UK financial AI adoption reached 86% of respondents in a Bank of England and Financial Conduct Authority survey published on October 26, 2026, compared with 75% in 2024. The findings show wider deployment of generative AI and agents, but limited use of broadly autonomous systems and growing reliance on outside technology providers.
Key Takeaways
- AI use increased among participating firms, although changes in the survey population limit comparisons with 2024.
- AI agents were widely deployed, but only 2% of full-survey respondents reported agents with broad autonomy.
- Third-party systems accounted for 53% of high-materiality AI systems, increasing the importance of vendor oversight.
Table of Contents
- UK financial AI adoption expands, with comparison limits
- Widespread agents do not mean widespread autonomy
- Outside suppliers take a larger role in material systems
- Governance is stronger before deployment than afterward
- Conclusion
UK financial AI adoption expands, with comparison limits
The fourth joint survey received 196 responses, including 101 larger firms and 95 smaller firms. Firms with fewer than 100 employees could choose a shorter questionnaire; 105 respondents completed the full version. Unless otherwise stated, the detailed findings use valid responses to the full survey.
The headline adoption figure covers both questionnaires. The authorities cautioned that changes in respondent composition and question scope mean the 2026 and 2024 percentages do not describe the same population. The results are evidence from participating firms, not a complete measure of the UK financial sector.
Generative AI represented 56% of reported systems. Foundation models represented 17% of reported use cases in 2024, but the authorities said those measures are not directly comparable. Operations and IT remained the largest business area, accounting for about 31% of systems, against roughly 22% in 2024.
| Item | Details |
|---|---|
| Firmwide AI policies, principles or standards | Reported by 94% of full-survey respondents answering the question. |
| Centralised AI or model-risk governance forums | Reported by 84% of respondents. |
| Applying model-risk principles to vendor models | Selected by 54% as an area of governance uncertainty or implementation challenge. |
| Frontier AI cyber capabilities | 88% expected a moderate or high impact on their firm's risk profile over the next two years. |
Widespread agents do not mean widespread autonomy
Among full-survey respondents, 80% reported deployed AI agents and a further 14% planned deployment within the next 12 months. Decision support dominated: 80% reported agents that recommend actions while humans make and execute decisions. Only 2% reported agents with broad autonomy.
Customer-facing deployment was more limited. Automated customer communications were reported by 16% of respondents, rising to 34% when plans were included. Even including planned deployment, no more than 9% expected agents to perform functions such as approving customer outcomes, executing trades, initiating or blocking payments, or changing limits and controls. Adoption therefore remained much broader than delegation of consequential financial actions.
Outside suppliers take a larger role in material systems
Third-party implementations accounted for 39% of reported AI systems, up from 33% in 2024. Their share of high-materiality systems rose more sharply, to 53% from 35%.
Firms classified about 7% of systems as high materiality, compared with 16% in 2024. That lower share did not mean fewer important systems: the absolute number classified as high materiality increased.
Provider relationships were concentrated. The three largest cloud and compute providers accounted for 79% of reported relationships, while the top three model API and hosting providers accounted for 54%. Information-security and operational-resilience concerns were the most prominent major constraints on third-party adoption, identified by 40% of respondents.
Governance is stronger before deployment than afterward
Human approval, pre-deployment review and vendor due diligence were each reported by more than 90% of respondents. Ongoing monitoring, leakage testing and periodic review or recertification were each reported by only around half. Formal approval processes were therefore more widespread than continuing checks after systems entered use.
Understanding also differed by origin. For high-materiality systems, 77% reported relying on deep knowledge of internally developed systems, compared with 24% for third-party systems. For external systems, 81% instead relied on vendor assurance and contractual arrangements.
The authorities’ 2022 discussion of how existing financial rules apply to AI examined requirements across data, models and governance. That policy framework provides context for the latest survey’s vendor-oversight challenges; the 2026 findings describe firms’ practices and perceptions rather than announce new rules.
Respondents expected AI risks to increase across all categories over the next one to three years. Cyber-risk amplification and critical third-party dependencies were their highest-rated systemic concerns, both currently and over the next three years.
Conclusion
The next stated deployment window is the 12 months following the survey, when a further 14% of full-survey respondents planned to introduce AI agents. Whether those plans move into production—and how firms monitor externally supplied systems—will be important measures of the next stage of UK financial AI adoption.
Read More News
Disclaimer
This FinanceVelix article provides general information, not personalized financial, investment, legal, or tax advice. Market prices, interest rates and economic conditions can change; investments carry a risk of loss. Facts reflect the material available when published and may change. Check important information independently and consult a qualified professional when needed. Mentioning an asset, company, or service is not an endorsement.
