WHITE PAPER

Transformation’s Edge:
The State of GenAI in Global
Financial Institutions
A BCG benchmarking survey

April 2024
By Ella Rabener, Stiene Riemer, Michael Widowitz, Michael Strauss, and Cindy Tan
Benjamin Rehberg, Anne Kleppe, Jeanne Bickford, Pim Hilbers, Kentaro Ogata, Jon Sugihara, Kevin Lucas, Marianna Leoni,
and Nina Kataeva
Transformation’s Edge: The State of
GenAI in Global Financial Institutions
Key survey insights:

• 85% of financial institutions in our survey believe GenAI will be highly disruptive or
transformational.

• But only 2% have a fully developed GenAI talent strategy.

• Just 26% are actively investing a significant proportion of their innovation budgets in GenAI
implementation.

• Almost three quarters of survey respondents are in the early stages of use case development.

• The most progressed GenAI use cases focus on boosting internal productivity, rather than
re-shaping critical functions or inventing new business models.

Introduction

D
ecision makers at some of the world’s leading financial institutions (FIs) believe Gen-
erative AI (GenAI) presents a transformative business opportunity, according to new
survey from BCG. But only a few have made significant progress in pursuing that
vision - for instance, by establishing delivery teams or developing detailed plans for use
cases. Indeed, many FI leaders say more groundwork is needed to assemble the tools and
capabilities that would foster a winning GenAI proposition. Acquiring specialized talent is a
key priority. Where FIs are using GenAI in practice, it is most often to serve support functions
such as call center services or software development, rather than transform business-critical
operations at scale.

These are some of the findings of BCG’s inaugural GenAI in Financial Institutions Bench-
marking Survey, designed to offer market participants an inside track on how FI leaders view
one of the most talked about technologies of the decade.

1 TRANSFORMATION’S EDGE: THE STATE OF GENAI IN GLOBAL FINANCIAL INSTITUTIONS

 Focusing on five distinct pillars of GenAI readiness (overarching strategy, use case
deployment, tech stacks and data architecture, the talent gap, and responsible AI
policy frameworks) the survey asked FI leaders in North America, Europe, the
Middle East, Latin America, and Asia Pacific to share their insights on the tail-
winds and headwinds that are shaping their GenAI journeys.

Taken together, these create a first-of-its-kind perspective on the progress of financial sector
GenAI adoption around the globe, rooted in FI’s own assessments of their achievements.

While our respondents have made it clear that GenAI adoption is generally a work in progress,
BCG’s experience shows what is possible when the technology is successfully (and strategical-
ly) implemented. In more than 170 global client cases across multiple industry sectors, we
have seen benefits including a 10-fold reduction in customer enquiry costs, a 25% decline in
time spent creating marketing content, and a 30% jump in content creation productivity – all
while improving customer satisfaction and accelerating issue resolution. All told, these add up
to significant potential impacts, if institutions can find the right formula for scaling.

One of our aims in launching our benchmarking survey is to create a tool for decision mak-
ers to gauge where they are on their GenAI journeys, and to highlight the issues, risks, and
opportunities shaping the technology’s evolution at a critical time in financial services. In
this article, we drill down into the inaugural survey’s insights and briefly offer our thoughts
on its implications for FI decision makers as they consider how to engage.

BCG’s experience in supporting GenAI

• Advising clients regarding AI has been a top BCG priority since 2015. Starting in
late 2022, Generative AI (GenAI) became a topic top of mind for CEOs and other
senior leaders across all sectors. In 2023, BCG performed hundreds of GenAI proj-
ects for industry-leading clients.
• BCG X, the innovative technology build and design unit of BCG, has unique capa-
bilities in AI/GenAI, bringing together more than 3,000 BCG tech experts. BCG X
has nearly 50 patents and patents pending designed to deliver integrated, end-to-
end solutions that maximize impact at scale.
• BCG is collaborating with many of the world’s leading global technology com-
panies, including AWS, Google, IBM, Microsoft, Salesforce, SAP, and a number
of AI-centered enterprises such as OpenAI, Anthropic, Articul8, LangChain, and
Palantir, to help clients strategically deploy AI/GenAI and digital solutions.
• In 2023, BCG invested heavily to train the firm’s consultants to harness AI/GenAI
internally, including on a range of proprietary tools for knowledge management
and content creation.

BOSTON CONSULTING GROUP 2

 Pillar 1: Few firms are GenAI ‘adoption ready’

While 85% of surveyed FI leaders expect the impact of GenAI to be either highly disruptive or
transformational, most are still in the early stages of articulating a vision, implementing a
dedicated operating model, and defining KPIs. (See Exhibit 1). Indeed, just 18% of respon-
dents say they have a well-defined GenAI strategy that they have started to implement.

Despite the generally significant gap between ambition and reality, larger banks1 appear to
be ahead of their smaller peers in developing their thinking. Some 63% of the large-bank
group say they have a well-defined GenAI strategy in place, and the same percentage are
investing a “significant proportion” of their innovation budgets.

Across our survey group, fewer than one in ten institutions (7%) have launched delivery
teams or developed accompanying structures and KPIs. This suggests that most are still
working to understand the technology’s potential, or are reluctant to dedicate the time,
money, and team resources to ramp-up delivery capabilities or establish detailed KPIs. Again,
larger institutions are ahead of the curve, with 63% saying they have set up GenAI teams and
are in the process of implementation.

One common challenge facing our survey respondents is in identifying where GenAI can add
most value. A full 74% of respondents say they find it “moderately or very challenging” to
assess the potential of individual use cases. In a similar vein, 85% say that they find it “mod-
erately or very challenging” to define relevant or actionable KPIs. One reason for these head-
winds may be that most institutions are still at a relatively low point on the investment
curve. In contrast to larger banks, just 26% of the wider cohort say they are investing a signifi-
cant part of their innovation budgets, with only 18% having drafted a detailed list of priori-
tized use cases supported by a clear view on sequencing.

Exhibit 1 - GenAI is expected to be highly transformational, but

institutions face a significant readiness gap
Of respondents believe GenAI
85% will be highly disruptive or
transformational to the industry

GenAI
Readiness
Gap 18%
Have a well-defined GenAI
strategy with implementation
7%
underway Defined and implemented
GenAI delivery teams / squads
with clear structure and KPIs

Source: BCG GenAI in FI Benchmarking Survey; n=62.

1. We define large banks as those with assets exceeding US$800 billion and with more than 10,000 FTEs.

3 TRANSFORMATION’S EDGE: THE STATE OF GENAI IN GLOBAL FINANCIAL INSTITUTIONS

What we think
Many financial institutions have made a welcome start on getting ready for the transforma-
tive impacts of GenAI. However, without a strategy signed off by the wider leadership team,
they risk missing out on many potential benefits. In fact, a scattergun approach could com-
pound operational inefficiencies, with some parts of the organization potentially on a differ-
ent capability plane to others. And without a clear strategy and consistent communication
across the organization, businesses face major challenges in identifying and measuring
potential value. This, in turn, may dampen appetite for experimentation, delay investment
decisions, and stall execution capabilities.

To forestall inertia, we recommend a more holistic approach, based on a view of how the
technology could transform key value chains. Once that view is established, FIs can make the
required changes to tech stacks and data architecture, think about how to address the talent
gap and employee upskilling, and ensure they have established a holistic set of Responsible
AI policy frameworks and systems.

Pillar 2: Most Institutions are experimenting with GenAI use case

development and implementation

GenAI was one of 2023’s buzziest technologies and we expect it to remain high on C-Suite
agendas over the coming year. But its impact can only be accurately measured once it is fully
embedded in business workflows. For now, institutions are experimenting with use cases
focused on everyday tasks, rather than thinking about transforming critical functions. Among
the most prominent use cases are those aiming to boost capabilities in internal productivity
(most actively), customer services, including engagement and retention, and customer acqui-
sition. (See Exhibit 2).

The majority of institutions we spoke to are at proof of concept (PoC) stage or are preparing
to launch PoCs. For example, in the case of IT/Engineering co-pilots for de-bugging and soft-
ware development, about 21% of respondents are still in initial discussions or prioritising this
use-case, with about 25% having established a PoC. Interestingly however, only 6% have
managed to implement their solutions at scale. Similar proportions apply to other productivi-
ty-focused use cases, for example relating to generating insights from summaries and do-
main-specific data.

Given their resources, it is not surprising that larger banks are ahead on use case develop-
ment, with 63% reporting that they are largely adhering to set timelines for GenAI use case
implementation, compared with 33% of respondents overall.

When it comes to customer service, financial institutions in our study are most advanced in
applications to support agents with GenAI-enabled co-pilots, ahead of use cases to serve
inbound customer enquiries, sentiment analysis, and personalization. (See Exhibit 2). Still, a
relatively sizable proportion of institutions (10–15%) are in the process of rolling out inbound
customer service chatbots at scale – thereby providing an excellent illustration of GenAI’s
capabilities. That is, the models are adept at finding and synthesizing information from
various (unstructured) data sources, making sense of written and spoken language, and
suggesting high-quality replies that often require only minimal edits. On the other hand,
there are limits to these endeavors: unlike in some other industries (retail, consumer goods),
full customer-facing personalization experiences are relatively underdeveloped, with more
than 90% of institutions still in the planning stage for PoC development.

BOSTON CONSULTING GROUP 4

Exhibit 2 - GenAI use cases focused on everyday tasks rather than
reshaping critical functions or inventing new business models
Establishing a Proof-of-Concept is (relatively) easy, but implementing at scale is proving more challenging

Internal Productivity
IT/Engineering co-pilot for
de-bugging & accelerated
development

Generating insights and

summaries from domain-
specific enterprise data

Generating optimized code

through automated reviews

Agenda and notes co-pilot for

effective meeting coordination

Generating scenarios for

software testing and QA via
synthetic data

Customer Service, Engagement and Retention

Co-pilot for customer service
agents to support with calls,
chat and email

Inbound customer service

chatbot to manage
questions & complaints

Sentiment analysis across inbound

customer communication
to detect churn risk

Hyper-personalized content
development for existing
customer engagement

Customer Acquisition
Hyper-personalized marketing
content development
(new customers)

0 10 20 30 40 50 60

N= respondents selecting this use case

# respondents implemented at scale # respondents PoC established # respondents initial discussions

# respondents scaling in progress # respondents prioritized # respondents not selected

Source: BCG GenAI in FI Benchmarking Survey; n=62.

Note: Chart shows responses selected across all 62 respondents incl banks, asset & wealth managers and payment companies.

5 TRANSFORMATION’S EDGE: THE STATE OF GENAI IN GLOBAL FINANCIAL INSTITUTIONS

What we think
In BCG’s experience, isolated delivery of PoCs stops misses an opportunity for accelerated
capability development across the organization. To unlock more value, businesses leaders may
wish to consider a structured approach to scaling, based on a three stage-gate framework:

• Deploy GenAI in everyday tasks. We have seen that many leading institutions focus
their deployment activities on productivity initiatives, including making meeting summa-
ries, coding, processing documents, and generating reports. In this context, the strongest
headwinds are realizing measurable value and working out which use cases to prioritize.
BCG recommends deployment in concentrated teams to generate early wins and use case
prioritization based on business case and pilot learnings.

• Reshape critical functions. Forward-looking institutions concentrate their efforts on

boosting their productivity and increasing the speed and quality of their outputs across
the value chain. Examples include transformations of the marketing and sales function,
systematic customer service enablement across channels, and empowerment of relation-
ship managers in both wealth and corporate banking to deliver more value to clients – and
spend less time on administrative or low-value tasks. Compliance monitoring is also a
growing area of interest. Among institutions we have worked with, common challenges
include reimagining the end-to-end transformation required, as well as the necessary
systems, process changes, and trainings. Finally, many institutions are unsure whether to
build or buy (or a combination of both), and face challenges in integrating new systems
with existing architectures.

• Invent new business models. We believe that institutions could add significant value
where they launch new Gen-AI powered business models focused on the customer experi-
ence, as well as platforms to support innovation and insight, or new financial products and
offerings. Challenges to these aspirations include finding the right combination of product,
market, and user experience to drive customer adoption, and identifying economics that
work at scale, as well as managing latency issues (the time delay between a user action
and a system response). Leveraging GenAI to overcome these hurdles could unlock new
business models, including value-based pricing.

Pillar 3: GenAI-ready tech and data stacks are being held back by
integration challenges

Like many tech innovations, GenAI needs to be able to work in concert with existing systems.
But that’s not easy to achieve. In fact, the majority of FIs in our survey say that integration
with existing infrastructure is their number one technology challenge, followed by ensuring
data security and privacy in their platform selection. Indeed, just 34% of respondents say
they are “well underway” in implementing their GenAI tech stacks, and just 10% say they
have made “significant progress” in upgrading data architecture to be GenAI-ready.

FIs have made progress on selecting large-language model provider(s), with 71% of respon-
dents saying they have made their choices. Still, only a half are implementing the necessary
infrastructure changes, and only one in three have adjusted data architecture as required.
(See Exhibit 3). The task ahead will be to accelerate the processes of marrying legacy assets
with new capabilities.

What we think
Organizations can benefit from setting up a GenAI ‘Centre of Excellence’ (CoE). This would
be a centralized structure that could support a consistent change management process as
GenAI use cases are developed and scaled. Winning organizations furnish the CoE with a
broad remit, including being able to define data enterprise architectures, select strategic
partners, and keep pace with emerging tech developments, vendors, and regulation. This
provides the best chance of effectively integrating large language models whilst ensuring a
consistent approach aligned with risk and regulatory guardrails throughout the business.
Over time, it may be appropriate, for larger institutions at least, to decentralize GenAI capa-
bility development and execution while still enabling a central oversight function. This would
help business units or dedicated incubation teams to experiment and deploy faster.

BOSTON CONSULTING GROUP 6

Exhibit 3 - The majority of respondents have settled on their LLM1
provider(s) and started deployments, but are taking longer to make
enterprise-wide technical changes

71%
Of respondents have settled
on their LLM1 provider(s)
and started deployments
48%
Of respondents have started
implementing required technical
37%
infrastructure changes Of respondents have started
implementing data
architecture changes

Source: BCG GenAI in FI Benchmarking Survey; n=62.

Large Language Model.
1

Pillar 4: FIs are overlooking the chance to future-proof their GenAI talent
strategies

Perhaps not surprisingly given GenAI’s nascent status, not a single respondent to our survey
says they have conducted an organization-wide assessment of GenAI’s potential impact on
talent and people. Moreover, 55% say they will not embed a GenAI talent strategy before
2025, of which 30% lack any target timeline at all. (See Exhibit 4). That said, there are signs
that decision makers are moving in the right direction, with 73% having started some form of
employee upskilling and 43% recognizing they will need to hire experts in future. At the
same time, 63% of large banks (compared with 37% of total respondents) have conducted a
partial impact assessment of their people and processes.

The capability gap is also manifested in ways of working, with just 7% of respondents saying
they have established mechanisms, feedback loops, and improvement cycles to understand
and promote GenAI usage. Once organizations move to wider adoption, these shortfalls in
readiness are likely to be a constraint on effective GenAI implementation, and on achieving
the productivity gains that many of our respondents believe are possible.

What we think
GenAI offers significant potential for institutions to boost productivity, but that is predicated
on strategic commitment at all levels of organizations. FIs must be willing to transform their
operating models end to end, and importantly invest in sufficient, dedicated talent resources.
Looking ahead, BCG estimates that GenAI will impact almost two thirds of the workforce,
with new roles emerging that can help institutions across the four key dimensions of build-
ing, shaping, using, and governing.

Winning talent strategies require excellent foundational steps. That means starting with an
impact assessment across the workforce, helping institutions speedily identify where they need
to make process changes and launch initiatives for up-skilling and reskilling. If companies can
roll out solutions to roles at all levels of the business, they will be best placed to graduate from
a series of targeted GenAI pilots to a coherent, organization-wide transformation.

7 TRANSFORMATION’S EDGE: THE STATE OF GENAI IN GLOBAL FINANCIAL INSTITUTIONS

Exhibit 4 - Majority of players underestimate the importance of talent
strategy in the GenAI race

Most respondents will not have their GenAI talent strategy in place before 2025;
one third have no clear timeline in place at all

Talent strategy fully

implemented for now

No target for
the next 2 years
2% Within 6 months
30%
13%

% of overall
survey
respondents,
single select

30%
25% Within 12 months
Within 24 months

Source: BCG GenAI in FI Benchmarking Survey; n=62.

Pillar 5. Responsible AI and risk policy frameworks are still under construction

When it comes to GenAI policies, most financial institutions in our survey are still working to
embed best practice. Just 13% have implemented a company-wide Gen-AI policy and practic-
es framework, while 23% have no target timeline within the next three years. This indicates
that a general lack of preparedness extends into strategic agendas. Still, the industry has
made more progress in developing individual risk policies and guidelines. For example, 74%
of respondents have data privacy and security policies in place, while 52% have drafted ethi-
cal guidelines for AI use and 40% have created AI regulatory and compliance frameworks.
Only 18% confess to having no risk policies at all.

As with the other pillars discussed here, large banks are generally ahead of their peers. Indeed,
88% of that group have either implemented a holistic GenAI policy and practice framework, or
intend to do so within the next 12 months. That compares with 57% of the wider cohort.

Even where companies have put in place guardrails for GenAI adoption, many institutions
continue to struggle with implementation. A significant 87% of survey respondents say that
managing risk in that context is a top three challenge. The other two are enforcing alignment
with GenAI risk and policy guidelines, cited by 82% of respondents, and ensuring compliance
with data protection and privacy regulations, cited by 78%. Meanwhile, just 13% are monitor-
ing global trends and working with regulators to shape new rules – a unique window of op-
portunity that will likely narrow in the medium term.

BOSTON CONSULTING GROUP 8

 What we think
Why do policies and practices matter? Our experience across numerous client engagements
shows that experimentation without a Responsible AI (RAI) framework tends to increase
institutional exposure to blind spots in process efficiency, quality checks, and risk manage-
ment. Thus, an RAI framework has a vital role to play, not just in expressing ethical values
but in reducing the risk of potentially costly errors, reputational damage, and regulatory
penalties – reflecting policymakers’ heightened awareness of potential financial industry
conduct issues.

Effective RAI frameworks combine ethical guidelines with contingencies for the potential
impacts of technology-supported decision making, the demands of data privacy, and model
risk management, where the roots of risk often lie. The best-prepared institutions are taking
active steps to engage with regulators and shape regulation, ensuring they have skin in the
game when decisions are made.

GenAI offers FIs a transformational opportunity to streamline to their operations, sharpen

their insights, and innovate across their products and services. Most decision makers in our
inaugural benchmarking survey recognize this potential - but only a few are ready to take full
advantage. Many institutions are stuck in second gear, as they struggle to transition away
from individual use cases and toward a more orchestrated and holistic approach. The chal-
lenge is often exacerbated by a lack of change management readiness, and deficits in risk
management and guardrails.

T
he antidote to these challenges lies in adopting a more expansive lens, and a strategy
that reflects GenAI’s potential impacts across many aspects of FI operations. In this
context, the first task for leaders will be to pivot to a more active articulation of GenAI’s
potential across value chains. From there, they can start in earnest to unlock potential in each
of the five pillars and create lasting value on the bottom line.

Our methodology
BCG’s inaugural GenAI benchmarking survey was conducted in December 2023.
Some 68 clients across the Americas, EMEA, Latin America, and Asia Pacific pro-
vided a detailed self-assessment of their capabilities, priorities, and progress to
date. A significant 44 percent of respondents had 10,000 or more FTEs, while 26%
had 5,000 to 10,000, and 21% had 2000– 5,000. By region, 60% of respondents
were based in EMEA while 16% were in Latin America, 13% in APAC, and 11% in
North America. By sector, 15% were in specialty FI sectors, including asset and
wealth management and payments, while 85% were banks.

9 TRANSFORMATION’S EDGE: THE STATE OF GENAI IN GLOBAL FINANCIAL INSTITUTIONS

About the Authors
Ella Rabener is a Managing Director & Partner with BCG’s Berlin
office and Global Lead for GenAI in Financial Institutions with a broad
perspective on GenAI trends for the sector. She supported several
clients with GenAI use cases, as well as advising on operating model
changes required to drive GenAI adoption. Ella is also an experienced
FinTech entrepreneur, having co-founded one of Europe’s few FinTech
unicorns. [email protected]

Stiene Riemer is a Managing Director & Partner with BCG’s Munich

office. She leads the AI & GenAI developments for Financial Institutions
globally and she is the Founding Member and Global Lead of Smart-
Banking AI by BCG. She is an expert in developing transformative
analytics and AI solutions. Her recent work includes the design, devel-
opment and implementation of AI driven fraud detection models, as
well as ongoing transaction monitoring. [email protected]

Michael Widowitz is a Managing Director & Partner with BCG’s

Vienna office. He is an expert in AI, Big Data and Advanced Analytics
at BCG X. He holds extensive expertise in in implementation of fully
customized, end-to-end, analytics and AI solutions, with a primary
focus on Financial Institutions. Recent work includes AI Operating
Model and Strategy, GenAI Center of Excellence Setup, GenAI-supported
Service Operations, and Personalization with support of GenAI capa-
bilities. [email protected]

Michael Strauss is a Managing Director & Senior Partner with

BCG’s Cologne office. He is part of the global FI and BCG X Leader-
ship Teams. Michael led several (Gen)AI strategy projects and helped
clients with high impact GenAI use case development and implemen-
tation across various domains. [email protected]

Cindy Tan is a Partner with BCG’s London Office. She is a core mem-
ber of the Financial Institutions team at BCG X in EMESA, supporting
clients in the region to embed (Gen)AI in future products and client
experience strategies. She specializes in innovation and complex
product build, with a focus on wealth management, banking and
insurance. [email protected]

For Further Contact

If you would like to discuss this report, please contact the authors.

Acknowledgments

Thank you for the contributions of our colleagues Benjamin Rehberg, Anne Kleppe, Jeanne
Bickford, Pim Hilbers, Kentaro Ogata, Jon Sugihara, Kevin Lucas, Marianna Leoni Nina Kataeva,
Emily Harrison, Karolina Weyzk, Petr Vodak, Sonali Maheshwari, Deepali Katyal, Tanmay Pathak
and Amit Sukhija.

10 TRANSFORMATION’S EDGE: THE STATE OF GENAI IN GLOBAL FINANCIAL INSTITUTIONS

Boston Consulting Group partners with leaders in business For information or permission to reprint, please contact
and society to tackle their most important challenges and BCG at [email protected]. To find the latest BCG con-
capture their greatest opportunities. BCG was the pioneer tent and register to receive e-alerts on this topic or others,
in business strategy when it was founded in 1963. Today, please visit bcg.com. Follow Boston Consulting Group on
we work closely with clients to embrace a transformational Facebook and X (formerly Twitter).
approach aimed at benefiting all stakeholders—empowering
organizations to grow, build sustainable competitive © Boston Consulting Group 2024. All rights reserved. 4/24
advantage, and drive positive societal impact.

Our diverse, global teams bring deep industry and functional

expertise and a range of perspectives that question the
status quo and spark change. BCG delivers solutions
through leading-edge management consulting, technology
and design, and corporate and digital ventures. We work
in a uniquely collaborative model across the firm and
throughout all levels of the client organization, fueled by
the goal of helping our clients thrive and enabling them
to make the world a better place.
bcg.com