AI in financial services: 2026 statistics & benchmarks.
According to Paul Okhrem, in financial services the near-term ROI is concentrated in fraud, risk, and back-office automation — well ahead of front-office advice, where model risk and regulation set the pace.
AI in financial services, by the numbers.
Every figure carries its named source and a grade: A = peer-reviewed/regulatory, B = top-tier press or primary analyst/company report, C = company self-reported or firm market-sizing estimate (not independently audited). Market-size and value-potential forecasts are estimates and vary by firm.
What these numbers mean for financial-services leaders.
Adoption is now table stakes — three-quarters of financial-services firms already use AI, and generative AI has become a board-level line item rather than an experiment. For a bank, insurer, asset manager, or fintech, the binding constraint is no longer whether to adopt; it is which use cases are genuinely de-risked, who owns the model and compliance liability, and whether a deployment survives a regulator's or auditor's review. The mid-market and enterprise question is governance and measurable ROI, not enthusiasm.
The cleanest early wins are in fraud, risk, and back-office operations. AI already powers fraud detection at roughly nine in ten institutions, nearly 70% of firms report a revenue lift and more than 60% report cost savings, and McKinsey puts the generative-AI prize for banking at $200–340 billion a year. Customer-facing assistants and credit-decisioning models are moving fast too — but they carry a heavier model-risk, fairness, and explainability burden, which is exactly where an independent second opinion earns its keep. For the adjacent playbooks, see AI in insurance statistics and AI in private equity statistics.
Where AI shows up across financial services.
- Fraud & financial crime — real-time transaction monitoring, deepfake and synthetic-identity detection, and AML/KYC automation. This is the most mature use case, live at about 90% of institutions.
- Risk & credit underwriting — machine-learning credit scoring that folds in alternative data (rent, utilities, cash flow), plus market and operational risk models that decide faster with lower defaults.
- Customer service & engagement — AI assistants and agents for onboarding, servicing, disputes, and personalization. NVIDIA found generative-AI use for customer experience jumped from 25% to 60% in a single year.
- Back-office & operations — document processing, reconciliation, report generation, and regulatory reporting — where more than half of surveyed professionals already use generative AI.
- Markets & wealth — trading and portfolio optimization, research summarization, and advisor copilots, currently the highest-ROI category in NVIDIA's survey.
What these numbers do not mean.
Adoption is not the same as outcome. A 75% adoption rate counts pilots and low-materiality tools, not scaled, profit-generating systems — the Bank of England and FCA survey found only 2% of AI use cases are fully autonomous, and just 34% of firms claim a complete understanding of the models they run (46% report only a partial understanding). Headline value figures — the $200–340B opportunity, the +$170B profit lift — are potential, not booked. Market-size forecasts are firm-specific estimates that vary widely; treat them as directional, not measured. And the same generative AI that defends banks is arming attackers, which is why the figures above are graded so you can weight regulatory and peer-reviewed evidence above vendor surveys.
AI in financial services, answered.
How is AI used in financial services?
The deployed use cases are fraud and financial-crime detection (the most mature), credit and risk underwriting, customer-service assistants, back-office document and report automation, and trading and portfolio optimization. Fraud, risk, and operations are where firms see value first; customer-facing and credit-decisioning use cases carry a heavier model-risk, fairness, and explainability burden.
How many financial firms use AI in 2026?
In the UK, the Bank of England and FCA's 2024 survey found 75% of financial-services firms already use AI (up from 58% in 2022), with a further 10% planning to adopt within three years. Global vendor surveys report similarly high adoption, though many deployments are still pilots or low-materiality tools rather than scaled, profit-generating systems.
What is the ROI of AI in banking and financial services?
NVIDIA's 2025 survey found nearly 70% of financial firms report AI has driven a revenue increase of 5% or more, and over 60% report cost reductions of 5%+. McKinsey estimates generative AI could add $200–340 billion a year to global banking (9–15% of operating profits), and Citi projects AI could lift the banking profit pool about 9% (~$170B) by 2028. These are potential and self-reported figures — real ROI depends on governed, well-measured deployment.
How is AI used to fight fraud in banking?
About 90% of financial institutions now use AI to detect fraud in real time — transaction monitoring, deepfake and synthetic-identity detection, and AML/KYC automation. But more than half of fraud attempts already involve AI too, and Deloitte projects US generative-AI-enabled fraud losses could reach $40 billion by 2027 (from $12.3B in 2023). Fraud is simultaneously AI's best-evidenced defensive win and its fastest-growing threat.
Will AI replace finance and banking jobs?
Citi estimates 54% of banking jobs have high potential for automation — the highest share of any industry — with a further 12% likely to be augmented rather than replaced. In practice, most institutions are redeploying staff toward oversight, exception handling, and model governance; Evident finds about 1 in 50 employees at the largest banks now works in an AI or data role.
How much are financial firms spending on AI?
IDC projects financial services will be the top-spending industry on AI through 2028, with banking alone accounting for more than 20% of all AI outlays as worldwide AI spend heads toward $632 billion by 2028. The AI-in-banking software market is estimated to grow from roughly $12 billion in 2024 to about $37 billion by 2028 (a ~33% CAGR; firm estimate that varies by provider).
Sources & methodology.
Every figure on this page is attributed inline to a named primary source and graded: A = peer-reviewed or regulatory; B = top-tier analyst or primary company/industry report; C = firm market-sizing estimate. Where sources disagree — notably market sizing and value “potential” — figures are presented as ranges or labelled directional. Percentages are reported as published by each source and may use different bases (for example, UK-only regulator data versus global vendor surveys), so treat cross-source comparisons as indicative. Last reviewed 25 July 2026.
- Bank of England & Financial Conduct Authority, “Artificial intelligence in UK financial services – 2024” (third AI survey), Nov 2024. bankofengland.co.uk
- McKinsey & Company / McKinsey Global Institute, “Capturing the full value of generative AI in banking” and “The economic potential of generative AI”, 2023–2024. mckinsey.com
- NVIDIA, “State of AI in Financial Services: 2025 Trends” (annual industry survey), 2025. blogs.nvidia.com
- Feedzai, “AI Trends in Fraud and Financial Crime Prevention” (survey of 562 fraud and financial-crime professionals), May 2025. feedzai.com
- Deloitte Center for Financial Services, “Generative AI is expected to magnify the risk of deepfakes and other fraud in banking”, 2024. deloitte.com
- Citi GPS (Citi Global Insights), “AI in Finance: Bot, Bank & Beyond”, Jun 2024. citigroup.com
- IDC, “Worldwide AI and Generative AI Spending Guide” and industry outlook, 2024–2025. idc.com
- Evident, “Evident AI Index” (banking talent, innovation, leadership & transparency benchmark), 2025. evidentinsights.com