AI in Finance Statistics: 29 Numbers That Matter in 2026
Updated August 2026
How widely is AI actually being used in finance? The answer depends on what you measure.
Large-enterprise surveys increasingly show AI embedded across finance and other corporate functions. Broader government data show a much lower adoption rate across the full population of U.S. businesses. Surveys of finance leaders report positive returns, while workforce research shows that companies are still figuring out how to convert time saved into durable business value.
That disagreement is useful. It suggests a market that has moved well beyond experimentation but is still uneven in maturity.
The 29 statistics below cover six questions that matter most to finance leaders: adoption, ROI and productivity, finance use cases, AI agents, talent and skills, and investment.
I have linked each figure to the organization that published it and preserved the relevant population or sample information where available.
How to Read These Statistics
The numbers should not be treated as if they came from one dataset.
A Gartner survey of senior finance leaders is measuring something different from Census Bureau data covering the broader U.S. business population. A survey asking whether a company is "using AI" will produce a different result from one asking whether AI is deployed extensively in finance. Likewise, reported ROI generally reflects respondents' assessments, not independently audited financial returns.
The more useful question is not whether one percentage is the "correct" adoption rate. It is what the evidence collectively says about where finance is heading.
My reading is fairly straightforward: AI use in finance is now common, but depth varies considerably; productivity benefits are increasingly measurable, but organizational value is harder to prove; agentic AI is moving quickly from interest toward deployment; and the harder problems are increasingly skills, workflow design, data, controls, and management rather than access to the technology itself.
AI Adoption in Finance
of finance leaders reported using AI in their finance function in Gartner's 2025 survey, essentially flat from 58% in 2024 after rising sharply from 37% in 2023.
That is an important qualification to the AI-adoption narrative: adoption is high, but the initial acceleration has slowed.
Gartner, 2025 AI in Finance Survey · 183 CFOs and senior finance leaders · 2025
of CFOs at large North American companies said their organizations use AI extensively or modestly across multiple key functions and operations.
This is broader than finance-specific AI use. It shows how difficult it has become to find a large enterprise where AI is absent altogether.
Deloitte, Q2 2026 CFO Signals · 200 North American CFOs at companies with at least $1 billion of revenue · 2026
of U.S. companies in KPMG's 2024 study reported using AI in finance to a moderate or large degree, while 58% were piloting or deploying generative AI.
The survey included 300 U.S. finance leaders within a larger global study of 2,900 companies.
KPMG, AI in Finance · 2024
of U.S. businesses in the finance and insurance sector reported using AI in May 2026, compared with 19.8% across U.S. businesses overall.
This is one of the most useful counterweights to corporate surveys because Census data cover the broader business population, not primarily large enterprises or senior finance leaders.
U.S. Census Bureau, Business Trends and Outlook Survey · 2026
of U.S. workers reported using generative AI for work to some degree in the first nationally representative U.S. survey of workplace generative AI adoption.
The underlying survey was conducted in 2024 and published by the Federal Reserve Bank of St. Louis in 2025.
AI ROI and Productivity
of respondents in KPMG's 2026 global finance study said AI ROI was meeting or exceeding expectations: 46% said it was meeting expectations and 28% said it was exceeding them.
That is encouraging, but it is self-reported ROI. It should not be interpreted as 74% of AI projects producing a verified financial return.
KPMG, AI in Finance: The Decision Advantage · 1,013 senior finance leaders across 20 countries · 2026
of surveyed companies adopting AI agents reported measurable productivity gains. Fifty-seven percent reported cost savings and 55% reported faster decision-making.
The distinction matters: the respondents were companies already adopting agents, rather than a representative sample of all companies.
PwC, AI Agent Survey · 300 senior executives · 2025
of frontline employees who regularly use AI reported saving at least eight hours per week, roughly one workday.
The more interesting finding may be what happens afterward: 66% said they receive limited or no guidance on what to do with the time AI saves.
BCG, AI at Work · nearly 12,000 workers, managers, and leaders · 2026
per 40-hour week was the average self-reported time savings among workers using generative AI in Federal Reserve Bank of St. Louis research.
That equals approximately 5.4% of working hours among users.
by 2029 is Gartner's forecast for organizations whose CFOs strategically deploy AI and broader finance technology portfolios.
This is a forecast, not an observed return. I would treat it as an indication of Gartner's view of the potential economic stakes rather than evidence that AI has already produced ten points of margin.
Gartner, Finance Technology Forecast · survey of 314 organizations · 2026
Where Finance Teams Are Using AI
of finance organizations that had implemented AI reported using it for knowledge management, Gartner's most common finance AI use case.
Accounts payable process automation followed at 37%, with error and anomaly detection at 34%.
of U.S. companies in KPMG's survey were piloting or using AI in financial planning.
Accounting was close behind at 76%, followed by treasury management at 64%.
KPMG, AI in Finance · 2024
of U.S. companies surveyed by KPMG reported using AI specifically in financial reporting.
In the same study, 92% said their finance AI initiatives were meeting or exceeding their ROI expectations.
KPMG, AI in Finance · 2024
of CFOs in Deloitte's Q2 2026 survey said their finance functions use AI for operational productivity tasks.
Forty-four percent reported using AI for planning and budgeting, while 41% were using it to analyze financial data for insights.
AI Agents in Finance
of surveyed U.S. companies expect to be deploying or scaling AI within their finance functions over the following 18 months.
Half were already planning to orchestrate or develop multi-agent systems across finance workflows.
This is a statement of planned adoption, not current deployment.
KPMG, AI in Finance: The Decision Advantage · global survey of 1,013 senior finance leaders, including 163 in the U.S. · 2026
of executives surveyed by PwC said AI agents were already being adopted somewhere in their companies.
Only 34% reported using agents in accounting and finance.
That gap is probably more useful than either figure by itself: enterprise interest in agents is ahead of deployment inside finance.
PwC, AI Agents for Finance · 2025
of executives surveyed by PwC agreed that AI agents would drastically transform existing roles within 12 months.
At the same time, 48% expected agentic AI to result in increased headcount.
That combination is a useful reminder that automation and employment are not necessarily simple substitutes.
PwC, AI Agent Survey · 2025
of CFOs said integrating AI agents into finance was a transformation priority for 2026.
That ranked ahead of improving data quality, access, and usability at 52% in Deloitte's survey.
Deloitte, Q4 2025 CFO Signals · 200 North American CFOs · published 2026
of CFOs expected AI to be extremely or very important to their finance department's operations in 2026.
Only 2% said AI would not be important.
Finance Talent and AI Skills
of senior finance and accounting leaders said AI would be the most transformative technology trend affecting the profession over the following 12 to 24 months.
Only 8% said their organizations were very well prepared to manage that change.
The gap between expected importance and organizational readiness is one of the clearest findings in this entire dataset.
AICPA & CIMA, Future-Ready Finance Survey · 1,446 senior finance and accounting leaders and managers · 2025
of finance and accounting leaders identified generative AI as their most prominent skills gap.
Half cited human capital, skills, and talent as the largest barrier to adopting new technology.
of leaders surveyed by Microsoft said their organizations were considering hiring for AI-specific roles over the following 12 to 18 months.
At the same time, 33% were considering headcount reductions.
Both can be true. Companies can automate portions of existing jobs while simultaneously creating demand for different skills and roles.
Microsoft, 2025 Work Trend Index · 31,000 workers across 31 markets · 2025
of CFOs and tax leaders surveyed by EY said fewer new accountants entering the profession would hurt their functions.
Sixty-one percent expected retirements among senior professionals to have a significant impact.
AI is therefore arriving while finance and accounting are already dealing with structural talent pressure.
EY, 2025 Tax and Finance Operations Survey · approximately 1,600 leaders · 2025
of frontline employees in BCG's 2026 survey said they now use AI every day or several times per week.
That was up 23 percentage points from the prior year.
The technology-adoption problem is therefore increasingly becoming a management problem: what work should change once employees have the tools?
BCG, AI at Work · 2026
of CFOs named automating processes so employees can perform higher-value work as their leading finance talent priority for 2026.
That framing matters. The objective is not simply automation. It is deciding what employees should do with the capacity automation creates.
AI Investment and Finance Budgets
was Gartner's forecast for worldwide AI spending in 2025, with spending expected to exceed $2 trillion in 2026.
This is not finance-function spending. It is useful as context for the broader investment environment surrounding finance AI.
of CFOs were increasing technology budgets for 2026, according to Gartner.
Nearly half were increasing those budgets by 10% or more.
Gartner, Finance Technology Survey · 314 organizations · 2026
of senior executives surveyed by PwC said their function planned to increase AI-related budgets over the following 12 months because of agentic AI.
More than one-quarter expected increases of at least 26%.
PwC, AI Agent Survey · 2025
more budget could be redirected toward strategic and value-generating activities through AI, according to finance and tax leaders surveyed by EY.
Separately, 86% said leveraging data, generative AI, and technology was a leading priority.
What the AI in Finance Data Actually Says
I would draw five conclusions from these statistics.
AI adoption in finance is no longer unusual. Whether the appropriate number is 34%, 59%, 62%, or substantially higher depends on the population and definition. The direction is much clearer than the precise percentage.
Using AI and transforming finance with AI are different things. Drafting emails, summarizing meetings, and performing research count as adoption. They do not necessarily indicate that the close, forecast, reporting process, or finance operating model has materially changed.
There is credible evidence of productivity gains, but ROI claims require more scrutiny. Workers consistently report time savings, and large shares of executives report positive returns. But self-reported ROI is not the same as measured incremental cash flow, margin, or enterprise value.
Agents are arriving faster than operating models are changing. Companies are investing in agentic systems and finance leaders expect them to become important. The harder question is who owns them, how they are controlled, where humans review their work, and what jobs look like after parts of a workflow become autonomous.
Skills are a major constraint, but not the only one. Finance organizations also cite data quality, security, legacy systems, governance, unclear use cases, and change management. Training matters most when it is connected to actual workflows rather than treated as a standalone technology exercise.
That last distinction matters to me.
The objective of AI adoption in finance should not be to create a department full of people who know more about AI. It should be to produce a finance organization that makes better decisions, completes lower-value work more efficiently, exercises stronger judgment, and redirects its people toward work where they add more value.
AI in Finance FAQ
What percentage of finance teams use AI?
There is no single defensible percentage because the major studies measure different things.
Gartner found that 59% of surveyed finance leaders were using AI in their finance functions in 2025.
KPMG found that 62% of U.S. companies in its 2024 survey were using AI in finance to a moderate or large degree.
The U.S. Census Bureau found that 33.9% of businesses in the broader finance and insurance sector were using AI as of May 2026.
Deloitte found that 93% of CFOs at large North American companies said their organizations were using AI across multiple key functions and operations in 2026.
Those findings are not directly contradictory. They cover different populations, company sizes, definitions of AI use, and organizational scopes.
The defensible conclusion is that AI has become common in large finance organizations, while adoption across the broader business population remains materially lower.
How are finance teams using AI?
Current use is spread across both general productivity and finance-specific workflows.
Common applications include knowledge management, research, accounts payable automation, anomaly detection, financial planning and budgeting, financial data analysis, financial reporting, treasury, and forecasting.
Gartner's 2025 survey found knowledge management to be the most common finance-specific use case among respondents that had implemented AI, while Deloitte's 2026 CFO survey found operational productivity to be the most common use among its respondents.
That is consistent with what I see in practice: the easiest starting point is usually knowledge work, while deeper finance transformation requires more integration, controls, and process redesign.
What ROI are finance teams getting from AI?
The evidence is positive but should be interpreted carefully.
KPMG reported in 2026 that 74% of surveyed finance leaders said AI ROI was meeting or exceeding expectations. Its 2024 U.S. study produced an even higher figure of 92%.
Worker-level studies provide a more concrete measure of one source of value. Federal Reserve Bank of St. Louis research estimated average self-reported savings of 2.2 hours per 40-hour week among generative AI users, while 42% of regular frontline AI users in BCG's 2026 survey reported saving at least eight hours per week.
The unresolved question is how much of that capacity becomes economic value.
Saving four hours is not the same as reducing four hours of labor cost. The organization has to decide how that capacity will be redeployed and whether it produces better output, faster decisions, additional revenue, lower cost, or reduced risk.
What is the biggest barrier to AI adoption in finance?
Skills are one of the clearest barriers.
AICPA & CIMA found that 56% of finance and accounting leaders identified generative AI as their most prominent skills gap, while 50% cited human capital, skills, and talent as their biggest barrier to technology adoption.
But I would not reduce the issue to training alone.
Other studies identify data quality, data security, legacy systems, governance, organizational adoption, unclear role-specific use cases, and lack of hands-on practice as significant constraints.
The finance organizations that struggle are often not missing access to an AI model. They are missing the operating conditions required to use it well.
How fast are AI agents coming to finance?
Quickly, although planned adoption is ahead of actual finance deployment.
PwC found that 79% of surveyed executives said agents were already being adopted somewhere within their companies, but only 34% reported agent use in accounting and finance.
KPMG reported that half of surveyed U.S. companies were planning to develop or orchestrate multi-agent systems across finance workflows.
Deloitte found that 54% of CFOs considered integrating AI agents into finance a transformation priority for 2026.
The more consequential transition will occur when agents move from assisting with individual tasks to performing multi-step finance workflows.
That creates a different set of management questions around authorization, review, evidence, segregation of duties, controls, accountability, and human judgment.
Will AI reduce finance and accounting jobs?
The evidence does not support a simple yes-or-no answer.
Microsoft found that 33% of surveyed leaders were considering headcount reductions, while 78% were also considering hiring for AI-specific roles.
Meanwhile, finance organizations are already facing shortages of accounting talent and expected retirements among experienced professionals.
The more likely near-term outcome is significant job redesign. Some tasks will require fewer human hours, while other work involving judgment, controls, business partnering, AI supervision, data, and workflow design becomes more important.
That can reduce demand for particular activities without eliminating the need for finance professionals.
What finance jobs are most likely to change because of AI?
Roles with large amounts of repeatable research, reconciliation, document review, variance analysis, data preparation, reporting, drafting, and information synthesis are likely to change quickly.
That does not mean those jobs disappear.
The work tends to move toward reviewing exceptions, interpreting outputs, exercising judgment, communicating conclusions, designing processes, and supervising automated workflows.
The distinction between doing the first draft and being accountable for the answer will become increasingly important.
What skills should finance professionals develop for AI?
I would separate AI skills into three levels.
The first is individual fluency: knowing how to work effectively with models, provide context, evaluate outputs, research, analyze, and automate personal workflows.
The second is workflow design: identifying which steps should be performed by a human, an AI model, traditional automation, or an agent, and designing the handoffs between them.
The third is judgment and governance: understanding when an output is reliable enough to use, what evidence is required, what controls should exist, and who remains accountable.
Finance professionals still need accounting, finance, commercial judgment, communication, and critical thinking. AI increases the leverage on those skills; it does not make them irrelevant.
How should CFOs measure AI ROI?
I would avoid measuring success primarily through licenses purchased, employees trained, prompts submitted, or applications built.
Measure changes in the underlying work.
Useful metrics can include:
- cycle time;
- hours required per recurring process;
- cost per transaction or deliverable;
- forecast accuracy;
- error and exception rates;
- time from data availability to management insight;
- percentage of work automated or augmented;
- employee capacity redirected to higher-value activities;
- adoption of reusable workflows;
- control failures or overrides; and
- incremental revenue, margin, or cost savings where they can be reasonably attributed.
The metric should follow the use case.
An AI forecasting tool should eventually improve forecasting economics or decision quality. An AI contract-review workflow should reduce review effort or improve identification of risk. A reporting tool should shorten the reporting cycle or improve the quality of analysis.
"People are using it" is an adoption metric, not an ROI metric.
The Question Finance Leaders Should Ask Next
The question in 2024 was whether finance professionals would use generative AI.
By 2026, that is increasingly settled.
The more useful questions are what work should change, which use cases produce enough value to scale, where agents should be allowed to act, how finance controls AI-assisted processes, and what people should do with the capacity the technology creates.
That is a harder problem than buying licenses.
It is also where most of the remaining value is.
For what the numbers mean for your team, my approach is the AI Adoption Method, the readiness baseline is the free AI Readiness Assessment, and the hands-on side is training.