How AI will transform

the CFO’s role

As smarter automation takes over rote tasks and AI
is deployed alongside analytics, finance leaders will
become instrumental in guiding business strategy.

www.pwc.com/cfo-artificial-intelligence
2 | PwC How AI will transform the CFO’s role

Contents

Introduction 3

1
Set clear aspirations for target outcomes and benefits 5

2
 Directly involve frontline employees in designing and
implementing solutions 7

3
Define your area of focus across people, processes and
performance 8

4
Put the right governance structure in place, both for data
and for AI overall 9

5
Prepare the workforce 10

Conclusion 11

Contacts 12
3 | PwC How AI will transform the CFO’s role

Introduction

In every industry, chief financial officers (CFOs) must contend with a

variety of challenges, including sluggish growth, new competitors, an
evolving workforce, changing business models, exploding amounts of
data, greater regulatory complexity and a persistent need to cut costs.
In this environment, it’s no longer enough for CFOs to be the financial
stewards in their organisations, ensuring finance reports are accurate and
complete. In addition, they now need to harness the data flowing through
the organisation to take on a more strategic role, helping fellow business
leaders see not only where the organisation stands at any given moment,
but also where it can and should be next week, next month and next year.

Technologies such as robotic process automation (RPA), intelligent process automation (IPA)
and artificial intelligence (AI) — combined with various analytics approaches and tools — can
help CFOs move forwards on this path and ultimately transform the entire finance function.
According to PwC’s Finance Effectiveness Benchmarking Report 2019, 61% of finance leaders
believe that finance functions could become more effective with improved technology.1

In fact, CFOs are uniquely positioned to lead a broader organisational shift into digitisation.
They have insight into all business units and how they interact with one another, helping
leaders from all areas of the organisation understand the why of finance data — not just the
what — and ultimately leading to smarter business decisions throughout the enterprise.

There’s a logical sequence for how CFOs can lead this change. It starts with basic automation
tools to handle routine, predictable processes like reporting and reconciliation (leading to
increased efficiency and lower costs), and to improve the performance of a given function
(improving accuracy and speed). After standardising and automating repetitive tasks, the
finance department can start applying more advanced tools such as AI and analytics to
interpret data, more accurately predict what will happen according to a variety of factors, and
plan for various scenarios. In other words, CFOs would follow a path starting with descriptive
analytics and proceeding through more accurate diagnostic, predictive and prescriptive
analytics (see exhibit, next page).

1 PwC’s Finance Effectiveness Benchmark Report 2019.

 4 | PwC How AI will transform the CFO’s role

Exhibit: The path to digitising the finance function

Backwards-looking Forwards-looking

Prescriptive analytics

Predictive analytics Recommend ‘right’

or optimal actions or
Diagnostic analytics decisions (What should
Predict future outcomes be done?)
Increasing business value

based on the past

Descriptive analytics (What could happen?)
Identify causes of
trends and outcomes • Real-time product and
(Why did it happen?) service propositions
Describe, summarise (graph analysis, entity
and analyse historical • Forwards-looking view resolution on data
data (What happened?) of current and future lakes to infer present
customer value customer need)
• Observed customer
behaviour • Sentiment scoring • Rapid
 evaluation
• Nontraditional
 data • Graph
 analysis and of multiple ‘what-if’
• Traditional data sources such as social natural language scenarios
sources listening and web processing to identify 
• Optimisation decisions
• Standard
 statistical crawling hidden relationships and actions
methods and themes
• Dual objective models
• Behavioural economics

Increasing sophistication of data and analytics

By progressing in this linear manner, the finance function can generate the momentum to apply
those solutions to other parts of the business. At public utility Duke Energy, based in Charlotte,
North Carolina, for example, CFO Steve Young says, “We started in finance, with a few
software robotics that now perform bank reconciliations, account reconciliations and financial
statement compilations. A lot of those tasks were done quickly. So, I could say, ‘Hey, we’re
doing it here in finance. You need to start doing it in your department.’”2

Every CFO will have a different approach, based on the industry that the person works in and
the level of digital maturity in the organisation. But you’re more likely to succeed if you focus on
five priorities:
• Set clear aspirations for target outcomes and benefits
• Directly involve frontline employees in designing and implementing solutions
• Define your area of focus across people, processes and performance
• Put the right governance structure in place, both for data and for AI overall
• Prepare your workforce

2 Minaya, E., and Shumsky, T., 13 May 2019. “How Automation Is Changing the Workplace at Duke Energy,” Wall Street Journal.
5 | PwC How AI will transform the CFO’s role

1
Set clear aspirations for target
outcomes and benefits

For many CFOs, plans for where to implement automation and AI will be closely linked to
broader digital transformation goals for the entire enterprise. As PwC’s latest Digital IQ survey
showed,3 those goals can be broken into four categories, which are increasingly complex but
also hold the potential for unlocking more value:
• Efficiency — executing existing processes faster and at lower costs
• Effectiveness — executing those processes better (for example, at higher levels of accuracy)
• Expansion — taking on new tasks or entering new markets
• Disruption — finding new ways to create value, such as business model innovation

RPA is well suited to the first two of these objectives — increasing efficiency and effectiveness
by automating time-consuming finance functions. For example, the subsidiary of a global
reinsurance firm implemented two bots to increase the speed and accuracy of certain routine
finance processes. One bot converts broker statements into a data format that allows them to
be processed automatically. The other supports quality checks in the consolidation process.

To tackle more complex objectives like expansion or disruption, organisations need to use
more advanced tools, such as AI and machine learning. These tools will help them make
better decisions and uncover new market opportunities. For example, Microsoft sought to
reduce its risk exposure and ensure compliance with the Foreign Corrupt Practices Act (FCPA)
in its dealings with global resellers.4 The resulting analytics-based solution conducts real-
time compliance reviews during the life cycle of a sale. Using RPA, IPA, machine learning,
AI and data visualisation, the company can flag potential corruption risks by identifying
trends, patterns, relationships and anomalies in individual sales. Human staffers can review
transactions flagged by the analytics solution and take action — including cancelling potentially
risky deals before they’re finalised.

3 PwC Digital IQ 2018: “The no-excuses way to win in a digital world.”

4 PwC Case Study: “Prioritising ethics and integrity: How Microsoft uses data analytics to fight corruption.”
6 | PwC How AI will transform the CFO’s role

Similarly, financial institutions are increasingly turning to AI to combat money laundering and
other financial crimes. In one case, regulators directed a global bank to review about 20m
business customer transactions going back several years. The bank partnered with an AI
software developer to analyse the transactions, identify patterns and flag outlier behaviours,
which human staffers could then investigate more closely.

Read more
Prioritising ethics and integrity: How Microsoft uses data analytics to fight
corruption

PwC’s Finance Effectiveness Benchmark Report 2019

7 | PwC How AI will transform the CFO’s role

2
Directly involve frontline
employees in designing and
implementing solutions

From a cultural perspective, CFOs need to understand that automation and analytics can seem
like a disruptive threat to many employees. During the implementation phase, organisations
need to position the new tools as an opportunity to do more interesting, value-creating work,
rather than something that will replace human staff. In many cases, organisations can overcome
resistance and generate buy-in by giving people the training, tools, autonomy, incentives and
appropriate governance structures to get involved in designing and implementing solutions.

Citigroup offers an example of employee-led innovation. As part of a broader strategy initiative,

the global financial-services company gave a subsection of its finance workforce a hands-on
role in developing and deploying a digital tool kit that used RPA, data visualisation and natural
language processing (NLP, a way for AI systems to understand human speech patterns and
enable human–machine communications). This has allowed employees to automate everything
from simple finance reporting to more complex activities, such as generating management
commentary about financial results.

Meanwhile, insurance companies are starting to tap the combined expertise of actuaries
and data scientists to build predictive analytics models that help them improve in various
areas, including underwriting and pricing life insurance plans, risk and capital management,
policyholder engagement, and reserves (the money set aside to pay policyholders who have
filed or are expected to file legitimate claims on their policies).5 These initiatives capitalise on
the insights of actuaries, who have always been an elite group at insurance companies, and
empower them to do more with data and analytics.

Read more
OrgDNA: The key to unlocking your company’s potential

Three governance considerations to unlock the power of AI

5 PwC, March 2018. “How do actuarial and data science skills converge at life insurers?”
8 | PwC How AI will transform the CFO’s role

3
Define your area of focus
across people, processes and
performance

Automation for its own sake isn’t the objective. CFOs should use automation as part of a broader
transformation to a modern finance function, balancing their focus across people (upskilling and
creating new career paths that improve the employer value proposition of the finance function),
processes (making that function more effective) and performance (generating bottom-line
financial results for the company). There is clearly a lot of room for improvement: as recently
as two years ago, 60% of the global organisations surveyed for PwC’s Finance Effectiveness
Benchmark Study were still working with Excel spreadsheets in an effort to gain year-end
financial insights.

The corporate tax team at medical device maker Boston Scientific took this challenge on
directly. The team had been burdened by manual tasks that were so time-consuming employees
often had to work overtime during busy periods. In addition to paying overtime expenses, the
company was finding it difficult to retain top tax talent. Seeing an opportunity to free their team
from repetitive work by automating a variety of tasks, tax leaders tested the possibilities with
a few processes at first, then built on those successes by automating others. The automation
software they implemented reduced the time required to complete manual tasks by 85% and
created a better work–life balance for employees in the corporate tax and finance functions.
Moreover, establishing an RPA centre of excellence helped ensure that tax leaders could scale up
automation and the company could replicate their team’s success elsewhere.6

Read more
PwC case study: Boston Scientific boosts efficiency and job satisfaction with
robotic process automation

How do actuarial and data science skills converge at life insurers?

6 PwC case study: “Boston Scientific boosts efficiency and job satisfaction with robotic process automation.”
9 | PwC How AI will transform the CFO’s role

4
Put the right governance
structure in place, both for data
and for AI overall

Data is fuel for automation initiatives — without data, nothing runs. PwC’s Finance
Effectiveness Benchmarking Report 2019 found that top-performing finance functions spend
75% of their time on data analysis.7 Your goal should be to create a governance structure that
lets finance pull data from the right sources, ensure that it is accurate and clean, and give
access to the right people at the right time — when that access can lead to better decisions.
Don’t overlook external, unstructured data that you can integrate with your internal data to help
you provide strategic guidance to the business in order to mitigate risk and ensure compliance,
for example, or economic data to run what-if scenarios and feed predictive analytics.

Helping business leaders use the data often requires presenting it in new ways, through
visualisations, dashboards and other tools. This may mean you need to provide dynamic self-
service dashboards that allow business users to create custom comparison views and link to
more detailed data to give them the necessary context for specific data sets. Governance also
requires restricting access to data in order to prevent sensitive information from getting into the
hands of people who don’t need it.

More generally, governance applies both to the broader concept of AI and to the models
themselves. As AI begins to affect all areas of the business, including marketing and HR,
companies need to provide structures for model governance. Several companies have
already generated headlines by deploying AI that replicated hidden biases in their data, was
difficult to control or was prone to being fooled. Part of responsible AI is ensuring end-to-end
governance, from definition of an organisation’s AI strategy, through training, testing, deploying
and monitoring AI and the data used to train it.

Read more
A practical guide to responsible artificial intelligence (AI)

Five standards for responsible AI use

7 PwC’s Finance Effectiveness Benchmark Report 2019.

10 | PwC How AI will transform the CFO’s role

5
Prepare the workforce

The best technology will not deliver results if a CFO doesn’t have a workforce with the right
capabilities. Why not? Because technology alone is never the solution. Employee adoption —
supported by appropriate training and a mechanism for continuous improvement — is what makes
the difference. Structured training and upskilling programmes can help give frontline employees the
skills they need in a finance function where more tasks and processes are automated.

Although few transaction-level employees are likely to be transformed into data scientists, their
expertise should nonetheless be mined as thoroughly as possible. Their real-world insights
into how processes and tasks actually get done can make them a useful interface between
the finance team and the technical team implementing the automation tools. Moreover,
such workers often have valuable suggestions about where to start. One German company
developed a ‘build your own bot’ programme in which nontechnical finance staff identified
specific processes that could be automated to increase efficiency. In other words, these
employees are not coding bots, but they’re highlighting ways to make bots more useful.

This is not to say you should assume a lack of technical skills in your ranks, either. We
surveyed the members of a client’s finance team to get a better understanding of their skill
sets and were pleasantly surprised to learn that 20% of the company’s finance employees had
some coding experience.

Some finance organisations may need a culture shift among their employees as well. In the
past, finance rewarded adhering to rote processes and avoiding risk. Finance has traditionally
been concerned with certainty and accuracy of historical data. The changes you want to
encourage, ideally by adjusting your own perspective, should help your team cultivate a more
exploratory, proactive mind-set.

Read more
Preparing for tomorrow’s workforce, today

PwC’s 22nd Annual Global CEO Survey: Talent Trends

11 | PwC How AI will transform the CFO’s role

Conclusion

When CFOs begin automating rote processes, the finance function runs
more efficiently and effectively, with greater speed and accuracy and
lower costs. And when CFOs start applying such advanced tools as AI,
they shift from descriptive to predictive and prescriptive applications,
helping finance anticipate changes and becoming a true source of
business intelligence in the organisation. These tools do not replace
human staff but rather augment them.

For CFOs, the only question regarding automation and AI is how to begin. Implementing
technology today is a different matter than it was in the past. It happens faster, and it prioritises
real-world pilots and experimentation rather than planning and deliberation. For CFOs, after
thinking through the points above, the final step is to embark on some initiatives: identify use
cases involving automation through straightforward tools like RPA, move into more ambitious
AI-powered initiatives and build on that experience over time. Only by doing so will you begin
to create and equip a truly modern finance function, with the capabilities to compete today and
into the future.
12 | PwC How AI will transform the CFO’s role

Contacts

Bob Woods
Partner, PwC US
+1-206-398-3341
[email protected]

Bob Woods is a finance transformation leader for PwC US in the

technology, media and telecommunications sector. He specialises in
consulting and risk services to public and private client CFOs, COOs
and CIOs in the high-tech, software, cloud and Internet industries. His
experience includes business and digital transformation initiatives across
finance, operations, IT, HR, customer support and infrastructure functions,
particularly their impact on the overall finance function and the role of the
CFO during business model transitions to SaaS and cloud services.

Gori von Hirschhausen

Partner, PwC Germany
+49-89-5790-6698
[email protected]

Gori von Hirschhausen is the initiative leader for PwC’s CFO think tank,
Future of Finance, and finance consulting leader in Europe. Throughout
his almost 20 years in management consulting, he has supported a large
number of global corporate transformations. He is a well-known and
esteemed keynote speaker at client symposiums, CFO conventions,
academic institutions and universities around the world.

Anand Rao
Principal, PwC US
+1 617 530 4691
[email protected]

As the global and US leader for PwC’s artificial intelligence practice, Anand
leads efforts to bring innovative AI methods in machine learning, deep
learning, natural language processing, simulation and reinforcement learning
to practical client problems to enhance and augment human decision
making. He advises clients in financial services, insurance, healthcare, auto,
and technology companies.
13 | PwC How AI will transform the CFO’s role

Brian Furness
Partner, PwC UK
+44-0-77409-23515
[email protected]

Brian Furness is the global leader of PwC’s finance consulting business

across all industry sectors. He advises CFOs and their finance teams on
a broad range of topics, including operating model design, technology
and automation within finance, how to build an efficient and effective
finance team, regulatory compliance, process redesign/standardisation,
developing talent within the finance team, internal and external reporting,
and financial control.

He regularly speaks at conferences and CFO events and is the coauthor

of PwC’s Finance Effectiveness Benchmarking Study.

Florian Puppe
Manager, PwC Germany
+49-160-530-3724
[email protected]

Florian Puppe is a management consultant with a strong background on

back-office transformation and digital finance. He is part of PwC’s Future of
Finance initiative and is leading the finance innovation lab in Europe.

The following PwC colleagues also contributed to this report:

Michael Baccala, Hamish Clark, Michael Graham, Kevin Kroen,
Abhishek Mitra, Austin Morris and Stephen Pillsbury.
www.pwc.com/cfo-artificial-intelligence
At PwC, our purpose is to build trust in society and solve important problems. We’re a network of firms in 158 countries with more than 236,000 people who are
committed to delivering quality in assurance, advisory and tax services. Find out more and tell us what matters to you by visiting us at www.pwc.com.
This publication has been prepared for general guidance on matters of interest only, and does not constitute professional advice. You should not act upon the
information contained in this publication without obtaining specific professional advice. No representation or warranty (express or implied) is given as to the
accuracy or completeness of the information contained in this publication, and, to the extent permitted by law, PwC does not accept or assume any liability,
responsibility or duty of care for any consequences of you or anyone else acting, or refraining to act, in reliance on the information contained in this publication or
for any decision based on it.
© 2019 PwC. All rights reserved. PwC refers to the PwC network and/or one or more of its member firms, each of which is a separate legal entity.
Please see www.pwc.com/structure for further details.