapter six in the IFZ
horizontal layers. The upper two layers represent the FinTech Study 2021 (Ankenbrand et al., 2021).

Figure 3.2: Architecture of financial ecosystems. Source: Ankenbrand et al. (2021)

 Technological Drivers for a Data-driven Future in Banking 8

Ideally, open banking results in a better experience for ready a number of platform providers covering a range
consumers through an improved customer journey (see of APIs for payment initiation, account information ac-
Figure 3.3). cess and wealth management services, among others
(Ankenbrand et al., 2021).
One of the first regulation-driven initiatives towards
open financial ecosystems was the Open Banking ini-
tiative launched by the UK Competition and Markets As a survey of the Swiss financial industry shows, how-
Authority in January 2018 (Open Banking, online), fol- ever, the local banks do not yet feel much pressure
lowed by the European Union’s PSD2 in September to open their interfaces to third-party providers, even
2019 (European Commission, online), which makes the though there is perceived potential for corresponding
opening of certain data to TTPs mandatory for banks. solutions, especially in the area of payments (Anken-
While these regulations do not apply to Swiss banks brand et al., 2021). In order to promote the success of
directly, it still is a strong motivation to adopt similar open financial ecosystems without the need of issuing
concepts and open up to corresponding standards, as corresponding binding regulations, there is a particu-
the benefits of connected or integrated banking appli- lar need to incentivize the participation of banks that
cations outweigh the potential threats like security risk hold large amounts of financial data and to reduce ex-
or loss of control. In order to create good conditions isting barriers. In this context, banks see the greatest
for a market-driven implementation of open banking benefit of open financial ecosystems in the simplified
in Switzerland, the Swiss Bankers Association has pub- collaboration with third-party providers, for example, to
lished a set of guidelines for implementation already in leverage specialized analytics capabilities from experts,
2020 (Swiss Bankers Association, 2020c). and the potential for an improved customer experience,
while significant effort and cost as well as a perceived
One development that is emerging in the context of lack of standards and security related to APIs are the
open financial ecosystems is the trend away from in- two biggest obstacles (Ankenbrand et al., 2021). In
dividual API-based collaborations between financial in- summary, however, it can be said that Swiss banks still
stitutions and third-party providers toward platform so- have a lot of untapped potential in the area of data ex-
lutions for a large number of market participants. This change with third-party providers, although the prereq-
can also be observed in Switzerland, where there are al- uisite for its use would actually be given. Improved data

Reviews

Research
Digital Value Chain

Review/ Feedback
Customer App Developer API API Team Backend

Support/ Return
Use

Get Product/ Service Order/ Subscribe

Figure 3.3: Customer journey in the digital value chain. Source: Google Cloud (2022a)
9 Data-driven Banking

availability and volume would provide an expanded ba- growing number of Swiss FinTech companies, which act
sis for harnessing the potential of data-driven analyt- as suppliers to the established financial industry, that
ics approaches, for example, through the use of data- use AI (Ankenbrand et al., 2022). However, only a few
intensive methods from the field of AI. of them make it a critical element of their strategic
planning. Given the high pace of innovations in the
3.3. Artificial Intelligence (AI) field of AI and the aggregation of data through initia-
tives like open banking, the significance of AI for the
A further change in the technological environment in Swiss financial industry will further increase.
general is the adoption of artificial intelligence in al-
most every industry. The leap in technical possibilities What needs to be in focus is developing AI applications
driven by AI research lead to a variety of new innova- that solve important problems and ultimately help cus-
tions and business cases. The main factors behind this tomers in their day-to-day financial lives. In general,
process are innovations in the field of deep learning, there is large potential for AI and other advanced tech-
the growing amount of available data (e.g., by open fi- nologies to empower people, widely benefit current and
nancial ecosystems) and the accessibility of relatively future generations, and work for the common good.
cheap computing power (e.g., via cloud computing). AI applications need to be assessed, also from a risk
The potential of AI in finance is estimated to add USD perspective, against the following objectives (Pichai,
1 trillion in annual value, of which USD 660 billion is ac- 2018):
counted for by traditional AI (e.g., clustering and basic
1. Be socially beneficial
regression) and around USD 360 billion by advanced AI
2. Avoid creating or reinforcing unfair bias
(e.g., neural networks) (McKinsey, 2022).
3. Be built and tested for safety
On a global level, a considerable number of banks are 4. Be accountable to people
already applying AI, as a survey by The Economist In- 5. Incorporate privacy design principles
telligence Unit shows. The most common use cases are 6. Uphold high standards of scientific excellence
listed in Figure 3.4. According to the survey, the major- 7. Be made available for uses that accord with these
ity of banks surveyed already heavily apply AI for fraud principles
detection (57.6%), optimizing IT operations (53.7%)
To create ethical and unbiased AI solutions, banks can
and digital marketing (50.2%) (The Economist, 2022).
adopt a number of best practices. The following con-
siderations can serve as examples:
0% 25% 50% 75% 100%

Fraud detection 57.6% • Use a human-centered design approach

Optimizing IT operations 53.7% • Identify multiple metrics to assess training and
Digital marketing 50.2%
monitoring
Risk assessment 48.3%
Personalising customer experience 43.9%
• When possible, directly examine your raw data
Credit scoring 42.4% • Understand the limitations of your dataset and
Optimizing product design 42.0% model
Sales and marketing optimization 40.0%
Personalizing investments
• Test AI solutions consistently and thoroughly
39.5%
Portfolio optimization 36.6% • Continue to monitor and update the system after
deployment
Figure 3.4: Proportion of banks making heavy use of
AI per use case. Source: The Economist (2022) If these ground rules are followed, AI can lead not only
to better business performance for the banks them-
selves, but also to improved products and services for
Also in Switzerland, many banks already use AI in one or end customers, for example through greater personal-
few business cases. This is underscored by the steadily ization.
 Regulatory Aspects 10

4. Regulatory Aspects
Financial services providers in Switzerland as else- 4.1. Data Governance
where need to carefully consider regulatory aspects in
their decision-making and continuously adapt to their There are many advantages to sharing data. However,
changes. Switzerland’s regulatory framework govern- due to the often high sensitivity of financial data, finan-
ing financial services activities consists of various fed- cial institutions tend to be reluctant to share them with
eral laws and implementing ordinances (Ankenbrand third parties. The potential value of data sharing must
et al., 2021). Two legal frameworks in particular are be weighed against the implications for customer pri-
the focus of interest as of November 2022. The Swiss vacy, data security, and control over competitive data.
Financial Market Architecture consists of the Financial New ways, e.g. privacy enhancing technologies, offer
Services Act (“FinSA”) and the Financial Institutions Act a possible solution to this conflict of objectives. These
(“FinIA”). Both FinSA and FinIA entered into force on 1 new approaches enable the protection of sensitive data
January 2020 governing the provision of financial ser- at rest, in transit and in use, which can reduce concerns
vices, offering financial instruments and the respective about data sharing and strengthen or, in extreme cases,
licensing requirements in Switzerland. even replace trust in institutions.

FinSA primarily sets out requirements applicable to the

In Europe, the trend is towards digital self-
provision of financial services and the offering of fi-
determination and consumer sovereignty, in com-
nancial instruments in Switzerland. FinIA provides for
parison to approaches in, for example, the United
a comprehensive supervisory licensing regime applica-
States and China (see Figure 4.1). The European Com-
ble to portfolio managers, trustees, managers of col-
mission laid another cornerstone to this end in the year
lective investment schemes, fund management com-
2020 (Swiss Bankers Association, 2020b) by publishing
panies and securities firms.
a new data strategy (European Commission, 2020a).
FinSA and FinIA impact both “traditional” financial ser- This strategy aims to make more data available for
vices providers (FSP) and FinTech companies. FinSA use in the economy and society. The focus is primarily
covers financial services providers, client advisers and on non-personal data. At the same time, the aim is to
manufacturers, and providers of financial instruments. ensure that data producers - i.e., citizens - retain control
This could be relevant if a fully licensed and regulated over their data. The European Commission’s strategy
FSP intends to collaborate with a FinTech company. also includes measures on how this is to be achieved.
They include the creation of transparent and fair rules
FinSA sets out rules of conduct, which namely cover A) for data access and usage and the equipment of users
information duties, B) suitability and appropriateness with rights, tools and competences so that they retain
checks, C) documentation and accountability duties as full control over their data. Another priority of the
well as D) duties regarding transparency and due care. strategy is the expansion of the European infrastruc-
The information duties aim at providing clients a com- ture, especially cloud capacities, in order to be able to
prehensive and transparent overview of the services process and store the ever-increasing amounts of data
and products offered by the Financial Services Provider directly in Europe. This is intended to reduce the cur-
(FSP). In the context of data-driven banking, a key fo- rent dependence on global, predominantly American,
cus for banks is data governance and compliance, both technology companies (Swiss Bankers Association,
of which are discussed in more detail in the following 2020b).
sections.
11 Data-driven Banking

Figure 4.1: Comparison of data protection cultures. Source: Swiss Bankers Association (2020b) with information
from Axon Active

In parallel with the data strategy, the Commission clients’ financial data in particular play a key role in this
published a communication entitled “Shaping Europe’s regard. It is important that Swiss banks are not at a
digital future” (European Commission, 2020c) and a disadvantage in the new global competitive environ-
white paper on artificial intelligence (European Com- ment of the data economy. A regulatory environment
mission, 2020b), in which the Commission sets out how that allows the benefits of data use to be optimally ex-
it intends to support and promote the development and ploited while at the same time taking the correspond-
use of AI across the EU. With this work, the EU has laid ing risks into account can help position Switzerland as
an important cornerstone for the future design of the a haven of security and trust.
European data economy, underscoring the importance
of data and artificial intelligence for the entire econ- Whether financial services providers rely on the pri-
omy. The EU’s ambitions are high and the timeframe it vate, public or multi cloud, the data hosting infrastruc-
has set for itself is tight. It will therefore soon become ture and control framework have to satisfy regulatory
clear whether the many words will be backed up by ac- requirements by the Swiss Financial Market Supervi-
tions. sory Authority (FINMA). FINMA defines the supervi-
sory requirements applicable to outsourcing solutions
There are no comparable regulations in Switzerland. at banks, securities dealers and insurance companies
However, due to the high level of integration of the in terms of appropriate organization and risk limitation
Swiss financial sector in international banking, it must (FINMA, 2018).
also adapt to international requirements and, if possi-
ble, align accordingly. FINMA does not prohibit customer data to be held
abroad (Walder Wyss, 2019). The Swiss Banking Act
doesn’t differentiate between domestic and foreign.
4.2. Regulatory Compliance
However, up to now banks have predominantly stored
In order to keep its capacity for innovation and compet- data domestically and prevented it from being ac-
itiveness as a business location high, Switzerland must cessed from abroad. In practice, there is often a mis-
make use of its growing wealth of data. Banks and their conception among banks that the moment data is pro-
 Regulatory Aspects 12

cessed or accessed from abroad, banking secrecy would contains requirements for an appropriate organization
automatically be compromised. The Swiss Bankers As- and aims to limit their risk. FINMA circular 2008/21
sociation takes a different view in its guidelines for defines CID (client identifying data) as customer data
cloud use by Swiss banks (Swiss Bankers Association, that contains personal data as defined in Art. 3a of
2020a). One of the main issues is to ensure that unau- the Federal Act on Data Protection (FADP) (FADP, 2022)
thorized parties cannot access the data. There are and makes it possible to identify the customers af-
also legal opinions that state explicitly that data can fected.
be stored and processed abroad provided appropriate
safeguards are adopted (Laux Lawyers, 2019; Walder
The existing regulations in Switzerland do not prevent
Wyss, 2019).
the outsourcing of business processes, for example in
FINMA Circular RS 2008/21 Annex 3 regulates the reg- the area of data analysis, or the storage of data in the
ulatory requirements for handling electronic customer cloud (even abroad). The prerequisites for data-driven
data from banks and securities dealers. For these, it banking would therefore also be met for Swiss banks.
13 Data-driven Banking

5. Use Cases for Data-driven Banking

To maintain its position as a financial innovation hub of telligence. The advantage of such innovations can be
the world, Swiss banks must continuously adapt to the demonstrated with the example of a fully autonomous
changing environment. With the emergence of FinTech onboarding process, one of the few processes every cus-
companies and challenger banks offering low-friction tomer of a bank is confronted with. An implementation
services that focus on a customer’s macro (savings) and of this use case not only reduces costs and increases
micro (payments) financial journey, traditional finan- efficiency, but also has a strong impact on customer
cial institutions must also increasingly embrace data- satisfaction. It furthermore shows a strong technology
driven banking. In general, this can create positive affinity of a bank in vital customer-facing processes.
added value for the business success of banks via the
following three levers: For a long time, visiting the bank with official docu-
ments was the only way to verify the user’s identity
1. Cost reduction: Technology and data can be and his or her eligibility to open a bank account. This
used to digitize and automate existing business process is not only inconvenient for the customer, as it
processes, for example, by reducing paper-based is time-consuming, but is also cost-intensive from the
or human effort. bank’s point of view due to the large amount of manual
work involved. As a consequence, many banks digitized
2. Risk reduction: While expanding access to prod-
the onboarding process by providing online identifica-
ucts and services, data can be used to ensure
tion via video calls, in which an employee of the bank
that risk is managed correctly, so that, for ex-
verifies the prospect, in recent years.
ample, the right products match with the right
customers, and that any corporate or regulatory With advances in AI, especially with regard to facial
compliance mandates are adhered to. recognition and the passage of regulations, a new gen-
eration of fully autonomous customer onboarding solu-
3. Revenue expansion: Data can be used to expand
tions has recently been made possible. By completely
access to products and services to a broader cus-
eliminating the personal involvement of a bank em-
tomer base and help customers achieve their per-
ployee, the potential of automation is fully exploited,
sonal financial goals.
and costs can be reduced accordingly. From the cus-
In the following sections, specific use cases for these tomer’s point of view, a fully automated onboarding
three levers are presented individually. Note that these process can be described as follows (see Figure 5.1):
use cases are not discussed in full depth but serve as ba-
• Step 1: Verify identity document: With the help
sic ideas for data-driven banking and are based on the
of modern machine learning models, the identity
technological and regulatory enablers discussed.
verification tasks are achieved without human in-
teraction.
5.1. Use Case 1 - Cost Reduction
• Step 2 & 3: Liveness detection by biometric
Swiss banks have kept their total costs stable over the identification: Liveness detection prevents on-
past few years, as has been shown in Figure 2.1. How- boarding of misrepresentation through imper-
ever, technology- and data-driven innovations have the sonation by leveraging advances in facial recog-
potential to reduce ongoing costs in certain areas, for nition. Without such mechanisms, a person
example by automating business processes and mak- could simply hold a victim’s social media picture
ing them more efficient through the use of artificial in- into the camera and the face comparison model
 Use Cases for Data-driven Banking 14

ID or passport Biometric Biometric Check address Check contract & Yeah! Done
identification identification transmit information

Select your identify- Check and validate Take photos in The address is Accept legal framework You are now a
cation document (ID the document with different poses to checked by conditions (AML, AIA etc.) using customer!
or passport) and scan a fake check and verify that you are comparing it with a confidential signature system. Your account is
it. read the a real person. the address data of Your data is stored and verified open and ready
information using the Swiss postal directly in the system via a core to use.
NFC. service. banking system connection.

Figure 5.1: Account opening process screenflow including KYC. Source: ti&m (2022)

would recognize a match with the victim’s stolen • Step 1: Generate data using bots: Setting up a
document. fully automated search bot that scours search en-
gines for a specific PEP, generates a dataset of
• Step 4: Address check: The address is checked by
relevant public information.
comparing it with the address data of the Swiss
postal service via a simple API call. • Step 2: Analyze data using AI: Artificial intel-
ligence is used to check the generated data
• Step 5: Sign contract digitally: Accept le-
for controversial content and verify whether the
gal framework conditions including Know-Your-
search hit is actually related to the person of in-
Customer (KYC) information (e.g., Anti-Money
terest. Whether a person is tested positively or
Laundering (AML) or Automatic Information Ex-
negatively can be done either fully automatically
change (AIE)) using a confidential signature sys-
(considering the ethical aspects in Section 3.3) or
tem based on a trust service provider (TSP). The
in combination with a human expert.
data is stored and verified directly in the system
via a core banking system connection. A digi- These introductory examples show that also very
tal signature replaces the classical signature on customer-facing processes can be optimized using
a contract. data-driven approaches. These automations are inter-
esting from a banking perspective, particularly because
In addition to customer onboarding, there are also of the potential efficiency gains, while customers also
other data- and AI-driven opportunities to reduce costs benefit, for example through an improved customer ex-
in the area of screening (potential) customers. For ex- perience.
ample, banks can install automated processes in the
area of politically exposed persons (PEPs) to screen
5.2. Use Case 2 - Risk Reduction
them for critical reporting in public sources, such as on-
line newspapers. Such a process could be implemented More recent FinTech-based banking approaches do not
in two steps: completely remove the human element from banking.
15 Data-driven Banking

At some point in the client relationship, there is a time tion of different scenarios for the development of the
where a personal interaction is required over the phone relevant drivers. Scenario analysis is highly relevant for
or in a branch. Such interactions are normally around risk management in banks. Typically, it involves highly
the sale of new products and services, and usually with complex, large-scale simulations that generate billions
some background checks or risk assessments involved. of data points. Traditional on-premise hardware solu-
These essentially still manual processes can be sup- tions have limited capacity, which is why such scenario
ported by data-driven practices to optimize risk assess- analyses are very time-consuming. With a cloud infras-
ment on a client-specific basis. tructure, such data-intensive risk management can sig-
nificantly accelerate decision-making. Potentially valu-
One area where data can be used to mitigate risk is in able input data in this context could be, for example,
the lending business. Banks currently use a number of the following information (Sigrist & Hirnschall, 2019):
techniques to mitigate the credit risks to which they are
exposed. For example, exposures may be fully or par- • Financial performance indicators from balance
tially collateralized with cash or securities through se- sheets and/or income statements
nior claims, a credit exposure may be guaranteed by • Repayment history
a third party, or a bank may purchase a credit deriva- • Loan characteristics (e.g., amount, maturity)
tive to offset various forms of credit risk. In addition,
• Public data (e.g., from social media platforms)
banks may agree to offset loans owed to them with
• Data from third parties such as credit bureaus, in-
deposits from the same counterparty. Non-financial
surance companies and/or government (e.g., tax
companies also continuously monitor and analyze their
filings)
credit risk exposure to avoid potential counterparty de-
faults. Besides assessments by rating agencies, credit In addition to the advantages for the bank through
default swaps are commonly used financial instruments optimized risk management and the resulting opti-
that provide information on the creditworthiness of a mized allocation of available financial capital, data-
counterparty. However, the underlying pricing dynam- driven credit risk management also offers advantages
ics often remain unknown and demand further investi- for clients, for example in the form of more accurate in-
gation (Roeder et al., 2022). Moreover, such risk mea- terest rates based on their individual financial situation.
surement and hedging tools exist for large companies
but are not available in the context of small business or
5.3. Use Case 3 - Revenue Expansion
personal loans.
In addition to measures on the cost side, optimization
For such kinds of loans, in particular, data-driven meth- on the revenue side can also have a positive impact
ods can be a key concept for identifying and quantify- on the business success of financial institutions. Rec-
ing potential risks and can therefore be used to support ommendation systems are a well-suited tool for this
and justify credit decisions but also as an early warn- purpose. In general, a recommender system uses the
ing system for existing loans. Although some banks al- process of information filtering to predict which prod-
ready use such methods, they tend to be rule-based and ucts a user might like and rates the products according
therefore poorly adaptive, especially in times of uncer- to the user’s preferences. While such systems are al-
tainty. The use of dynamic and self-learning methods ready widespread in other industries such as entertain-
from the field of artificial intelligence can further in- ment (e.g., “people who liked this movie also liked that
crease the forecasting accuracy of, for example, loan movie”), there is still a lot of untapped potential in most
default probabilities and thus improve risk manage- financial institutions.
ment at banks. If a scalable infrastructure is available,
a wide range of internal and external data can poten- Specifically, recommendation systems can be used in
tially be used for such modelling, including considera- banking to provide existing customers with additional
 Use Cases for Data-driven Banking 16

Collaborative filtering Content-based filtering

Owned/used by both clients Owned/used by client

Similar clients Similar products/

services

Owned/used by her, recommended Recommended to client

to him

Figure 5.2: Recommender system in banking. Source: Based on Wu et al. (2018)

financial products or services based on their personal

needs and circumstances. The individual preferences of that of a randomly selected person. For example,
clients could be determined by using existing internal information about a customer’s portfolio alloca-
and possibly external data including: tion can be used to make further investment sug-
gestions to that customer based on the portfolio
• Current product ownership allocations of similar bank customers.
• Transactional behavior • Content-based filtering: Method for automati-
• Credit scores cally predicting a user’s interests by using simi-
larities in the features of products and services, as
• Demographics and socioeconomic factors
well as information accumulated about the user
• Location
to make recommendations. The underlying as-
• External social media profiles sumption is that if a person A has an opinion
about a particular financial product, it is more
In general, a distinction is made between collabora-
likely that A will have the same opinion about an-
tive filtering and content-based filtering to identify the
other financial product with similar features than
most appropriate recommendations (see Figure 5.2).
about a randomly chosen product. For example,
Both approaches are based on methods from the field
information about a bank customer’s sustainabil-
of artificial intelligence, such as clustering, neural net-
ity preferences, as measured by the ESG rating of
works, or decision trees, and are situationally prefer-
the investments in his/her current portfolio, can
able depending on the context of use and the available
be used to recommend only products with a cor-
data. The two approaches can be described as (Mole-
responding ESG rating.
naar, 2022):
A high-performance recommender system offers con-
• Collaborative filtering: Method for automati-
siderable advantages for a bank. These include higher
cally predicting a user’s interests by collecting
revenues through up- and cross-selling, higher conver-
preferences from many users. The underlying
sion rates and reduced customer churn. But customers
assumption is that if a person A has the same
also benefit directly, for example through increased
opinion as a person B about a particular finan-
personalization and customer experience, resulting in
cial product, it is more likely that A will have B’s
increased customer satisfaction.
opinion about a different financial product than
17 Data-driven Banking

6. Pathway to Becoming a Data-driven Bank

The pathway to becoming a data-driven bank can seem clear agreement on the collaboration model during this
daunting in its complexity and time to value. However, period. It is critical to identify a business use case that
both can be broken down via a sound approach. One of can show meaningful results over the agreed number of
the most important points is that a bank has a vision of weeks. The first phase is about the creative process of
a data-driven business model that can be built as a se- generating, developing and communicating new ideas.
ries of modules or Minimal Viable Products (MVPs) to For many use cases, it can be beneficial to include, for
reduce complexity as well as incrementally realize time example, experts with regard to user experience in the
to value. Such a gradual approach can also be used MVP team to create a positive customer experience. In
to continuously gain experience and sharpen the un- the second phase, the joint MVP team iterates on the
derstanding of the possibilities of data-driven banking. solution in line with the Agile principle of “fail fast”. Fail-
The creation of a specific MVP, for example for one of ing fast means to have a process of starting work on
the use cases described earlier, is often done in collabo- a project, immediately gathering feedback, and then
ration with a third party provider (TPP) like a specialized determining whether to continue working on that task
technology company and can generally be divided into or take a different approach. The MVP is thus contin-
three main phases, which are shown in Figure 6.1. uously adapted in short steps to the generated feed-
back. Finally, the MVP team implements the solution
An MVP typically runs over a period up to 12 weeks, identified and provides the information required for in-
starting with the definition of success criteria and a ternal stakeholder management and decision-making.

Ideation Iteration Implementation

Week 2-4

Week 1 Week 5-9 Week 10-12

Set
Define goal Identify functional Design Build Run
of stage deliverables success solution Validate
environment retrospective
criteria

Bank & Bank & Bank & Bank & Bank & Bank & Bank &
TPP TPP TPP TPP TPP TPP TPP

Based on Define the

success Design the Building the Validate the
Identify
Define the the business solution to environment what could
goals, define criteria that outcome of
business need to be meet all can be sped phase
have been
deliverables requirement up by our done better
goals of next and met for against
phase phase s and blueprints success and was
specialists reiterate as and good for the
needed successful criteria
and finished necessary experience next phase

Figure 6.1: MVP approach. Source: Based on Google Cloud (2022b)

 Pathway to Becoming a Data-driven Bank 18

The incremental creation of such MVPs and their scal- such cooperation, but also gain relevant experience,
ing and linking ultimately leads to the realization of the build up internal competences and get used to a new
vision of a data-driven bank. culture of innovation.

One obstacle for many Swiss banks to develop such In general, the technological and regulatory conditions
MVPs and thus to move towards data-driven banking is for a transformation to a data-driven bank would be
their limited innovation capacity in the IT area (Anken- in place in Switzerland. However, in order to practice
brand et al., 2022). This can be countered by work- data-driven banking successfully, Swiss banks need a
ing with external partners. Such a service is offered by fundamental change in mindset. The often prevail-
various providers in Switzerland, typically with strong ing compliance mindset, which in many cases prevents
IT competences, and usually includes the entire pro- or at least slows down innovation, must give way to
cess chain for creating specific MVPs in a short time a technology- and data-friendly culture in order to be
using specialized teams. Banks can therefore not only able to exploit the full potential of data-driven banking
achieve quick successes in data-driven banking through within the given legal framework.
19 Data-driven Banking

7. Conclusion
The present white paper aimed to analyze and discuss Great potential in many areas of the banking busi-
the drivers and potential of data and corresponding an- ness model. The application of data-driven banking
alytics for the Swiss financial sector in a structured way. can have a positive impact on the business success of
The core findings are summarized in the following four banks through three main levers. First, concrete use
statements and theses: cases, such as fully automated customer onboarding or
checks of potentially politically exposed persons, can
Comparatively low penetration of data-driven bank- reduce costs for financial institutions. Secondly, busi-
ing in the Swiss financial industry. Most Swiss banks ness risks in banking can also be reduced through the
have not yet explored deeply into data-driven banking. use of data-driven insights, e.g., in the area of lend-
This is despite the fact that the potential of data ana- ing through more accurate default predictions. Thirdly,
lytics and artificial intelligence for their business model in addition to potential improvements on the cost and
has been generally recognized by bank representatives risk side, the revenue side can also benefit from data-
in Switzerland, as existing surveys on this topic show. In driven banking. Concrete applications such as recom-
order for Switzerland to remain one of the world’s lead- mendation systems can help financial institutions to
ing financial centers in the future, the industry must, increase their revenues through up- and cross-selling,
however, continuously and dynamically adapt its busi- higher conversion rates and lower customer churn. But
ness models to the changing environment so that it re- customers also benefit directly, for example through
mains competitive in the long term. This increasingly better personalization and customer experience, lead-
requires the broad and consistent use of data and data ing to higher customer satisfaction.
analytics, as is already common in other industries and
is also more and more practiced by the Swiss FinTech
sector. Just do it. Due to the perceived high complexity and
long implementation times for data-driven banking,
Technological and regulatory developments are smaller banks in particular are discouraged from ben-
pushing banks. Important enablers or drivers for data- efiting from its great potential. However, the trans-
driven banking can be found in current technological formation to a data-driven bank can be done step by
and regulatory developments. The former relate in par- step with an organized approach, also in order to con-
ticular to the slowly emerging transition towards cloud- tinuously gain experience. An approach via the devel-
based IT infrastructures for Swiss banks, their opening opment of minimal viable products (MVPs) in cooper-
to third-party providers via open financial ecosystems ation with specialized technology providers is particu-
(e.g., open banking), as well as the continuous advance- larly suitable if the required competencies are not avail-
ments in the field of artificial intelligence to generate able in-house. In a short period of a few weeks, this
new insights for banking based on internal and external can already produce initial results in certain use cases
data. Relevant regulatory aspects in connection with that the bank considers to be particularly value-adding,
data-driven banking relate in particular to data man- which in turn can serve as the basis for further devel-
agement and compliance, which, like the technologi- opment in the direction of a data-driven bank. The
cal requirements, cannot in principle be seen as an ob- banks should simply dare to take the first step, because
stacle to the use of corresponding approaches by Swiss the necessary conditions are in place in Switzerland, be
banks. they technological or regulatory.
 Authors 20

Authors

This white paper was prepared in collaboration with the following individuals who contributed in the form of text,
discussion, document reviews and other forms of feedback (in alphabetical order):

Thomas Ankenbrand Timon Kronenberger

Head Competence Center Investments Master’s Assistant
Lucerne University of Applied Sciences and Arts Lucerne University of Applied Sciences and Arts

Denis Bieri Daniel Ott

Research Associate Head Product Management
Lucerne University of Applied Sciences and Arts ti&m

Fabian Braunwalder Ranja Reda Kouba

Head Products & Banking Innovations Head of Customer Engineering FSI
ti&m Google Cloud, Germany

Ursin Brunner Fabio Sigrist

Head Machine Learning Professor of Applied Statistics and Data Science
ti&m Lucerne University of Applied Sciences and Arts

Andreas Dietrich Andrew Viejra

Head Institute of Financial Services Zug IFZ Principal Architect Financial Services
Lucerne University of Applied Sciences and Arts Google Cloud, Switzerland

Gregor Kaelin Tanja Walser

Head of Financial Services Head of Customer Engineering FSI
Google Cloud, Switzerland & Austria Google Cloud, Switzerland

For more information about this report, please contact us at:

Thomas Ankenbrand Daniel Ott Gregor Kaelin

HSLU ti&m Google Cloud
[email protected] [email protected] [email protected]
21 Data-driven Banking

Bibliography

Ankenbrand, T., Bieri, D., Frigg, M., Grau, M., & Lötscher, D. (2021). IFZ FinTech Study 2021. An Overview of Swiss
FinTech. Retrieved 19/10/2022, from https://zenodo.org/record/5109653#.Y1AJtnZBzZt

Ankenbrand, T., Bieri, D., Frigg, M., Kronenberger, T., & Reichmuth, L. (2022). IFZ FinTech Study 2022. An Overview
of Swiss FinTech. Retrieved 19/10/2022, from https://hub.hslu.ch/retailbanking/wp-content/uploads/sites/7/2022/
03/IFZ-FinTech-Study-2022.pdf

Ankenbrand, T., Dietrich, A., & Bieri, D. (2019). IFZ FinTech Study 2019. An Overview of Swiss FinTech. Retrieved
19/10/2022, from https://blog.hslu.ch/retailbanking/files/2019/03/IFZ-FinTech-Study-2019_Switzerland.pdf

Blattmann, U., Buschor, F., Estermann, F., & Ettlin, J. (2022). IFZ Sourcing Studie 2022. Retrieved 19/10/2022, from
https://drive.switch.ch/index.php/s/n3QZPfyCqFmzYxT

Bloem, J. W., van Doorn, M., Duivestein, S., van Manen, T., van Ommeren, E., & Sachdeva, S. (2013). No more secrets
with big data analytics. Sogeti.

Burger, K. (2019). Assurance according to ISAE 3000 for ti&m hosting. Retrieved 19/10/2022, from https://news.ti8m
.ch/blog/en/Assurance-nach-ISAE-3000-f-r-das-ti-m-Hosting.html

CambridgeCentre. (2020). Transforming Paradigms – A Global AI in Financial Services Survey. Retrieved 19/10/2022,
from https://www.jbs.cam.ac.uk/wp-content/uploads/2020/08/2020-ccaf-ai-in-financial-services-survey.pdf

European Commission. (2020a). A European strategy for data. Retrieved 19/10/2022, from https://ec.europa.eu/info/
sites/default/files/communication-european-strategy-data-19feb2020_en.pdf

European Commission. (2020b). On Artificial Intelligence - A European approach to excellence and

trust. Retrieved 19/10/2022, from https://ec.europa.eu/info/sites/default/files/commission-white-paper-artificial
-intelligence-feb2020_en.pdf

European Commission. (2020c). Shaping Europe’s digital future. Retrieved 19/10/2022, from https://ec.europa.eu/
info/sites/default/files/communication-shaping-europes-digital-future-feb2020_en.pdf

European Commission. (online). Strong customer authentication requirement of PSD2 comes into force. Re-
trieved 19/10/2022, from https://finance.ec.europa.eu/publications/strong-customer-authentication-requirement
-psd2-comes-force_en

FADP. (2022). Federal Act on Data Protection (FADP). Retrieved 19/10/2022, from https://www.fedlex.admin.ch/eli/
cc/1993/1945_1945_1945/en

FINMA. (2018). Outsourcing – banks, insurance companies and selected financial institutions under FinIA.
Retrieved 19/10/2022, from https://www.finma.ch/en/~/media/finma/dokumente/dokumentencenter/myfinma/
rundschreiben/finma-rs-2018-03-01012021_de.pdf

Gartner. (2021). Gartner Identifies Five Emerging Trends That Will Drive Technology Innovation for the Next Decade.
Retrieved 19/10/2022, from https://www.gartner.com/en/newsroom/press-releases/2021-08-23-gartner-identifies
-key-emerging-technologies-spurring-innovation-through-trust-growth-and-change
 Bibliography 22

Google Cloud. (2022a). Customer journey in the digital value chain. (unpublished)

Google Cloud. (2022b). MVP approach. (unpublished)

Hakanen, E. (2021). Ecosystems and anti-rival goods – Greetings from project ATARCA. Presentation at HSLU.

Hire Digital. (2021). How DBS Made Data Its Culture. Retrieved 19/10/2022, from https://hiredigital.com/blog/
how-dbs-made-data-its-culture

Laux Lawyers. (2019). Use of Cloud Offerings by Banks: Admissibility under Art. 47 Banking Act. Retrieved 07/12/2022,
from https://www.lauxlawyers.ch/wp-content/uploads/2019/04/Swiss-Banking-Secrecy-and-Cloud-Computing.pdf

McKinsey. (2020). The executive’s AI playbook. Retrieved 19/10/2022, from https://www.mckinsey.com/business

-functions/mckinsey-analytics/our-insights/the-executives-ai-playbook?page=industries

McKinsey. (2022). Artificial Intelligence in Business. Retrieved 19/10/2022, from https://www.mckinsey.com/

capabilities/quantumblack/our-insights/the-executives-ai-playbook?page=industries/banking/

Molenaar, C. (2022). Demand-driven Business Strategy: Digital Transformation and Business Model Innovation.
Routledge.

Open Banking. (online). UK’s Open Banking to launch on 13 January 2018. Retrieved 19/10/2022, from https://
www.openbanking.org.uk/news/uks-open-banking-launch-13-january-2018/

Pichai, S. (2018). AI at Google: our principles. The Keyword (Google blog). Retrieved 05/12/2022, from https://
blog.google/technology/ai/ai-principles/

PwC. (2017). Sizing the price. PwC’s Global Artificial Intelligence Study: Exploiting the AI Revolution. Retrieved
19/10/2022, from https://www.pwc.com/gx/en/issues/data-and-analytics/publications/artificial-intelligence-study
.html

Roeder, J., Palmer, M., & Muntermann, J. (2022). Data-driven decision-making in credit risk management: The in-
formation value of analyst reports. Retrieved 19/10/2022, from https://www.sciencedirect.com/science/article/pii/
S0167923622000410

Schroeck, M., Shockley, R., Smart, J., Romero-Morales, D., & Tufano, P. (2012). Analytics: The real-world use of big data.
Retrieved 19/10/2022, from https://www.bdvc.nl/images/Rapporten/GBE03519USEN.PDF

Sigrist, F., & Hirnschall, C. (2019). Grabit: Gradient tree-boosted Tobit models for default prediction. Journal of Banking
& Finance, 102, 177–192.

SNB. (2022a). SNB Annual income statement. Retrieved 19/10/2022, from https://data.snb.ch/en/topics/banken/
cube/baerfrechua?fromDate=2010&toDate=2021&dimSel=D0(ERZ1,WBZ,BEZ1,ZDK,ZDH,ZDF,ZAU,ERK1,KWA
,KKG,KDL,KAU,ERH,UER1,ERV,BER,LER,AOE,AOA,GAU1,PAF1,GEH1,SOL,WAV,UEB,SAF,GER,AEG,AAU,VRB,STE
,EGV),BANKENGRUPPE(A30)

SNB. (2022b). SNB data portal. Retrieved 19/10/2022, from https://data.snb.ch/en

SNB. (2022c). SNB Securities holdings in bank custody accounts, by security category. Retrieved 19/10/2022, from
https://data.snb.ch/en/topics/banken/chart/bawebekatach

Swiss Bankers Association. (2020a). Cloud Guidelines. Retrieved 19/10/2022, from https://www.swissbanking.ch/
_Resources/Persistent/a/4/f/b/a4fb72f65d7f039abe516a40a078554523c31f90/SBA_Cloud-Guidelines_2020
_EN.pdf
23 Data-driven Banking

Swiss Bankers Association. (2020b). Data strategy –the path to digitalself-determination. Retrieved
19/10/2022, from https://www.swissbanking.ch/en/news-and-positions/news/datenstrategie-der-weg-zur-digitalen
-selbstbestimmung

Swiss Bankers Association. (2020c). Open Banking: An overview for the Swiss financial cen-
tre. Retrieved 19/10/2022, from https://www.swissbanking.ch/_Resources/Persistent/8/8/2/8/
88286724aa4fdc3dd8bb8ecd9b9c0d7af659a803/SBA_Overview_OpenBanking_en.pdf

Swiss Bankers Association. (2022). Swiss Banking Barometer 2022. Retrieved 19/10/2022, from https://publications
.swissbanking.ch/banking-barometer-2022/

The Economist. (2022). Banking on a game-changer: AI in financial services. Retrieved 19/10/2022, from https://
impact.economist.com/perspectives/sites/default/files/aiinfinancialservices.pdf

ti&m. (2022). Account opening process screenflow including KYC. (unpublished)

Walder Wyss. (2019). Permissibility of disclosure by Swiss banks of bank client information to agents in foreign countries
under article 47 of the Banking Act (BA). Retrieved 07/12/2022, from https://www.walderwyss.com/user_assets/
publications/190216-SBVg-Gutachten-Auslagerung-durch-Banken-V020b-EN.pdf

Wu, C.-S. M., Garg, D., & Bhandary, U. (2018). Movie recommendation system using collaborative filtering. In 2018
IEEE 9th International Conference on Software Engineering and Service Science (ICSESS) (pp. 11–15).