Accenture Age AI Banking New Reality

The age of AI:
Banking’s new reality

2
Embracing generative AI
for a competitive edge
Generative AI has taken the world by storm. to be impacted by generative AI—39% by
The rates of adoption and enhancement automation and 34% by augmentation. Only
are more rapid than any major technology 27% of employees’ time currently has a low
innovation in the history of human potential to be transformed.1 Similar metrics
development, and few doubt its potential have been observed in banks worldwide.2
to have a transformative effect on business
and society. It’s no surprise, therefore, that within a year
of the launch of ChatGPT, almost every
Banks are at the sharp end. Our analysis bank had started to explore and evaluate
indicates that—due to the importance of early use cases. The leaders in this field
language throughout the value chain—the have gone further: they are already seeking
industry has a greater potential to benefit to maximize the value they have generated
from the technology than any other. In fact, by scaling their implementation across the
we concluded that 73% of the time spent organization.
by US bank employees has a high potential
3
Our latest financial projections indicate that the gains over

the next three years will be substantial for the early adopters:
From this it becomes apparent that while generative AI is likely to dramatically improve the efficiency of the banking operating model,
its potential to differentiate and drive growth by enhancing the customer experience is what excites bankers the most.
22% to 30% 600 bps 300 bps

productivity improvement rise in revenue growth increase in return on equity
Source: Accenture
Research analysis
4
The business of banking has barely services. By the time banks have executed
changed in hundreds of years. At its heart, their generative AI strategy, they will have
it’s about taking and safeguarding deposits reinvented and modernized most parts of
and lending money. Generative AI—like the bank.
By the time banks
the internet and the smartphone that have executed
transformed customer engagement—will Achieving these goals will be neither easy
not change the fundamentals of banking. nor automatic. Consumers in their millions
their generative AI
But no one doubts that its impact on may already be using ChatGPT, Microsoft’s strategy, they will
the industry will be seismic. Thanks to Copilot, Google’s Bard and other models
its inherent ability to learn, advance and to good effect, but for organizations like
have reinvented and
create it will, over the next few decades, be banks to maximize the benefits a number modernized most
a driving force for continuous reinvention of obstacles need to be overcome. Given
across the enterprise. It will be widely the pace at which many are seeking to
parts of the bank.
deployed throughout the value chain, scale generative AI, it is important that
and will radically transform virtually every they plan their journey strategically and
facet of how banking gets done and how holistically.
customers experience their bank and its
5
Generative AI has a
role to play in every
part of banking
Our analysis of the potential impact of generative
AI on the banking industry3 concluded that every
role in every bank is likely to benefit in some way
from generative AI. Through this study and our
ongoing work with leading banks worldwide, we
have identified hundreds of promising use cases
that span the banking value chain.
From the back and middle offices through to

the tellers, advisors, relationship managers
and contact center agents in the front office,
the ability of generative AI to automate routine
manual tasks and augment workers’ capabilities
will make a profound difference (see Figure 1
on page 7).
6
Generative AI will transform banking roles in different ways and to different degrees,
depending on the specific nature of their tasks and the time that each takes.
Automation Augmentation All-round support
In our analysis of US banks, we discovered Employees whose work involves a high We determined that 25% of all employees
that occupations representing 41% of measure of judgment, such as credit will be similarly impacted by both
banking employees are engaged in tasks analysts, or who need to understand automation and augmentation. Customer
with higher potential for automation. Roles customers’ needs and circumstances service agents, who spend their time
such as tellers, whose jobs primarily involve and personalize their interactions, such explaining products and services to
collecting and processing data, would as relationship managers, could be customers, responding to inquiries,
benefit greatly from automation—60% of empowered by generative AI tools that preparing documentation and maintaining
their routine tasks could be supported by help them prepare for and run meetings sales and other records, are a good
generative AI. —34% of banking employees fall into this example. Of these tasks, 37% could be
category. automated while 28% could be augmented.
7
Figure 1: How the top 20 banking industry roles are like to benefit from generative AI.
Industry
60% average = 39% Front office Middle & back offices
Personal Financial
Advisors Credit Analysts
50% Securities, Commodities,

Higher potential for augmentation
and Financial Services Sales Agents
Software Developers
Market Research Analysts Financial Managers
40%
and Marketing Specialists
Loan Officers Financial Examiners Industry
average = 34%
Loan Interviewers and Clerks Tellers
30%
Customer Service Representatives
Financial and
Investment Analysts Bookkeeping, Accounting,
First-Line Supervisors of and Auditing Clerks
20%
Non-Retail Sales Workers
Office Clerks and General New Accounts Clerks
Accountants and Auditors
First-Line Supervisors of Office
10% and Administrative Support Workers Bill and Account Collectors
Management Analysts
0%
15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75%
Higher potential for automation Bubble size: Relative
number of employees in US
Source: Accenture Research analysis and estimates on (Process work)
BLS and O*Net data as of December 2022.
8
In the back and middle offices, generative AI will be The same will happen in the front office, but here Productivity is the most obvious benefit of
used to transform the operating model. Many tasks employees will have access to intelligent tools that widescale adoption of generative AI. However, the
will be automated. This will improve speed and allow them to personalize customer experiences ability to better understand and meet customers’
accuracy, reduce costs and relieve employees of and sell more effectively. With both the time and needs, thereby increasing satisfaction, retention
the more tedious aspects of their jobs. By allowing the means to engage more meaningfully with and share of wallet, is likely to make a much bigger
them to take on more or other tasks that add value, customers, they will be able to restore the human contribution to banks’ bottom line.
they could have a bigger impact on the bank’s touch that was lost over the past two decades as
overall performance. banks digitalized their experiences.
59%
of banking employees
are already using AI
every day. 4
9
How generative AI
can be put to work
While the potential applications for generative AI
in banking are almost limitless, our experience
and analysis show there are mainly three ways
the technology is currently being employed: by
using tools in which generative AI is embedded
(e.g. email), by using the technology to transform
the operating model (e.g. call centers, code
development), and by using it to innovate and
differentiate the bank’s experiences and offerings.
10
Embed Transform Innovate and differentiate
Many software vendors whose platforms are used One of the most immediate ways banks could The greatest, most enduring impact of generative
by banks to run their business are incorporating put generative AI to work is to integrate it with AI will likely be in equipping banks to innovate and
generative AI into every aspect of what they do. middle- and back-office operations to drive differentiate their products, marketing and customer
For example, Microsoft began integrating large efficiency and effectiveness gains. Just one interactions. On the product side, banks are using
language models (LLMs) into its Microsoft 365 example is the transcription and summarization generative AI to produce thousands of scripts that
suite of apps back in March 2023 with the launch of customer call recordings. Generative AI could are tailored for individual customers. In marketing,
of Copilot.5 Adobe’s Firefly tool can generate also enable transformation that has been put off they are beginning to adopt the technology to
images from simple text prompts.6 Salesforce due to financial or talent constraints—such as achieve levels of personalization which, until now,
offers a CRM assistant called Einstein that gets core system modernization. It is still early days, have been economically impossible. They are
its intelligence from generative AI,7 and Workday but we are seeing some banks use generative AI combining internal and external customer data
recently started integrating the technology into its to dissect and reverse-engineer their legacy code, with behavioral economics to generate curated
tools.8 All of these are intended to both automate and rewrite it in a modern language. Westpac, for experiences similar to that of the latest vehicle
and augment banking tasks and roles. example, is pairing its engineers with a generative sat-nav systems. Customer intent has become
AI companion to help fast-track software more apparent, allowing banks to become more
development projects, resulting in a 20%+ empathetic, proactive and relevant. The ability to
increase in code written by its programmers.9 tailor customer interactions, recommendations and
pricing may very well be the most important benefit
banks gain by using generative AI.
Banks are understandably cautious about the reputational and other risks associated with this leap in innovation. However, given the opportunity
to reinvent their customer experiences and drive growth, most are working hard to ensure they take advantage in a responsible way.
11
How to lead
in the era of
generative AI
Accenture research10 shows that banks have
Many banks are asking how they can unlock identified a number of factors that are key to
the incredible potential of generative AI. the success of their adoption of generative AI.
The key is to develop a holistic strategy Their priorities are:
that identifies the most promising use
cases, but then commits to moving beyond
isolated proofs of concept to scaled,
responsible deployment in a way that is
36% 46%
aligned with the bank’s business goals and cloud infrastructure data strategy
reassures regulators.
34% 25%
talent acquisition overcoming worker
resistance to change
12
Our conversations with industry front-runners

reveal common themes.
01 Lead with value
We see five key imperatives which C-suite

02 Understand and develop a secure AI-enabled digital core
executives should address to reinvent in the

age of generative AI. We have also identified 03 Reinvent talent and ways of working
the key steps for each that will help your bank
become a leader.
04 Close the gap on responsible AI
05 Drive continuous reinvention
Understand and develop a secure Reinvent talent and Close the gap on Drive continuous Measuring the ROI
Lead with value
AI-enabled digital core ways of working responsible AI reinvention of generative AI
13
I M P E R AT I V E O N E
1
Lead with value
Since there are more potential use cases for
generative AI than any bank could possibly explore
at any one time (see Figure 2 next page), the big
question is not what to do but rather what not to
do—and therefore, how to prioritize adoption.
Lead with value
14
Figure 2: Leaders are moving forward on use case development from front to back office.
Unsecured Secured Wealth

Deposits Commercial Banking
Lending Lending Management
Front-Back Book Offer Intent Treasury

Transaction Banking Trade Finance
Management Investment Mgmt.
Optimization Optimization Identification
Commercial Commercial
Intelligent RM Real Estate Credit Origination Advisor Mgmt.
Commercial Advisory Services
Products
Products Product Development
Lead Generation & Optimization Account Servicing & Pricing
Portfolio
Underwriting / Credit Assessment Optimization
Application Processing & Fulfilment Trading
Lead Nurturing & New Offer Brand Brand Marketing Segmentation & Client & Sales Digital Content
Sales / Marketing Lead Origination Qualification Management Management & Campaigns Targeting Insights Creation & Mgmt.
Client Engagement Customer Experience

& Value Mgmt. Loyalty Program Correspondence Branch & Advisor Call Center Client Tools & Portals Client Onboarding Channel Mgmt.
IT Planning & Application Doc. & Knowledge Data Sourcing Data Structuring BI Reporting &
Technology & Data Coordination IT Engineering Enterprise Testing Management Management Data Governance Strategy & Processing Self-Serve
Deposit, Cash & Collections & Account & Investment Fund Services & Custody & Asset
Operations Payments Mgmt. Default Mgmt. Cards Mgmt. Portfolio Mgmt. Fraud Management Management Administration Safekeeping
Risk, Compliance & Financial Crime &

Audit Compliance Risk Management KYC Finance Procurement HR Legal
Enterprise AML
Collaboration & Knowledge Mgmt. & Marketing Enterprise Resource Supply Chain Business Process
Enterprise Applications Commun. Software Doc. Mgmt. System CMS CRM
Automaton Project Management Planning Management Management
Estimated impact of generative Al Low impact Medium impact High impact

Source: Accenture analysis
Lead with value
15
The key is to balance the need for rapid Banks can achieve the necessary balance by
diffusion throughout the organization with doing three things simultaneously:
the accompanying cost and with the relevant
• Lead with top-down support and funding for
regulatory requirements. Some use cases are
the prioritized initiatives;
simple and relatively inexpensive while others,
like building a digital twin of the bank’s mortgage • Establish an operating and steerco model that
function, are complex and require a lot of ensures adoption and is compliant with all
expertise, data and computation. With regard to relevant regulations;
computation, the rapid decline in the cost of using
most generative AI models makes prioritization • Drive multi-speed implementation and
especially challenging. Many banks are starting adoption across business segments, functions
their generative AI journey with simple, no-regrets and enterprise applications.
applications while planning the timing of their
more complex initiatives by calculating where the
By broadening the scope beyond single
shifting cost and return curves are likely
applications, banks can integrate generative AI
to intersect.
more holistically into their value chains, leading
to transformative improvements across business
functions. However, this broader integration
requires strong C-level sponsorship and a broad
business strategy, all underpinned by a robust
governance mechanism.
Lead with value
16
• Developing and enforcing standardized which projects to scale, and how to optimize their
Establishing a strong strategy and oversight approaches, assets, best practices and transformation journey. It will also help gain the
team is critical. This team should include leaders principles for the deployment of solutions. support of everyone in the bank for the holistic
from the business, risk and technology sides of adoption of generative AI.
the bank, and its mandate should encompass • Establishing the frameworks and approaches for
strategy, policy, talent, technology, regulatory model risk management, to ensure compliance Some banks may consider having separate
compliance and data. with not only the law but also corporate infrastructures for generative AI and traditional
governance standards and requirements. AI / data functionalities. However, this could
One of its first priorities should be the cause strategic conflicts and make it more
establishment of a generative AI center of • Supporting vendor assessments. difficult to capture efficiencies. A single
excellence (CoE). This dedicated group will focus structure will drive the initiative from a unified
• Assessing the talent impact and supporting
on generating business value by implementing the platform and facilitate synergy.
change management and upskilling efforts to
bank’s generative AI strategy and cross-pollinating
minimize disruption and encourage adoption. Not all implementations are economically viable
the technology throughout the bank. It will
prioritize use cases, clear the way for generative at this time. However, the overall trend in
A vital and ongoing role of the CoE would be
AI to be scaled up in a federated model, and generative AI implementation and consumption
to accurately measure the ROI of the bank’s
catalyze innovation. The responsibilities of the is towards lower costs and greater feasibility.
generative AI applications. This should not be
CoE could include: The challenge for banks is to position themselves
limited to immediate cost and revenue gains
to capitalize on new use cases as they quickly
but should also consider long-term strategic
• Collaborating with the business units to become economically feasible. This will require a
benefits (see Measuring the ROI of generative AI).
develop proofs of concept and roll out the strategic approach to prioritization, focusing on
A clear and empirical view of these benefits will
successful ones throughout the organization. current objectives while keeping an eye on the
help banks decide where to allocate resources,
longer-term investment horizon.
Lead with value
17
CASE STUDY
European bank expands its CoE to scale AI benefits

A leading European bank started to build an AI they have skin in the game. A value assurance
CoE six years ago to ensure strategic alignment group, composed of planning, HR, legal and other
and facilitate allocation of resources for AI departments, assesses proposed use cases for
projects and programs. The CoE, with sponsorship their potential value.
from the CEO, CDO and CFO, is set to expand
from a small structure into a 300-strong team This approach has created a fertile environment
over the next three years. It has responsibility for for cross-unit collaboration. The C-suite has
scaling AI and generative AI use cases and setting enough confidence in the goals of the AI projects
technical standards. currently underway to have committed publicly
to increase the bank’s 2025 operating income by
Operating in a hub-and-spoke model, the CoE hundreds of millions of euros.
ensures economies of scale as well as the
consistency and quality of AI and generative
AI applications. Business units are responsible
for proposing use cases, securing funding and
quantifying the value derived from AI. They
retain ownership and accountability to ensure
Lead with value
18
Leading with value
ACTIONS
• Develop a comprehensive generative AI • Implement rigorous ROI measurement
integration strategy across all business protocols that include both quantitative
functions, moving from isolated use cases to a financial metrics and qualitative assessments of
more holistic and connected approach. strategic impacts such as customer satisfaction
and competitive differentiation.
• Establish a strong C-level governance
framework to ensure that generative AI • Monitor the market as the capabilities of
initiatives are aligned with the organization’s generative AI models evolve and the cost of
strategic goals and effectively integrated using them drops, changing the business case.
across departments.
• Foster cross-departmental collaboration to

break down silos, facilitating a unified approach
to generative AI implementation that leverages
diverse expertise and insights.
Lead with value
19
I M P E R AT I V E T W O
2
Understand and develop a
secure AI-enabled digital core
Firms that showed early success in developing
a strong digital core and data foundation have Many banks’
been leading in generative AI. A modern digital
digital architecture,
core leverages the powers of cloud, data and AI
through an interoperable set of systems across the infrastructure and data
bank—including enterprise platforms, automation, capabilities are likely to
integration and security—that allow for rapid
development of new capabilities. This core, with its impede their successful
architectures, infrastructure, capabilities and talent, adoption of generative
is essential to making the most of generative AI—
whose capabilities are most often cloud-based. AI at scale.
Lead with value
20
However, our analysis of the current banking This spread highlights the fact that many LLMs is their ability to consume and work with
landscape reveals significant variation in the banks’ digital architecture, infrastructure and data huge volumes of data in different formats (images,
caliber of banks’ digital cores. This observation capabilities are likely to impede their successful text files, video and audio recordings, etc.) and
emerges from our global study of 240 banks. It adoption of generative AI at scale. This is confirmed covering multiple topics (from prospectuses
ranges broadly from 0.2 to 0.8 on a scale of 0 to 1, by our survey finding that 47% of executives across and policy documents to relationship managers’
with 82% of banks having a score between 0.2 all industries list ‘getting their data strategy right’ meeting notes). The problem is that much
and 0.6. as one of their greatest challenges as they strive to of it today is not only unstructured but also
implement and use generative AI.11 unorganized, unlabelled and dispersed
throughout the enterprise.
82% of banks have a score
between 0.2 and 0.6 Additionally, our research indicates that
54% approximately 35% of banks globally have For LLMs to work, all of this unstructured data
migrated less than 5% of their workloads to the needs vectorized databases. These currently exist
cloud. This is a substantial constraint, because the alongside traditional data lakes and enable the
evolution of generative AI is increasingly geared parsing and extraction of key information so that
28%
towards cloud-native technologies. Banks with LLMs can leverage it at speed. A key consideration
limited cloud integration are likely to miss out on is whether these databases will converge, and
12%
cloud-native AI functionalities. if so, how quickly and at what cost? Database
6%
0%
convergence will have a significant impact
0-0.2 0.2-0.4 0.4-0.6 0.6-0.8 0.8-1 The goal of scaling generative AI will impose new on banks’ ability to utilize the full potential of
Low High requirements on the digital core, so it’s important generative AI to produce precise and personalized
for banks to understand the status of their core outputs.
Figure 3: Variation in the caliber of banks’ tech
today. For example, one of the advantages of
development (0-1 scale)
Lead with value
21
Forward-thinking banks have an advantage The AI landscape is in a state of constant flux,

because they started migrating from data lakes marked by the continual emergence of new
to decentralized data meshes before generative models, enhanced capabilities and an expanding
AI was on the horizon. In such a structure, array of tools and providers. This dynamic
domains within a bank take ownership of their environment could make previously unviable use
data, including responsibility for data quality and cases feasible. This will require centralized but
accessibility. These domains manage and provide connected models management. This strategy
their data as products, making it easier for other will standardize critical capabilities, such as the
parts of the business to use it. selection and customization of foundational
models, allowing for their efficient and transparent
Regardless of the underlying LLM strategy integration across various business functions.
employed—be it licensing pre-built models,
adapting existing ones through retrieval-
augmented generation (RAG), fine-tuning or
developing models from the ground up—it
is crucial that the bank has a data strategy This dynamic environment
and approach that allows it to be flexible and could make previously
future-ready. This approach will be crucial in
gaining the best possible outputs.
unviable use cases feasible.
Lead with value
22
CASE STUDY
Scaling an enterprise-wide LLM and generative AI capability

A leading global financial services firm is capabilities. The bank has implemented a secure
developing an enterprise-wide AI strategy, and compliant platform that can host ChatGPT and
supported by a strengthened IT infrastructure OpenAI services.
and a transformative approach to talent. The bank
built a team of thousands of AI and data experts. This platform will enable the bank to scale
This AI strategy has, to date, brought in tangible generative AI and future-ready, cloud-neutral
benefits totalling around 3% of its expected net functionality across its business. By modernizing
income. its digital core, the bank not only kept up with
existing tech trends, it also positioned itself to
In its move towards a modern digital core, the integrate advanced AI with its core banking
bank focused on building a scalable multi-cloud operations at scale and keep one step ahead of
environment that complies with banking regulations. new technologies as they emerge.
This infrastructure has enabled it to fast-track its
adoption of next-gen, cloud-neutral generative AI
Lead with value
23
Understand and develop a secure AI-enabled digital core
ACTIONS
• Assess the complete inventory of unstructured • Start moving towards data-as-a-service models, • Test various architectures to find the best
data across the organization and analyze how it where employees and developers can access fit for each use case and plan the model’s
could help power generative AI. data from internal marketplaces to use in their lifecycle from experimentation to scaling and
own applications and tools. phase-out. This could be achieved through a
• Start to move this data into vector databases ‘model switchboard’ where banks can select a
and scale them with the precision required for • Elevate data governance standards to combination of models based on the business
real-time analytics and the unique demands of effectively manage unstructured data in context or on factors such as cost or accuracy.
generative AI. generative AI.
• Evolve security protocols to address the

complexities that come with diverse data
access.
Lead with value
24
I M P E R AT I V E T H R E E
3
Reinvent talent and ways of working
Generative AI will be pervasive, affecting virtually As work and roles are transformed, waste is taken
every process, task and role in the bank. To out and banks have the choice of boosting the
maximize its transformative impact, banks will bottom line or investing to increase value. Those
need a culture that not only anticipates but that opt for investment are using the savings
champions change. for talent to transition to newly created roles
such as managing generative AI, or in strategic
Key to achieving this impact is the strategic areas like business development or relationship
integration of generative AI into the bank’s management. These new and transitioned roles
processes. This goes beyond the adoption of new are aligned with the desired strategic customer
technology; it’s about reimagining workflows to and business outcomes. In the mortgage sector,
increase efficiency and innovation and aligning for example, we envisage a workforce that will
them with the goals of the business. include several new roles and has the increased
capacity needed to unlock significant productivity
gains (see Figure 4 next page).
Lead with value
25
Figure 4: An illustrative example of how work and roles in the mortgage business can be realigned with generative AI.
Today Tomorrow
Category Roles No. of People
Category Roles in the Role
^ĂůĞƐͬĚǀŝƐŽƌǇ Loan Officers Fewer
^ĂůĞƐͬĚǀŝƐŽƌǇ Loan Officers
^ĂůĞƐͬĚǀŝƐŽƌǇ Loan Officer Fewer
^ĂůĞƐͬĚǀŝƐŽƌǇ Loan Officer Assistants Assistants
^ĂůĞƐͬĚǀŝƐŽƌǇ Loan and Credit Same
^ĂůĞƐͬĚǀŝƐŽƌǇ Loan and Credit Managers Managers
KƉĞƌĂƚŝŽŶƐ Loan Processors ^ĂůĞƐͬĚǀŝƐŽƌǇ Loan Product More

Managers
KƉĞƌĂƚŝŽŶƐ Underwriters ^ĂůĞƐͬĚǀŝƐŽƌǇ Sales Prompt More
Engineers
KƉĞƌĂƚŝŽŶƐ Today
Loan Closers and Funders Tomorrow
KƉĞƌĂƚŝŽŶƐ Ops Engineers More
Category Roles No. of People
KƉĞƌĂƚŝŽŶƐ Category Roles
Customer Service in the Role
KƉĞƌĂƚŝŽŶƐ Loan Processors Fewer
^ĂůĞƐͬĚǀŝƐŽƌǇ Loan Officers Fewer
KƉĞƌĂƚŝŽŶƐ ^ĂůĞƐͬĚǀŝƐŽƌǇ Loan/ Officers
Collectors Default KƉĞƌĂƚŝŽŶƐ Underwriters Fewer
Management ^ĂůĞƐͬĚǀŝƐŽƌǇ Loan Officer Fewer
^ĂůĞƐͬĚǀŝƐŽƌǇ Loan Officer Assistants Assistants
KƉĞƌĂƚŝŽŶƐ Loan Closers and Fewer
^ĂůĞƐͬĚǀŝƐŽƌǇ Loan and Credit Same
^ĂůĞƐͬĚǀŝƐŽƌǇ Loan and Credit Managers Managers Funders
What this chart shows: Loan officer
KƉĞƌĂƚŝŽŶƐ assistants Loan Processors KƉĞƌĂƚŝŽŶƐ
^ĂůĞƐͬĚǀŝƐŽƌǇ CustomerMore
Loan Product Service Fewer
Managers
(second role in the left-hand table) will, following the
KƉĞƌĂƚŝŽŶƐ Underwriters KƉĞƌĂƚŝŽŶƐ
^ĂůĞƐͬĚǀŝƐŽƌǇ Sales PromptCollectorsMore
/ Default Fewer
implementation of generative AI in the mortgages Engineers Management
Loan Closers and Funders
function, either retain their roleKƉĞƌĂƚŝŽŶƐ
within a smaller group KƉĞƌĂƚŝŽŶƐ
KƉĞƌĂƚŝŽŶƐ Ops Engineers More
Loan Fulfillment More
of loan officer assistants (greenKƉĞƌĂƚŝŽŶƐ
arrow), be upskilledCustomer Service KƉĞƌĂƚŝŽŶƐ
Prompt Engineers
Loan Processors Fewer
and deployed to an expanding KƉĞƌĂƚŝŽŶƐ
group of sales prompt Collectors / Default
Retained Roles in Mortgages KƉĞƌĂƚŝŽŶƐ Servicing / Default More
New Roles in Mortgages KƉĞƌĂƚŝŽŶƐ UnderwritersManagement
Fewer
Prompt
engineers (purple arrow), or potentially be retrainedManagement
for a
Different Roles Outside Mortgages Engineers
different role elsewhere in the bank (yellow arrow). KƉĞƌĂƚŝŽŶƐ Loan Closers and Fewer
Eͬ Funders Different Roles More
KƉĞƌĂƚŝŽŶƐ Outside Mortgages
Customer Service Fewer
KƉĞƌĂƚŝŽŶƐ Collectors / Default Fewer

Management
Understand and develop a secure Reinvent talent and KƉĞƌĂƚŝŽŶƐ

Close the gap on Loan
Drive Fulfillment
continuous More Measuring the ROI
Lead with value Prompt Engineers
KƉĞƌĂƚŝŽŶƐ Servicing / Default More
Management Prompt
Engineers
26
The introduction of generative AI in banking User adoption and competence are make-or-
brings a crucial need for a workforce learning
strategy centered on adapting to the changes in
break factors in the successful implementation of
generative AI, as they are with other technologies.
A holistic
the nature of work brought about by automation
and augmentation. Upskilling transcends mere
Many companies face the challenge of low
adoption rates for data-driven tools, often
approach ensures
familiarity with new generative AI interfaces. It due to deployment issues. This highlights the that learning is a
requires a deep understanding of when and how to importance of a human-centered design and the
engage in AI-driven processes and an awareness of creation of supportive learning experiences. Such collaborative effort
the associated risks. This holistic approach ensures an approach not only addresses the technical
that learning is a collaborative effort, involving aspects of generative AI, but also considers the
everyone in the organization and making the human element. Together these ensure that
transition smoother and more effective. employees are not just users on whom change is
inflicted, but become active participants in the
AI-driven transformation.
Lead with value
27
Reinvent talent and ways of working
ACTIONS
• Conduct an in-depth analysis of all banking • Develop a robust strategy for workforce
processes to ascertain where generative AI upskilling, focusing on evolving roles and
deployment is best aligned with strategic continuous learning to keep pace with rapid
business goals. technological advancements.
• Redefine workflows to maximize the efficiency • Engage line and middle managers in shaping
and innovation potential of generative AI, use cases for generative AI and future ways of
breaking down silos for a more collaborative working that align with the strategic vision set
and integrated approach. by senior leaders. Encourage team members to
propose and validate generative AI solutions,
fostering a cycle of iterative improvement and
deeper integration into banking operations.
Lead with value
28
I M P E R AT I V E F O U R
4
Close the gap on responsible AI
Responsible AI (RAI) isn’t merely a compliance The good news is that most banks have for
issue or an ethical imperative; it is also a strategic decades had their responsibility frameworks
asset that enhances brand equity. In our global and extensive model risk management
analysis of job postings by more than 2,200 approaches in place. As they accelerate their
companies—including 176 banks—banking job use of generative AI, they need to enhance
postings for RAI roles grew from 1.1% of the total their existing frameworks and approaches to
in 2019 to 8% in 2022. The average across other incorporate the emerging risks of generative AI
industries lingered at 2.8% in 2022. and LLMs. Some of the potential risks can be
found on the next page.
This indicates that banks are leading the way in RAI,
partly because of the high regulatory burden they
bear. Failure to build a RAI capability could result in
reputational damage and legal consequences.
Lead with value
29
Bias and harm Unreliable outputs disinformation and the enablement of fraud at
There is a greater possibility of unfair or biased Generative AI has already acquired notoriety for scale could impact the socio-economic arena and
decisions relating to marketing offers, credit its hallucinations, the limited traceability of the create negative perceptions of the technology.
decisions, service standards, collection / recovery information it provides, and its reluctance to explain
options, and approaches to customers in its outputs. Insufficient disclosure and transparency Workforce transition
unmonitored environments. The potential for the could result in customers feeling confused and In addition to concerns about job losses within
spread of misinformation, disinformation and toxic alienated and, at worst, being harmed. the workforce, there is a risk that the lack of
responses should be closely monitored. appropriate, universal training could lead to a skills
Confidentiality and security gap similar to the digital divide that, years ago,
Liability and compliance With insufficient controls, generative AI could prevented many from benefitting from the advent
The standards and regulations pertaining to facilitate unauthorized access to and use of the of the internet.
generative AI in different parts of the world are confidential information which is used to train
continuing to be shaped, yet adherence is essential. LLMs. Widespread or inconsistent implementation
Current standards for model risk management of the technology will increase vulnerabilities,
don’t fully address the risks and complexities of which could lead to data leakage and lapses in Inconsistent
generative AI, which could include fines, legal action
and reputational damage due to the improper use
consumer protection.
implementation of
of intellectual property owned by others. Sustainability the technology will
Intensive reliance on generative AI could increase
banks’ carbon footprint, putting their ESG targets
increase vulnerabilities
at risk. As other players adopt the technology, bias,
Lead with value
30
CASE STUDY
An Asia-Pacific bank at the forefront of responsible AI

A leading bank in Asia-Pacific has positioned itself models, with tight controls to minimize bias and
as an AI-first organization, benchmarking its AI maximize the likelihood that the AI’s outputs will
customer experience capabilities against industry be fair and unbiased.
leaders beyond banking. The bank has woven
the principles of responsible AI into its corporate The bank has taken a steadfast approach to
DNA, with the goal of remaining in lockstep with transparency and accountability. It has deployed
societal values and ethical imperatives. over 1,500 AI models, each ingrained with
responsible AI practices and structured to
The bank is shaping its responsible AI agenda in explain the reasoning behind AI-driven decisions.
collaboration with industry regulators, seeking Techniques like prompt engineering and model-
their feedback as it tests its AI ethics and to-model Q&A, coupled with human oversight,
principles. It uses proprietary data to optimize AI ensure that any AI decisions are personalized,
understandable and justifiable.
Lead with value
31
Close the gap on responsible AI
ACTIONS
• Enhance existing RAI governing principles with • Amplify and broaden existing model risk • Put in place change management systems that
clear taxonomy, accountability and governance management measures to mitigate the risks engage with the bank’s compliance program to
for responsible generative AI usage. specific to generative AI. positively address how generative AI will impact
the workforce, sustainability, and privacy and
• Assess the risks of generative AI use cases, • Continuously monitor generative AI systems security across the enterprise.
applications and systems across the and initiatives in conjunction with mitigation
enterprise through qualitative and quantitative and compliance activities.
assessments.
Lead with value
32
I M P E R AT I V E F I V E
5
Drive continuous reinvention
In the process of adopting generative AI There is much about generative AI that is novel
throughout their organizations, in all their functions and exciting. One aspect that stands out is how
and processes, banks will end up reinventing much it has succeeded in democratizing artificial
of how they do business today. To accomplish this intelligence and making technology more human-
will require not only execution excellence but also like. Banks that recognize its potential to redefine
a culture of innovation, a core value of which will entire value chains, and that succeed in blending
be curiosity. the technology with their people’s ingenuity,
will capture long-term value and build lasting
The challenge will be to balance reinvention with resilience.
the ongoing operation of the bank, maximizing the
opportunities while limiting the disruption. This is
why agility, a culture that embraces curiosity and
innovation, and expertise in change management
are so important.
Lead with value
33
CASE STUDY
BBVA embarks on a bold new future12

BBVA launched an agile transformation program to minutes using AI-based facial recognition and text
break down organizational barriers and enable cross- analytics to verify applicants via its mobile app.
functional collaboration. Coaches worked with the BBVA is also able to offer ‘one-click loans,’ which
bank’s people to embed agile principles throughout potentially provide same-day funding for both
the business. Today, it has a flexible talent pool from customers and non-customers, and can onboard
which people can be assigned to the highest-priority customers in a few minutes. This has contributed
projects or processes at any time. to a 150% growth in new customers.
This, among other changes, has enabled BBVA Today, nearly 50 million BBVA customers interact
to evolve from simply offering digital services with the bank through digital channels, seven out
to developing and selling new digital banking of 10 sales are made digitally, and cost-to-income
products and offering leading financial health has dropped to 43%.
tools. In addition, BBVA onboards clients in mere
Lead with value
34
Drive continuous reinvention
ACTIONS
• Establish a holistic strategy for the entire • Employ proven techniques such as Six Sigma,
enterprise. The focus should be on process, kaizen and kanban to effectively understand
data and people, while developing change how and where generative AI can be leveraged
management and agility as core competencies. at scale.
• Balance boldness with nuance. Evaluate

potential use cases carefully, but do take bold
and strategically scaled bets.
• Create a culture of curiosity and innovation by

removing barriers such as functional silos, rigid
roles and discouragement of risk-taking.
Lead with value
35
SIDEBAR
Measuring the ROI of generative AI

Scaling generative AI is not just about immediate Another significant area in which banks will Simple applications might be easy to assess, but
financial gains but preparing for the future. need to invest is the transformation of their as banks delve into more complicated projects,
Approaches for evaluating return on investment organizational structure, processes and culture. like digital banking assistants or digital twins, the
should extend beyond immediate economic This investment is often underestimated. cost of data management, talent and technology
returns and lay the groundwork for continuous In the most recent wave of our recurring Pulse will rise.
reinvention. Leading banks will use a 360° value of Change survey, one in four banking executives
framework to define and track a range of near-, said change management is a challenge.
medium- and far-term objectives, including talent In the previous wave, only 8% held that view.13
development, enhanced customer and employee This makes it the fastest growing concern as
experiences, sustainability and responsible AI. the implementation of generative AI gathers
momentum.
On the investment side of the equation, banks will
need to account for a range of costs. For many— Finally, regarding the generative AI tools
especially those which have not progressed themselves, banks will need to provide for the
very far along their modernization journey— costs of creating, industrializing, maintaining and
strengthening their core will consume a significant using generative AI. These costs will be directly
portion of their overall investment. linked to the complexity of the applications.
Lead with value
36
Banking on AI
Few technologies have evolved at the pace of Leading banks that embrace the generative
generative AI—or have the potential to drive such AI opportunity as a strategic imperative and
a profound reinvention. Leading banks recognize move aggressively from proof of concept to
that the question is no longer whether generative industrialization will likely secure a durable
AI will transform their industry, but how. The competitive edge.
next question is: How soon can they capture the
massive value generative AI promises?
Generative AI will affect nearly every bank

Leading banks recognize
employee in a short time horizon—in fact, the that the question is no
technology is already available in many of the longer whether generative
applications they use daily. However, to harness its
AI will transform their
full potential, banks will need to purposefully and
methodically deploy it and encourage its adoption industry, but how.
in those parts of the organization that will benefit
most from ‘waste out, value in’.
37
References
1. Accenture Research analysis and estimates 7. Salesforce press release, “Salesforce

on BLS and O*Net data as of December 2022. Announces Einstein GPT, the World’s First
Generative AI for CRM”, 7 March 2023.
2. Accenture, “Work, Workforce, Workers:
Reinvented in the age of generative AI”, 8. Workday press release, “Workday Announces
16 January 2024. Fiscal 2024 Third Quarter Financial Results”,
28 November 2023.
3. Accenture Research analysis and estimates
on BLS and O*Net data as of December 2022. 9. FST Media, “Westpac Buddies Up Engineers
with Gen AI Partner”, 11 August 2023.
4. Avanade, “Banks Focus on Efficiency with AI”,
2024. 10. Accenture, Pulse of Change Survey, 2024.
5. Microsoft Blog, “Introducing Microsoft 365 11. Accenture, Reinvention by the Numbers,
Copilot – your copilot for work”, 16 March 18 January 2024.
2023.
12. Accenture case study, “Banking on BBVA’s
6. Adobe press release, “Adobe Unveils Firefly, a bold new future”, 2023.
Family of New Creative Generative AI”,
21 March 2023. 13. Accenture, Pulse of Change Survey, 2023.
38
About the research

Assessing generative AI’s impact on productivity, We tagged each task and the time employees The estimation of generative AI’s effects on
Accenture conducted extensive research on the spend in performing it into one of four categories: revenue generation was informed by insights
impact of generative AI across industries. We used High potential for automation, high potential gleaned from expert consultations and relevant
data from the US Bureau of Labor Statistics and for augmentation, low potential, or no language case studies. Our assumptions included expected
the Occupational Information Network to analyze tasks. We analyzed data for 2.7 million banking improvements in various financial metrics related
descriptions of 900 occupations across industries employees in the US as well as the 170 roles to lending, deposit rates and other banking
in the US, the 19,000 tasks they perform, and the and 3,500 tasks they perform. This enabled us revenues. Specifically, we considered the impact
time employees spend on each of those. This to assess the potential of generative AI across of generative AI on revenue streams such as
allowed us to understand the likely impact of various job roles. wealth and asset management, as well as service
generative AI on labor productivity. charges.
Assessing generative AI’s impact on
We also used a combination of expert judgement banking financials Our assessment of generative AI’s impact
and generative AI itself (GPT-4). We started by To explore the influence of generative AI in the on banking staff expenses was based on an
identifying which of the granular tasks require banking sector, we conducted a comprehensive analysis of over 200 language-related tasks,
intensive use of language. Then, we assessed analysis using financial data from 150 of the using data from the Bureau of Labor statistics.
in which way the task requires the use of largest banks worldwide. This data, sourced from We linked these tasks to various banking roles
interpersonal or cognitive skills to be resolved. S&P Capital IQ, spanned a 12-month period ending and evaluated the potential for task automation
This enabled us to understand whether generative September 2023. and augmentation in different occupational
AI could automate a task or augment the categories, leading to insights on possible
employee’s skills. productivity gains across diverse roles in the
banking sector.
39
About the research

continued
This was contextualized within a scenario where
generative AI’s adoption rate over the next three
years may exceed that of other groundbreaking
technologies from the previous century.
Additionally, our analysis extended to non-staff

costs, where we projected a decrease in certain
expenses, such as marketing, due to the efficiency
improvements attributed to generative AI. We also
considered growth in IT-related expenses, guided
by specific insights from industry analysts.
To assess the impact on risk management,
we focused on the potential of generative AI
to enhance the recovery of consumer loans.
Our methodology incorporated a conservative
estimate of improvement in loan recoveries
attributable to the implementation of generative
AI tools to reduce credit risk.
40
Authors
Keri Smith Michael Abbott Mauro Centonze

Global Banking Data & Al Lead, Global Banking Lead, Thought Leadership Sr. Principal,
Accenture Accenture Accenture Research
The authors would like to acknowledge the following people for their contributions to this report: Adrian Tapp, Alejandro Borgo, Ashish Garg, Avinash Rao, Dariusz
Orynek, David Woods, Francesca Caminiti, Jessica Wolfe, John Newlin, Leslie Wong, Marco Siciliano, Patrick Goth, Rebecca Parish, Rocio Perez-Auba del Campo,
Sarah Schofield, William Knapik.
Read more at accenture.com/GenAIBanking

41
About Accenture About Accenture Research
Accenture is a leading global professional services company that helps the world’s leading Accenture Research creates thought leadership about
businesses, governments and other organizations build their digital core, optimize their the most pressing business issues organizations face.
operations, accelerate revenue growth and enhance citizen services—creating tangible value Combining innovative research techniques, such as
at speed and scale. We are a talent- and innovation-led company with approximately 733,000 data science-led analysis, with a deep understanding of
people serving clients in more than 120 countries. Technology is at the core of change today, industry and technology, our team of 300 researchers
and we are one of the world’s leaders in helping drive that change, with strong ecosystem in 20 countries publishes hundreds of reports, articles
relationships. We combine our strength in technology and leadership in cloud, data and AI and points of view every year. Our thought-provoking
with unmatched industry experience, functional expertise and global delivery capability. research developed with world leading organizations
We are uniquely able to deliver tangible outcomes because of our broad range of services, helps our clients embrace change, create value and
solutions and assets across Strategy & Consulting, Technology, Operations, Industry X and deliver on the power of technology and human ingenuity.
Song. These capabilities, together with our culture of shared success and commitment
to creating 360° value, enable us to help our clients reinvent and build trusted, lasting For more information, visit www.accenture.com/research
relationships. We measure our success by the 360° value we create for our clients, each
other, our shareholders, partners and communities.
Visit us at www.accenture.com
Copyright © 2024 Accenture. All rights reserved. Accenture and its logo are registered trademarks of Accenture.
This content is provided for general information purposes and is not intended to be used in place of consultation with our professional advisors.
This document refers to marks owned by third parties. All such third-party marks are the property of their respective owners. No sponsorship,
endorsement or approval of this content by the owners of such marks is intended, expressed or implied.

Accenture Age AI Banking New Reality

Uploaded by

Copyright:

Available Formats

Accenture Age AI Banking New Reality

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Accenture Age AI Banking New Reality

Uploaded by

Copyright:

Available Formats

The age of AI:

Banking’s new reality

Our latest financial projections indicate that the gains over

22% to 30% 600 bps 300 bps

From the back and middle offices through to

Automation Augmentation All-round support

50% Securities, Commodities,

and Financial Services Sales Agents

Embed Transform Innovate and differentiate

Our conversations with industry front-runners

We see five key imperatives which C-suite

executives should address to reinvent in the

05 Drive continuous reinvention

Unsecured Secured Wealth

Front-Back Book Offer Intent Treasury

Application Processing & Fulfilment Trading

Client Engagement Customer Experience

Risk, Compliance & Financial Crime &

Estimated impact of generative Al Low impact Medium impact High impact

European bank expands its CoE to scale AI benefits

Leading with value

• Foster cross-departmental collaboration to

Forward-thinking banks have an advantage The AI landscape is in a state of constant flux,

Scaling an enterprise-wide LLM and generative AI capability

Understand and develop a secure AI-enabled digital core

• Evolve security protocols to address the

KƉĞƌĂƚŝŽŶƐ Loan Processors ^ĂůĞƐͬĚǀŝƐŽƌǇ Loan Product More

KƉĞƌĂƚŝŽŶƐ Collectors / Default Fewer

Understand and develop a secure Reinvent talent and KƉĞƌĂƚŝŽŶƐ

Reinvent talent and ways of working

An Asia-Pacific bank at the forefront of responsible AI

Close the gap on responsible AI

BBVA embarks on a bold new future12

Drive continuous reinvention

• Balance boldness with nuance. Evaluate

• Create a culture of curiosity and innovation by

Measuring the ROI of generative AI

Generative AI will affect nearly every bank

1. Accenture Research analysis and estimates 7. Salesforce press release, “Salesforce

About the research

About the research

Additionally, our analysis extended to non-staff

Keri Smith Michael Abbott Mauro Centonze

Read more at accenture.com/GenAIBanking

About Accenture About Accenture Research

You might also like