Put AI to work

Finance

IBM
Contents

There’s a unique opportunity 01 → 04 →

for AI in the enterprise finance Financial services Take the next step
in the age of AI
function. Despite finance’s distinct 05 →
restrictions and multiple overlapping 02 → Ready to begin?
responsibilities, bold firms have Why AI in finance matters
already realized ROI by applying AI
solutions. We’ve learned from them 03 →
Put AI to work for
to outline the why, what and how of financial services
getting real value from AI in finance.

2
01 Financial services
in the age of AI
With unique responsibility in every part of
the business, as well as specific demands for
dependability and accountability, finance
pros are feeling squeezed from all directions.
AI has the potential to change this status quo.
01

Financial services
in the age of AI
Today’s CFOs and finance teams are
facing challenges in multiple areas.
To name a few, they need to drive
digitization and make purchasing
decisions for the entire enterprise.
They’re tasked with maximizing returns
on digital investments. And they need
to adapt the enterprise to changing
customer spending habits.

CFOs surveyed rank the balance sheet

as their top concern, but also face a
triple squeeze of achieving profitable
growth, managing inflation and
navigating a talent shortage.¹

4
01 Financial services in the age of AI

Add these factors all together, and the CFOs are expected to be both innovative
task of a CFO can seem not only difficult, and stable; visionary and detail oriented.
but impossible. And on top of it all, CFOs In short, where once CFOs reported on
face specific challenges and expectations, past activity and provided high-level
even within the C-suite. Think about a strategic input, they’re now expected to
few compliments you might hear about be very granular, weighing in at the tactical
different C-suite executives: level. If you’re a CFO today, everything is
your problem.²
– Our CEO is a real maverick.
– Our CTO encourages us to push the Generative AI (gen AI) offers finance teams
boundaries of what’s possible. a new tool for addressing this complex and
– Our CMO changes the game by embracing challenging situation. It’s clear that gen
disruptive technologies. AI is going to have an impact in finance, as
in virtually every enterprise function. This
Are these examples still compliments guidebook explores how CFOs can get the
that a CFO would like to hear? most from gen AI, including how to prepare
for it, where to apply it and what they
need to make it a valuable addition to their
finance function.

5
02 Why AI in
finance matters
Finance pros aren’t used to going out on a
limb. But there are many areas of finance
where dependable automation and intelligent
acceleration can really deliver value—and
already has.
02

Why AI in
finance matters
CFOs have a particular remit and good
reason to be cautious, particularly when
embracing new technologies. Perhaps
that’s why most organizations have only
adopted AI within 1 or 2 of the 4 key
finance workflows as discussed on
page 19.³

Nevertheless, there remain many areas in

finance where quicker and more accurate
ingestion of information can mean major
productivity gains and avoidance of
errors. Gen AI has the potential to be
applied to these hard-to-automate areas,
empowering faster and better data-driven
decisions based on historical data, market
trends and AI foundation models that
identify patterns and anomalies often
missed by traditional analysis methods.

7
02 Why AI in finance matters

At its core, AI technology can support Financial analysis and reporting teams The benefits are expected to compound
finance professionals so they can focus on have long made use of AI tools and will be as gen AI is applied to more ambitious
higher-value activities. And surveys show one of the areas most impacted by gen AI, use cases in areas, such as regulatory
that AI adoption means better performance bringing self-service capabilities into many compliance, forecasting and designing new
by the finance organization. new realms of the business. There are text- accrual methods. Recently, 600 surveyed
based productivity use cases in several global finance managers agreed that gen AI
For example, a finance organization that sectors where gen AI can be adopted now, can expand the reach of AI in finance.
effectively supports the enterprise as and the benefits for mature AI adopters
a whole is 6 times more likely to have in finance are already significant. Benefits The potential is clearly there. But to be
mature AI adoption than simply piloting include 33% faster budget cycle time successful, finance leaders will need to get
AI.³ And for all 4 key finance processes, for planning and monthly close, a 43% actively involved in the analysis process;
a significantly higher percentage of reduction in uncollectible balances, and weigh risks, materiality and potential
organizations that have implemented 25% lower cost per invoice paid.⁴ financial exposures; and partner with other
AI end to end with an IT strategy that business units from the beginning.
integrates AI, cloud and app modernization These benefits are so compelling that
have achieved top quartile ROI—more than the opportunity cost of not adopting AI
30%—compared to other organizations.³ is significant. Organizations that adopt
a wait-and-see approach risk rapidly
falling behind.

8
02

Scaling AI across
the enterprise,
safely

Despite the widespread enthusiasm for

using gen AI for business, especially
from board members and investors,
many business leaders still have
concerns about its adoption.

9
 61%
In a recent study, the IBM have concerns about data lineage
Institute for Business Value or provenance.
found that responding
executives have 4 top concerns
about gen AI adoption.3

57%
have concerns about data security.

53%
have concerns about the constraints
introduced by regulations
and compliance.

45%
have concerns about data privacy.

10
02 Why AI in finance matters

How your organization succeeds with To continue this practice in the age of AI,
gen AI is influenced by how you select, IBM has developed a multidisciplinary,
govern, analyze and apply data across multidimensional approach that embeds
it. Huge volumes of data from different ethical principles into AI applications and
sources are used to train gen AI models, processes. With IBM’s Principles for Trust
so, implementing governance, management and Transparency and Pillars of Trust as the
and ethical frameworks that operate end foundation for our AI ethics initiatives, we're
to end is key if you wish to adopt AI safely helping people and organizations adopt AI
and responsibly. responsibly, and with clear purpose.

IBM has long followed core principles

grounded in commitments to trust,
transparency and fairness to guide how we
handle client data and insights—and how
we develop and deploy new technologies.

11
02

IBM’s guiding Principles for Trust Pillars

principles for and Transparency of Trust
AI ethics
The purpose of AI is to augment Explainability
human intelligence. Good design does not sacrifice transparency in creating
We believe AI should make all of us better at our jobs, and a seamless experience.
the benefits of the AI era should touch the many, not just
the elite few.
Fairness
Properly calibrated, AI can assist humans in making fairer choices.
Data and insights belong to their creator.
Clients’ data is their data, and their insights are their insights.
We believe government data policies should be fair and Robustness
equitable, prioritizing openness. As systems are employed to make crucial decisions, AI must be
secure and robust.

Technology must be transparent and explainable.

Companies must be clear about who trains their Transparency
AI systems, what data is used and what goes into their Transparency reinforces trust, and the best way to promote
algorithms’ recommendations. transparency is through disclosure.

Privacy
AI systems must prioritize and safeguard consumers’ privacy
and data rights.

12
02 Why AI in finance matters

In the past, enterprises have approached So, does this shift in technology mean you
AI as an add-on, with the end goal being need to replace your traditional AI solutions
digital transformation. with the latest gen AI for finance? IBM AI
leaders say no.
Now, AI is becoming the centerpiece of
business transformation—75% of business Traditional AI models that use conventional
leaders surveyed believe competitive machine learning (ML) and rules-based
advantage will depend on who has the models have different capabilities and serve
most advanced gen AI.4 But harnessing the different functions: predictive analysis,
potential of AI to fundamentally transform security and compliance, automation, and
finance requires a mix of vision and more. Gen AI models, on the other hand,
technology. Enterprises need to put AI to use foundation models to autonomously
work at the strategic core of the business— generate content based on the data they
not just add it on to existing systems—to were trained on.
solve challenges and help achieve their
business objectives. It’s time to move from
+AI to AI+.

13
The AI Ladder®
in the modern day AI+

Let AI help with the work

Replace your workflows

Automate your workflows

Add AI to your applications

Collect, organize, grow data

+AI
14
02 Why AI in finance matters

Gen AI and traditional AI

are, in fact, complementary Traditional Generative
AI ↷ AI ↷
technologies; use them together
to help accelerate and achieve
your strategic goals.

Applies predefined rules and Uses both small language models

algorithms to specific sets of data to (SLMs) and large language models
help solve problems, make predictions (LLMs) to generate new content
and automate tasks. based on patterns learned from
the data they were trained on.

Potential IT tasks using traditional AI Potential IT tasks using gen AI

– Predictive maintenance – Issue summarization
– Problem detection – Action recommendations
– Incident response – In-depth conversation
– Service desk automation – Code generation
– Process automation – Insight extraction
– Intelligent workflows – Classification
– Capacity planning – Testing and debugging

15
03 Put AI to work for
financial services
The 4 key finance workflows—O2C, FP&A, R2R
and P2P—can each benefit in different ways from
the addition of gen AI. Based on our experience
with finance and AI transformation, here are the
areas to focus on within each workflow.
03 Put AI to work for financial services

Where finance managers

see gen AI impacting
financial processes³

33% 48% 34% 39% 30% 54% 35% 40%

Generate Analyze and Identify potential Provide self-service Generate financial Conduct financial Analyze external Detect fraud.
management and process bills. risks and provide to answering reports and analysis and data to help c
 reate
performance reports recommendations financial questions. visualizations. create forecasts. business cases.
and visualizations. to manage them.

17
03

Different finance workflows have different

areas of applicability for gen AI. An
experienced partner can help you identify
areas where gen AI can have the most
valuable impact and link together the
technology pieces needed to support
those use cases.

Based on IBM’s experience through the

whole combined ecosystem of finance
and AI transformation services, the
following are the applicable areas, key
processes and benefits in each of the
4 key finance workflows.

18
03 Put AI to work for financial services

1 2
Order to cash (O2C) Financial planning and analysis (FP&A)
AI-driven innovations in O2C help with AI and advanced analytics rank as
credit scoring, pricing decisions and key components of the FP&A process,
the prevention of payment frauds. AI galvanizing and orchestrating planning
collects, prepares and distributes data and performance management. AI can
from documents and unstructured data, partially automate the labor-intensive
surfacing insights at key moments to process by parsing through data on
optimize decision-making and accuracy. the market, company performance,
In the key metric of uncollectable competitor information, pricing and
balances, half of AI adopters credit AI with operations. The AI model can identify
a decrease of at least 2% and one-quarter anomalies, enhance forecasting, optimize
credit AI with a decrease of at least 8%.⁴ pricing and provide recommendations, as
well as apply trend analysis, correlation
analysis—including pattern and anomaly
detection—and neural networks for
financial forecasting. With neural
networks, AI determines the relationship
among data and uses it to predict
new data, resulting in higher forecast
accuracy. When AI is used for market
performance comparison, FP&A teams
can factor in more variables and internal
and external influences.

19
03 Put AI to work for financial services
Key processes Benefits4

O2C – Manage sales orders. – 43% reduction in

3 4
– Process customer credit. uncontrollable balances
– Invoice customers. – 32% decrease in days sales
– Process accounts receivables. outstanding (DSO)
– Manage and process collections. – 28% decrease in the cycle time from
– Manage and process transmission of invoice to receipt
adjustments and deductions. of payment
Record to report (R2R) Procure to pay (P2P)
In the general accounting and reporting An AI-powered workflow, underpinned FP&A – Perform integrated – 25% lower cost for planning,
area, AI-powered workflow and data by data models, optimizes touchless financial planning. budgeting and forecasting
models could include a reconciliation processing and provides a unified interface – Perform planning and budgeting. – 33% faster budget cycle time
module that helps users aggregate for buyers, suppliers, procurement and – Perform management and – 4% higher overall forecast accuracy
subledger transactions and perform finance staff. Invoices are validated against performance reporting. – 57% lower sales forecast error
risk-based reconciliations and business rules, coded and matched to – Perform forecasting
cognitive forecasting. the purchase order automatically. Spend and modeling.
and pricing intelligence provides insights
during sourcing. AI automates procurement R2R – Process journal entries. – 31% lower cost per journal entry
operations and manages inquiries from – Reconcile the general ledger. – 33% faster cycle time for the
buyers and suppliers. Automating P2P with – Post and reconcile monthly close
AI has been shown to increase productivity intercompany transactions. – 2% more journal entries that are error
and permit finance to detect more – Perform consolidations and free first time
fraudulent invoices. process eliminations.
– Close the books.
– Perform fixed assets accounting.
– Perform financial reporting.

P2P – Procure products and services. – 25% lower cost per invoice paid
– Process accounts payables. – 32% faster cycle time from receipt of
invoice until approved and scheduled
for payment
– 3% more purchase orders processed
error free first time

20
03 Put AI to work for financial services

To meet the demands of today’s finance business That’s why IBM has significantly invested
in elevating and supporting our business
leaders, organizations must be continuously
partners. We combine their industry domain
innovating. But innovation doesn’t happen in silos. expertise with our innovative solutions to
better meet the needs of our clients. Our
business partners are using gen AI to deliver
turnkey solutions with the simple goal of
making it easier for finance and accounting
teams to report, plan and forecast. IBM
Gold Partner, Breakaway Technologies,
provides an out-of-the-box solution for
managed markets contract analysis, discount
forecasting, and reserve reporting—this
solution can help clients better predict and
plan, while improving productivity.

21
04 Take the
next step
Once you understand what to do with gen AI,
understanding how to do it is another thing entirely.
We recommend focusing on these 3 tasks: defining
the right business case, selecting high-value tasks
for automation, and applying FinOps.
04 Initiatives finance organizations have undertaken to support their AI strategy³

Take the
next step
As enterprises transition from the hype to the
how of gen AI, it becomes even more critical
to understand the costs of—and the value
derived from—investments in the technology.
CEOs and their CFOs know they need to
tread carefully while adopting it, but they
also feel the need to act fast. Still, according
to an IBM Institute for Business Value study,
60% of organizations surveyed have not yet
developed a consistent, enterprise-wide
approach to gen AI.5

AI and gen AI initiatives can only be as

successful as the underlying data permits,
so enterprises undertake a variety of data
initiatives to support their AI strategy,
ranging from process mining to data
governance. Respondents were asked “which
initiatives has your finance organization
undertaken to support its AI strategy?”

23
04 Take the next step

1
In addition, finance departments deploying Define a business case for your O2C metrics include total cost, cost as a Finance leaders noted significant results
AI should follow a structured approach to AI improvement initiatives. percent of revenue, payment processing against performance metrics, which
identify and assess potential risks: This business case should include time, collections efficiency, customer improve over time.
3 justifications: satisfaction, DSO, percent of invoices
– Comply with privacy and security processed without error, average revenue
regulations, policies and practices to – Prioritize potential benefits and risks. per customer and delinquency rates.
ensure that data collection and usage – Ensure that there’s alignment with the “Before implementing [AI], we tracked
practices meet requirements. broader strategy. FP&A metrics include margins, budget KPIs, such as supplier lead time, purchase
– Establish clear governance structures, – Secure funding. cycle time, accuracy of forecasts for sales order cycle time and supplier quality
including the definition of roles and and volume, and compliance. ratings. After implementation, we saw
responsibilities of all stakeholders involved To quantify these justifications, you’ll significant improvements in these metrics,
in AI implementation. need to use project-specific, high-level R2R metrics include ROI, close cycle time, with a 20% reduction in supplier lead
operational and financial metrics that query response time, accuracy of financial time, a 30% reduction in purchase order
But even as you lay this data, infrastructure measure cost, speed and quality. Tracking statements, percent of manual journal cycle time and a 15% improvement in
and process groundwork, how can you and reporting the appropriate metrics is entries, percent of error-free journal entries supplier quality ratings.”⁵
be sure your approach to gen AI will key to demonstrating the ROI from AI. and the intercompany reconciliation rate.
accommodate organizational and The following are some examples. Head of Finance
technology changes? P2P metrics include cost per invoice, Retail, United States
working capital, cycle time to process an
Here’s where you can benefit from a solid AP invoice, credit cycle time and percent
strategy based on experience drawn from of error-free purchase orders.
thousands of successful AI engagements.
We’ve provided 3 tested recommendations
to get you started.

24
04

IBM Business Partners are proactively “We at QuantumStreet AI are excited

working to help their clients improve
their risk management processes.
about expanding our use of Watson
QuantumStreet, an IBM Business Partner
that specializes in investment and
Discovery and Watson Studio into IBM’s
wealth management, is embedding AI foundation models watsonx.aiTM and
technology to improve the workflows
of its investment managers: watsonx.dataTM. These capabilities
will allow nontechnical investment
managers much needed access to
AI/ML tools, enabling them to identify
market opportunities and manage risk.”
Chida Khatua
CEO of QuantumStreet

25
04 Take the next step

2
Attack labor-intensive tasks that are ripe IBM research shows that the finance areas “Purchase order processing
for gen AI automation. poised to realize the greatest value from time used to be 7 days and
Use gen AI tools that automate manual gen AI include predicting anomalies (47%),
and mundane tasks. Focus on use cases explaining variances (41%), generating now it is done in hours.”⁵
that identify and mitigate risks before scenarios (40%), creating reports (39%)
making investments that might impact and managing accounts—both payables Head of Finance
core finance functions. (38%) and receivables (38%).⁵ Potential Automotive, United States
use cases abound, but could include:
– Acknowledge the short-term limitations of
gen AI while planning to capitalize on its – Capture and leverage institutional
full potential. While gen AI isn’t proficient knowledge such as recommending
at numeric analysis yet, it likely will be accounting treatments.
soon. Develop value-adding use cases now – Streamline audits by analyzing data to
to prepare for this future gen AI capability. identify discrepancies and anomalies.
– Drive targeted gen AI adoption in – Aid in contracts and negotiations by
day-to-day activities. Get individuals analyzing legal language to identify
to use gen AI in specific daily tasks. potential issues, risky clauses or
Aligning gen AI adoption to real-world opportunities for improvement.
responsibilities will help overcome
employee resistance and prove the value
of this transformative technology.

26
04

“The farther along organizations are

in their AI journey, the more value is
delivered. … Organizations that have
advanced to the optimization phase

51%
see [the highest] ROI:”6

Head of Finance
Automotive, United States

27
04 Take the next step

3
Fine-tune your FinOps practices with gen Gen AI offers you an opportunity to improve Most organizations that have
AI and apply FinOps across the enterprise value along 2 axes at once: fine-tuning activated a FinOps model
to make technology even more valuable. FinOps with gen AI and applying FinOps
FinOps, or financial management for across the enterprise. Gen AI can make using AI-powered software
cloud-based investments, should play a FinOps more accurate and efficient, report cost savings greater
big part in gen AI investment decisions. improving cost-benefit analyses and
It increases visibility into cloud-related helping teams prioritize use cases. than 20%.⁷
financial data, finding operational efficiencies
and measuring returns. It formalizes and – Apply FinOps practices to your gen AI
structures what might otherwise be a largely investments by implementing a cost
random process—and reduces the potential estimation and tracking framework
for unexplained cloud expenses. that provides clarity on the costs of
gen AI projects.
– Use gen AI to enhance FinOps
capabilities. Simulate financial data and
scenarios that can help increase the
accuracy of financial models, improve
risk management and support strategic
decision-making.

28
05 Ready
to begin?
Now is the time to solidify competitive advantage by
working with a partner that can help not just pilot and
adopt but scale AI across the finance function.
05

The right partner can help,

from financial planning through
record to report, to optimize
productivity and innovation
Join an IBM AI strategy AI expertise AI platform
while prioritizing compliance briefing for finance
and governance. IBM Consulting® can help you strategize, IBM watsonx is our data and AI platform,
IBM has the technology and the modernize, build and manage applications native to hybrid cloud, that’s designed to
consulting expertise to help you with the power of AI and hybrid cloud. IBM provide you with the ability to train, tune
move forward quickly with AI. has formed a center of excellence (CoE) and deploy gen AI across your organization,
for gen AI, specialized in applying the IBM using your critical, trusted data wherever
In just a single 1–4-hour watsonxTM platform and IBM ecosystem it resides.
session you’ll learn: partner technologies to rapidly create
coherent strategies and accelerate how IBM watsonx OrchestrateTM
– How to adopt and scale AI safely yet we solve business problems. The CoE
efficiently across your enterprise complements more than 21,000 IBM data IBM GarageTM
– How organizations such as yours are and AI consultants with a strong track
gaining business value from gen AI record in AI who are already collaborating To discover how IBM helps finance leaders
in priority use cases with thousands of global clients and deploy and scale AI confidently, visit
– How IBM technology and consulting partners to shape the future of AI. ibm.com/consulting/finance.
expertise is helping clients in industries
such as yours transform with AI IBM Partner Plus is an extensive partner
– How to progress to a pilot network, backed by our AI and hybrid
cloud solutions, armed to meet client
demands for quick tech adaptation and
deep understanding. We drive innovation
and growth, solving complex challenges
with speed and value.

30
1. Understand Your Customer: Finance Leader © Copyright IBM Corporation 2024 Examples presented as illustrative only. Actual results
Priorities, Gartner, March 2024. will vary based on client configurations and conditions
IBM, the IBM logo, IBM Consulting, IBM Garage, and, therefore, generally expected results cannot be
2. According to Monica Proothi, Partner and Leader of IBM Partner Plus, IBM Watson, IBM watsonx, The provided. THE INFORMATION IN THIS DOCUMENT
Global Finance Transformation at IBM. AI Ladder, watsonx.ai, watsonx.data, and watsonx IS PROVIDED “AS IS” WITHOUT ANY WARRANTY,
Orchestrate are trademarks or registered trademarks EXPRESS OR IMPLIED, INCLUDING WITHOUT ANY
3. Performance data and benchmarking, of International Business Machines Corporation, in the WARRANTIES OF MERCHANTABILITY, FITNESS FOR
IBM Institute for Business Value, 2024. United States and/or other countries. Other product A PARTICULAR PURPOSE AND ANY WARRANTY OR
and service names might be trademarks of IBM or CONDITION OF NON-INFRINGEMENT. IBM products
4. AI impact in finance benchmark report, other companies. A current list of IBM trademarks is are warranted according to the terms and conditions
IBM Institute for Business Value, January 2024. available on ibm.com/legal/copytrade. of the agreements under which they are provided.

5. Insights into AI’s impact on finance, This document is current as of the initial date of No IT system or product should be considered
IBM Institute for Business Value, February 2024. publication and may be changed by IBM at any time. completely secure, and no single product, service
or security measure can be completely effective in
6. The CEO’s guide to generative AI / Platforms, data, Not all offerings are available in every country in which preventing improper use or access. IBM does not
and governance, IBM Institute for Business Value, IBM operates. warrant that any systems, products or services are
12 September 2023. immune from, or will make your enterprise immune
from, the malicious or illegal conduct of any party.
7. The CEO’s guide to Generative AI / Finance,
IBM Institute for Business Value, 2 April 2024. The client is responsible for ensuring compliance with
all applicable laws and regulations. IBM does not
provide legal advice nor represent or warrant that its
services or products will ensure that the client is in
compliance with any law or regulation.

31