The European Banking Authority (EBA) observes 92% of EU banks are currently deploying AI, and 8% are pilot testing or discussing AI use cases. Among large banks under European Central Bank (ECB) supervision, more than 85% already use AI in some form. These AI in banking statistics come from regulators that each count a different population.
In the UK, the Bank of England and Financial Conduct Authority (FCA) found 75% of firms that responded to the survey are already using AI, up from 58% in the 2022 survey. Launches are accelerating, as the 50 banks tracked in the Evident AI Index for Banks announced 93 new AI use cases in Q2 2026. Yet only 27% of new use cases carried a disclosed outcome in Q2, down from 41% in the previous quarter.
Key Takeaways
- The insurance sector reported 95% current AI use, closely followed by international banks at 94%, according to the Bank of England and FCA survey.
- 55% of surveyed EU banks are already using General-Purpose AI (GPAI) or agentic AI in consumer-facing processes, according to the EBA.
- Agentic AI hit a record 31% of new use cases in Q1 2026, up from 15% in Q4 2025, according to Evident.
- Only 2% of use cases have fully autonomous decision-making among UK financial firms surveyed by the Bank of England and FCA.
- JPMorganChase, Capital One and Royal Bank of Canada (RBC) remain the top three of 50 banks in the October 2026 Evident AI Index.
- 83% of consumers who used an AI assistant to research or make a personal financial decision verified the information with another source or professional, according to TD Bank.
- US Treasury fraud processes, including machine learning AI, prevented and recovered over $4 billion in fiscal year 2024, up from $652.7 million in fiscal year 2023.
Editor’s Choice
- New AI use cases at tracked banks in Q2 2026: 93.
- Erica client interactions since launch: surpassing 3 billion.
- Bank of America employees using EricaAssist: more than 18,000.
- AI patents filed by Capital One: more than 1,700.
- Bank of America annual technology spend: $14 billion.
- Evident AI Index banks putting a number on group-wide AI value: 16 of 50.
AI Adoption Rates in Banking
- The EBA observes 92% of EU banks are currently deploying AI, and 8% are pilot testing or discussing AI use cases.
- More than 85% of large banks under European supervision already use AI in some form, and adoption rates continue to increase year after year, according to the ECB.
- 75% of firms that responded to the survey said they are already using AI, with a further 10% planning to use AI over the next three years, in the Bank of England and FCA study.
- In the 2022 survey, a further 14% of firms were planning to use AI.
| Measure | Population | Share | Period |
|---|---|---|---|
| Deploying AI | EU banks (EBA) | 92% | September 2025 |
| Pilot testing or discussing AI | EU banks (EBA) | 8% | September 2025 |
| Using AI in some form | Large banks under European supervision (ECB) | More than 85% | February 2026 |
| Using GPAI or agentic AI with consumers | Surveyed EU banks (EBA) | 55% | September 2025 |
| Using AI | UK financial services respondents | 75% | 2024 survey |
| Using AI | UK financial services respondents | 58% | 2022 survey |
Source: European Banking Authority, European Central Bank, Bank of England and FCA
About This Data
Figures are compiled from fifteen sources (five regulator and government publications, four bank press releases and six industry research and banking association reports), published between October 2024 and October 2026. Most figures come from primary publications and named survey or index owners, with two American Bankers Association Banking Journal reports cited alongside the CSBS and TD Bank releases they cover. Figures are reviewed on a rolling basis and updated when sources publish new editions.
What percentage of banks are using AI?
92% of EU banks are currently deploying AI, according to the EBA, and more than 85% of large banks under European supervision already use AI in some form. In the UK, 75% of firms that responded to the survey said they are already using AI. Each figure covers a different group of firms, so the rates are best read side by side rather than averaged.
UK Financial Services AI Adoption by Sector
- The insurance sector reported the highest percentage of firms currently using AI at 95%, closely followed by international banks at 94%.
- Financial market infrastructure firms are the sector with the lowest percentage of firms currently using AI.
- The survey received 118 responses, and UK banks was the largest sector with 50 respondents.
- Large UK and international banks reported a median of 39 and 49 use cases respectively, against an overall median of nine use cases.
- Respondents expect the median number of use cases to more than double over the next three years, from 9 to 21.
View the data
International banks sit 19 points above the all-firm average.
Recent Developments: AI in Banking Statistics
- October 2026: JPMorganChase, Capital One, and RBC remain the top three in the Evident AI Index for Banks, and AI deployment across the sector moved nearly three times faster than in previous years.
- October 2026: In the 2026 index, Bank of America rose 3 places to rank 7 and Citigroup rose 4 places to rank 8.
- October 6, 2026: A TD Bank US survey found that 69% would allow a bank-provided AI assistant to complete at least one financial task on their behalf.
- September 2026: The Conference of State Bank Supervisors released a new tool to help state examiners understand and assess risks from artificial intelligence at financial institutions, and the framework was approved by the CSBS State Supervisory Processes Committee (SSPC) and the CSBS Non Depository Supervisory Committee (NDSC) in August 2026.
- July 21, 2026: Bank of America added Generative AI (Gen AI) capabilities to EricaAssist that deliver contextual guidance in under three seconds.
Evident AI Index 2026 Bank Rankings
- JPMorganChase, Capital One, and RBC remain the top three, and every bank in the top 10 has placed there before.
- JPMorganChase ranks 1 on leadership and transparency and 2 on both talent and innovation.
- Capital One ranks 1 on talent and innovation but 40 on leadership, down 20 places.
- Morgan Stanley fell 4 places to rank 9, while HSBC fell 3 places to rank 11 and Goldman Sachs fell 3 places to rank 12.
| Rank | Bank | Change vs 2025 | Talent rank | Innovation rank |
|---|---|---|---|---|
| 1 | JPMorganChase | No change | 2 | 2 |
| 2 | Capital One | No change | 1 | 1 |
| 3 | Royal Bank of Canada | No change | 8 | 5 |
| 4 | CommBank | No change | 3 | 11 |
| 5 | Wells Fargo | Up 1 | 6 | 6 |
| 6 | UBS | Up 1 | 4 | 17 |
| 7 | Bank of America | Up 3 | 7 | 8 |
| 8 | Citigroup | Up 4 | 5 | 3 |
| 9 | Morgan Stanley | Down 4 | 16 | 4 |
| 10 | TD Bank | Up 3 | 9 | 9 |
Source: Evident AI Index for Banks, October 2026
Which banks are leading in AI banking?
JPMorganChase ranks first in the Evident AI Index, followed by Capital One, Royal Bank of Canada, CommBank and Wells Fargo. The index measures AI maturity across 50 of the world’s largest banks, and every bank in the top 10 has placed there before. The top four held their places in 2026, while every bank from rank 5 to 10 moved by one to four places.
AI Talent and Patents at Leading Banks
- In the 2025 Index, the AI headcount at the 50 banks Evident tracks grew more than 25%, the largest increase since the Evident AI Index launched in early 2023.
- In the 2025 Index, about one of every fifteen Capital One employees worked on AI, compared with roughly one in forty at JPMorganChase.
- In the 2025 Index, Capital One had filed more than 1,700 AI patents, the most of the 50 banks tracked, and had grown its AI team by 73% in the last year.
- JPMorganChase rolled out its generative AI platform LLM Suite to more than 200,000 people, 125,000 of whom use it every day.
- In the 2025 Index, seven of the top-10 banks, and 11 of the top 20, were headquartered in North America.
- Talent makes up 45% of a bank’s score in the Evident AI Index.
View the data
Talent carries the heaviest weight of the four pillars.
Bank AI Use Case Announcements by Quarter
- The 50 banks tracked in the Evident AI Index for Banks announced 93 new AI use cases in Q2 2026, up 45% from last quarter.
- Generative AI use cases almost doubled in Q2 2026, spreading across six of the nine application areas, led by Customer Engagement.
- Six banks announced their first agentic use case in Q2 2026.
- Commercial Banking and Wealth Management use cases tripled quarter-on-quarter, with Commercial Banking rising from 8 to 22.
- Microsoft’s share of bank AI use case announcements fell from 37% in early 2024 to 18% by Q2 2026.
- Anthropic was the most referenced vendor in Q1 2026, while specialized players beyond the hyperscalers account for 68% of all deployments.
View the data
AI ROI and Value Disclosures by Banks
- 16 banks on the Evident AI Index now put a number on group-wide AI value, with State Street setting the fourth-largest to date at $1 billion.
- Only 27% of new use cases carried a disclosed outcome in Q2, down from 41% in the previous quarter, with just 1 in 10 of those reporting an uplift in revenue.
- 38% of Q1 2026 use cases reported outcomes, and 15 of the 50 banks Evident tracks disclosed aggregate AI ROI.
- Bank of America spends $14 billion annually on technology.
| Bank | Disclosed AI figure | Basis |
|---|---|---|
| State Street | $1 billion | Group-wide AI value |
| JPMorganChase | Heading more towards $2 billion | AI ROI guidance from its president |
| Santander | €1 billion | 2028 goal, reported quarterly |
| Bank of America | More than $4 billion a year | Technology spend on new initiatives, including AI |
Source: Evident AI Use Case Tracker Q1 and Q2 2026, Evident AI Index 2025, Bank of America
Key finding: Only 27% of new use cases carried a disclosed outcome in Q2, down from 41% in the previous quarter, according to Evident’s use case tracker. Banks are announcing AI projects faster than they report what those projects earn, which makes group-wide value claims harder to check against individual deployments.
The Q1 and Q2 tracker editions state the Q1 outcome share differently, so the quarter-on-quarter comparison is best read within the Q2 edition.
Bank of America Erica Statistics
- Erica has assisted nearly 50 million users since launch at Bank of America, surpassing 3 billion client interactions and averaging more than 58 million interactions per month.
- More than 98% of users find the information they need, and the system has undergone over 75,000 updates since launch.
- Clients interacted 676 million times with Erica in 2024, bringing total interactions since launch to more than 2.5 billion.
- Merrill clients logged a record 11.5 million interactions in 2024, and Private Bank clients interacted with Erica 690,000 times.
- More than 18,000 employees use EricaAssist as a human-assisted AI agent, which reduces average call times by nearly one minute per interaction.
| Erica metric | Value | As of |
|---|---|---|
| Users assisted since launch | Nearly 50 million | August 2025 |
| Client interactions since launch | Surpassing 3 billion | August 2025 |
| Average monthly interactions | More than 58 million | August 2025 |
| Proactive insights delivered | More than 1.7 billion | August 2025 |
| Hours of client conversation | More than 18.7 million | August 2025 |
| Client interactions in the year | 676 million | 2024 |
Source: Bank of America newsroom, February and August 2025
Consumer Attitudes Toward AI in Banking
- 69% of Americans would allow a bank-provided AI assistant to complete at least one financial task on their behalf.
- Nearly half (49%) used an AI assistant in the past year to research or make a personal financial decision, and 83% of those users verified the information with another source or professional.
- Consumers were most willing to allow an AI assistant to alert them to potentially better products or rates (38%).
- The September 2026 TD AI Omnibus was conducted by Big Village from September 18 to 20, 2026, among 1,008 respondents.
View the data
Human oversight was the leading factor in building consumer confidence in a bank’s use of AI, with 36% citing human review of important financial decisions as a key safeguard. The safeguards respondents named are compared below.
View the data
AI Use Cases by Business Area and Materiality
- Operations and IT is the largest area of reported use cases, twice the proportion of retail banking with 11%.
- Foundation models account for 17% of all use cases, and operations and IT holds around 30% of foundation model use cases.
- Of the total number of use cases reported by respondent firms, 62% were rated as low materiality.
- High materiality use cases were most common in general insurance, risk and compliance, and retail banking.
- Anti-money laundering (AML) ranks among the top benefits: The areas with the highest perceived current benefit are data and analytical insights, AML and combating fraud, and cybersecurity.
View the data
How is AI being used in banking?
The most common uses of GPAI and agentic AI in consumer-facing processes at EU banks surveyed by the EBA include detection and notification of fraudulent or suspicious activities and assisting customer service agents and call center operators to handle service requests. In the UK, operations and IT accounts for around 22% of all reported use cases, ahead of retail banking with 11%.
Third-Party AI Providers and Automated Decision-Making
- A third of all AI use cases are third-party implementations.
- The top three third-party providers account for 73% of all reported cloud providers.
- 55% of all AI use cases have some degree of automated decision-making, with 24% of those being semi-autonomous.
- Only 2% of use cases have fully autonomous decision-making.
- 46% of respondent firms reported having only partial understanding of the AI technologies they use, versus 34% of firms with complete understanding.
View the data
Worth noting: Only 2% of use cases have fully autonomous decision-making in the Bank of England and FCA survey of UK financial firms. Most bank AI still passes its output through a person, which keeps accountability with staff even as use case counts climb and model supply concentrates among a few providers.
AI in Fraud Detection and Payment Integrity
- The US Department of the Treasury prevented and recovered over $4 billion in fraud and improper payments in fiscal year 2024, up from $652.7 million in fiscal year 2023.
- Identifying and prioritizing high-risk transactions resulted in $2.5 billion ($2,500 million) in prevention.
- Expediting the identification of Treasury check fraud with machine learning AI resulted in $1 billion ($1,000 million) in recovery.
- 68% of consumers are comfortable with banks using AI to detect potential fraud and send alerts, while 50% expressed trust in AI-powered technology to help protect their money from fraud and cyber threats.
- Detection and notification of fraudulent or suspicious activities is among the most common uses of GPAI and agentic AI in consumer-facing processes at EU banks surveyed by the EBA.
View the data
By the numbers: The US Department of the Treasury prevented and recovered over $4 billion in fraud and improper payments in fiscal year 2024, up from $652.7 million in fiscal year 2023, an increase it credits to its Office of Payment Integrity. Prevention before payment, not recovery after it, accounts for most of that total.
Regulatory Constraints on AI Adoption in Banking
- Data protection and privacy was noted as the greatest regulatory constraint, with 23% identifying it as a large constraint, 29% as a medium constraint and 10% as a small constraint.
- 12% of firms consider resilience and cyber security rules a large constraint.
- The CSBS Artificial Intelligence Supervisory Framework is a discretionary tool for state examiners to identify and understand AI at financial institutions.
- Financial institutions can use the framework to assess their own AI programs, establish sound AI governance and risk management, and prepare for examinations.
- Each state financial regulatory agency will determine the extent to which this framework is incorporated into their supervisory programs.
- At the ECB, supervisory attention has traditionally focused on AI in areas such as credit risk assessment, fraud detection and transaction monitoring.
View the data
Can AI replace human bankers?
Current evidence points to AI assisting bank staff far more often than replacing them. Only 2% of use cases have fully autonomous decision-making in the Bank of England and FCA survey, and more than 18,000 employees use EricaAssist as a human-assisted AI agent at Bank of America.
Headcount is still in play. JPMorganChase forecast that AI could help the bank cut at least 10% from its headcount across a number of divisions, while 83% of consumers who used an AI assistant for a financial decision verified the information with another source or professional.
Conclusion
92% of EU banks are currently deploying AI, and the 50 banks tracked by Evident announced 93 new AI use cases in Q2 2026. Yet only 27% of new use cases carried a disclosed outcome, so adoption is running ahead of the evidence that it pays.
The next markers to watch are how many banks put an audited figure on AI value and how many states fold the CSBS framework into their examinations.






























































