Skip to main content

The implications of generative AI in Finance

A new frontier in artificial intelligence and for Finance

By the end of 2022, generative artificial intelligence (AI) had fully sparked public imagination about the future. Consumers and enterprises alike have been using artificial intelligence for years: speaking to voice assistants like Siri or Alexa, automating routine tasks, using algorithms to recognise patterns and correlations in data. But applications that could produce original text or digital art that quickly became household names, heralded a new age where AI can mimic human creative processes.

How much will generative AI affect our lives? That remains to be seen; however, research has suggested that breakthroughs in generative AI could increase global GDP by 7%—nearly $7 trillion—and boost productivity growth by 1.5 percentage points.1

Generative AI might start by producing concise and coherent summaries of text (e.g., meeting minutes), converting existing content to new modes (e.g., text to visual charts) or generating impact analyses from, say, new regulations. Producing novel content represents a definitive shift in the capabilities of AI, moving it from an enabler of our work to a potential co-pilot. It might augment human capability and then advance it.

CFOs and Finance leaders should start to develop strategies today for how generative AI will affect both their functions and their businesses tomorrow.

Getting to know generative AI

To consider the potential impact of generative AI in Finance, it’s helpful to explore what generative AI is and how it operates.

Generative AI:

  • Creates original content such as text, images, audio, code and video. Up until now, these types of content required solely human skill and expertise to create.
  • Can add contextual awareness and human-like decision-making to enterprise and finance workflows, potentially dramatically changing how work is conducted.
  • Is powered by foundation models, which run on deep-learning algorithms modelled on the organisation of neurons in the human brain.

What does this mean for Finance?

A lot. Generative AI is powered by data. And since Finance draws upon enormous amounts of data, it’s a natural fit to take advantage of generative AI. 

Here we consider areas within Finance primed for generative AI-enabled transformation:

Generative AI has the potential to transform Finance and business, as we know it. It will take upfront investment in time and money. According to a Gartner study, 80% of CFOs surveyed in 2022 expected to spend more on AI in the coming two years.2 With that investment, however, around two-thirds think their function will reach an autonomous state within six years.

To make sound decisions, it will be crucial that leaders consider the use of generative AI from an enterprise-wide approach with a clear understanding of where this technology will have an impact on operating expenditures, capital expenditures, market capitalisation and a lot more. CFOs and Finance leaders can play a pivotal role in driving strategic collaboration among key C-suite leaders to enable greater success—and return on investment—of AI deployment and adoption. The journey should begin with a sound strategy and a few use cases to test and learn with well-governed and accessible data. It doesn’t have to be perfect, but it should be controlled.

While the future looks promising, generative AI has some current limitations that Finance professionals should consider. 

Humans train the foundation models that power generative AI and humans bring to the work inherent biases. Bias in equals bias out. If the underlying training data is skewed (e.g., an over- or underrepresentation of a population cohort), then these biases may be amplified in the content generated with results counter to an organisation’s diversity, equity and inclusion commitments. Other types of bias, such as anchoring bias in forecasting, where training data relies too heavily on certain pieces of information to make subsequent judgements, can specifically affect finance processes using generative AI.

Finance leaders must remain on guard for biases in training data and regularly evaluate content. As detailed in our Trustworthy AI™ framework, organisations can design new processes to break down bias and utilise audit trails that can trace the lineage of data used to generate content.

Foundation models may produce incorrect but confident responses to prompts. Referred to as hallucinations, the incorrect content arises from the models still learning. But they are like overconfident humans: They don’t recognise when they could be wrong. This is less likely in simple equations, but when subjectivity is introduced, when the model must make choices, the potential for generative AI to produce inaccurate content may increase.

One expects that this will improve with time. You may want to restrict initial usage to increase accuracy of inferences and then expand and scale models. Perhaps, one day, models will even have to pass a certification test to provide finance-related insights, advice and engagement. Regardless, Finance leaders would do well to remain vigilant in validating and certifying content.

Today, most generative AI applications do not guarantee privacy of data. Public consumer applications use existing content and prompts as part of their ongoing learning process. Information shared might show up again with another user altogether.

Many enterprises may end up with private models trained in secure environments that can help negate this risk. Regardless, safeguarding sensitive financial information will be critical. Finance leaders should partner with their enterprise technology teams to ensure that generative AI security approaches align to companywide standards.

Generative AI does not use discretion in how it shares content. Nor can it necessarily comply with different company policies and country privacy laws, especially for global companies where those might differ widely. For example, enterprises might have contracts that restrict the types of third-party data they can use. By itself, generative AI cannot necessarily comply with these terms; however, user controls might offer safeguards in restricting access to data.

So, leaders must continuously ask themselves: Is generative AI being used in a manner consistent with its purpose?

Implications on Finance talent

For all its tantalising potential to automate and augment processes, generative AI will still require human talent.

But workers are sceptical. In a recent Harris Poll of workers, about half do not trust the technology.3 Finance leaders should consider change management carefully, leaning into the idea that generative AI can support our lives, transforming from an enabler of our work to a potential co-pilot.

Certainly, roles and responsibilities will change and new types of workers may be needed. But enduring human capabilities like curiosity, empathy, critical thinking, teaming and more will still be needed. Finance professionals may need to develop new generative AI fluency and enhance skills such as:

  • How to engineer prompts (i.e., ask good questions) to get desired results.
  • How to recognise potential bias.
  • How to confirm the quality and validity of generated output and monitor performance of models over time.

Generative AI has quickly cemented its foothold in the public consciousness. It has sparked excitement around productivity increases and cost savings but also warrants caution.

Yet, generative AI may completely transform the Finance function. Leading organisations have launched pilot programmes and are scaling fast.

In the short term, generative AI will allow for further automation of financial analysis and reporting, enhancement of risk mitigation efforts and optimisation of financial operations. The impact is unlikely to stop there, though. With its ability to process vast amounts of data and quickly produce novel content, generative AI holds a promise for progressive disruptions we cannot yet anticipate.

Generative AI is here and it’s adapting and learning.

CFOs and Finance executives who adapt and learn with it can lead the future.

Ready to discuss generative AI in Finance?
Ready to discuss generative AI in Finance?

Contact us to learn more

Gina Schaefer
Managing Director, Analytics and Cognitive
Deloitte Consulting LLP
+1.404.631.2326
gschaefer@deloitte.com

Christine Ahn
Principal, Finance and Performance
Deloitte Consulting LLP
+1.213.553.1691
chrisahn@deloitte.com

James Glover
Principal, Finance and Performance
Deloitte Consulting LLP
+1.212.313.1916
jglover@deloitte.com

Robyn Peters
Senior Manager, Finance and Performance
Deloitte Consulting LLP
+1.214.840.1475
robynpeters@deloitte.com

Ranjit Rao
Principal, Finance and Performance
Deloitte Consulting LLP
+1.404.631.3661
ranjrao@deloitte.com

Cameron Andriola
Senior Manager, Controllership Transformation
Deloitte & Touche LLP
+1.702.893.5148
candriola@deloitte.com

Court Watson
Partner, Controllership Transformation
Deloitte & Touche LLP
+1.206.716.7082
cowatson@deloitte.com

Geoff Kovesdy
Principal, Internal Audit
Deloitte & Touche LLP
+1.212.436.5149
gkovesdy@deloitte.com

Prashant Patri
Principal, Treasury
Deloitte & Touche LLP
+1.212.436.7568
prpatri@deloitte.com

RJ Littleton
Partner, Tax
Deloitte Tax LLP
+1.214.840.1781
rilittleton@deloitte.com

Mark Gustafson
Senior Manager, Human Capital
Deloitte Consulting LLP
+1.213.688.4101
margustafson@deloitte.com

Arjun Krishnamurthy
Managing Director, SAP
Head of AIOPS.D™ Autonomous Finance
Deloitte Consulting LLP
+1.404.942.6967
arkrishnamurthy@deloitte.com

Sonal Sood
Senior Manager, Oracle
Deloitte Consulting LLP
+1.312.486.5096
ssood@deloitte.com

Goldman Sachs, “Generative AI could raise global GDP by 7%,” 5 April 2023.

Gartner, “Gartner predicts three ways autonomous technologies will affect the FP&An and Controller functions in Finance,” press release, 1 March 2023.

Andrew Laningham, “What’s missing in the conversation about generative AI and jobs,” The Harris Poll, 6 February 2023.

Did you find this useful?

Thanks for your feedback