FCA
Financial Conduct Authority (UK)
FCA publishes consumer research on AI use and trust in UK retail financial services
Published
Jul 3, 2026
Topics
Artificial intelligence, Retail financial services, Consumer Duty, Consumer understanding, Consumer protection, Data and privacy, Product governance, Operational risk, Financial advice and guidance, Vulnerable customers
Executive Summary
The FCA has published FCA-commissioned consumer research, conducted by Yonder in April 2026 among 5,026 UK retail financial services consumers, to inform the Mills Review on the long-term impact of AI on retail financial services. The research finds that AI is already being used in personal finance: 16% of consumers use AI to support at least one personal finance activity, and 17% of consumers who shopped for or used a financial product in the last 12 months used AI to assist them. Current use is mainly assistive, such as summarising, explaining, simplifying and comparing information, rather than fully delegating decisions. However, some consumers are willing to give AI deeper access to financial data: 13% of all consumers and 18% of current AI users would be willing to give AI real-time access to banking and financial data. The research highlights material consumer-risk themes relevant to AI product governance and Consumer Duty controls: consumers are most concerned about misuse of personal or financial data, lack of protection if something goes wrong, scams/fraud, mistakes that cost money, loss of control and opaque decision-making. The publication does not amend FCA rules, templates or reporting requirements, but firms developing or deploying AI-enabled retail propositions should treat it as an important reference point for AI risk assessments, customer testing, disclosures, human escalation design, complaints/recourse messaging, data-access permissions and monitoring of outcomes across customer groups.
What Changed
New
The survey covers 5,026 UK retail financial services consumers and reports evidence on AI usage, trust, perceived protections, data-sharing attitudes, and consumer responses to assistive, agentic and autonomous AI financial-services concepts.
New
16% of consumers use AI to assist with at least one personal finance activity; among those using AI for personal finance, 72% use it to summarise, explain or simplify information, 61% use it to suggest what to do based on information provided, 24% upload personal data or documents, and 14% give AI ongoing access to personal data, accounts or software.
New
Of consumers who shopped around for, took out or switched a financial product in the previous 12 months, 17% used AI to assist or carry out related tasks. Among these users, 72% used AI to summarise, explain or simplify information, 58% used AI for suggestions, 20% uploaded personal data or documents, and 9% gave AI ongoing access to personal data, accounts or software.
New
Around a quarter of consumers view AI tools such as ChatGPT as a reliable source of financial information or advice, while only 40% correctly recognise that there is no formal route for recourse if something goes wrong after relying on investment advice from a general-purpose AI tool.
New
Willingness to use AI declines as autonomy increases: 36% of consumers said they would definitely or probably use AI that reviews documents/data/accounts and makes recommendations, 30% would use AI that can take action with permission each time, and 20% would use AI that takes autonomous actions based on pre-set instructions or goals.
New
Firms should treat the publication as supervisory intelligence and implementation context for AI governance, Consumer Duty outcome monitoring, customer communications, product testing and operational controls, not as a binding rule instrument.
Business Impact
Who is affected
UK retail financial services firms and their third-party technology providers involved in AI-enabled consumer journeys, including banks, insurers, consumer credit firms, investment platforms, pensions and retirement providers, mortgage and motor finance firms, debt-advice providers, robo-advice/guidance providers, open-banking or personal financial management providers, and firms using generative AI in customer support, product comparison, eligibility, onboarding, switching, renewal, advice, guidance or money-management tools.
Jurisdictions
United Kingdom
Business processes
AI use-case inventory and approval, Product governance and customer-outcomes testing, Consumer Duty monitoring, Customer communications and financial promotions review, Advice/guidance boundary controls, Data-access consent and permissioning design, Third-party AI/vendor due diligence, Model risk management and validation, Operational resilience and incident management, Complaints, redress and customer escalation handling, Vulnerable customer impact assessment, Fraud, scams and cyber-risk controls
Estimated effort
Medium
Compliance risk
Medium
Affected Reports
| Field | Validation rule |
|---|
Recommended Actions
- 1
Update the firm’s AI use-case inventory to flag retail-consumer applications that match the FCA-tested concepts: moving money between accounts, switching financial products, managing insurance, borrowing and credit, and saving/investing. Prioritise review of use cases that move from information support to action with permission or autonomous action.
- 2
For each AI-enabled retail journey, document the customer control model: whether the tool only explains/summarises information, makes recommendations, takes action with case-by-case permission, or acts autonomously within pre-set instructions. The FCA research shows consumer willingness decreases as autonomy increases, so governance thresholds should increase accordingly.
- 3
Review customer disclosures and in-journey prompts to make clear when a customer is interacting with AI, what data the tool uses, what it can and cannot do, whether outputs are guidance or regulated advice, and what complaints or redress routes apply. The FCA research identifies confusion around recourse where consumers rely on general-purpose AI tools.
- 4
Strengthen data-permission controls for AI tools that request bank statements, personal financial documents, open-banking access or ongoing account access. The research shows some consumers already upload documents or allow ongoing access, while data misuse is the highest-ranked concern.
- 5
Build or refresh human escalation and opt-out routes for AI-supported journeys, especially where AI affects borrowing, insurance, investments, pensions, debt advice, switching or money movement. The research identifies human oversight, opt-out and clear protection as important consumer adoption triggers.
- 6
In Consumer Duty assessments, test AI customer journeys for consumer understanding, foreseeable harm, vulnerable customer impacts, data-use transparency, ability to challenge outcomes, and whether the tool may lead consumers to over-rely on outputs that are inaccurate, incomplete or not suitable for their circumstances.
- 7
For autonomous or agentic AI propositions, require enhanced pre-launch controls: scenario testing, fail-safe limits, transaction caps, fraud/scam risk testing, customer notification design, audit trails, reversibility analysis, incident playbooks and post-launch monitoring.
- 8
Review third-party and general-purpose AI dependencies. Where a firm integrates third-party models or tools, confirm accountability, data handling, explainability, monitoring, model-change notification, audit rights, security, resilience and customer redress arrangements.
- 9
Monitor FCA outputs connected to the Mills Review and the FCA’s broader AI work programme, because this research is explicitly described as part of the evidence base for the review and may inform future supervisory or policy expectations.
Timeline
other
Apr 1, 2026
Yonder conducted online fieldwork for the FCA-commissioned AI consumer research among 5,026 UK retail financial services consumers.
publication
Invalid Date
FCA publication of the AI consumer research PDF. No effective date or implementation deadline is specified in the supplied publication.
other
Invalid Date
FCA Mills Review on the long-term impact of AI on retail financial services. The research states that it was commissioned as part of the Mills Review and is intended to strengthen the review’s evidence base.
Sources
AI-generated analysis is based on the following primary sources. Always verify against the official publication.
- External research PDFFinancial Conduct Authority / YonderInvalid DateYonder / FCA AI Consumer Research: Consumer quantitative research ↗
https://www.fca.org.uk/publication/external-research/ai-consumer-research.pdf
- Press releaseFinancial Conduct AuthorityInvalid DateFCA publishes landmark review into impact of AI on retail financial services ↗
https://www.fca.org.uk/news/press-releases/fca-publishes-landmark-review-impact-ai-retail-financial-services
- Finalised guidanceFinancial Conduct AuthorityJul 1, 2022FG22/5: Final non-Handbook Guidance for firms on the Consumer Duty ↗
https://www.fca.org.uk/publication/finalised-guidance/fg22-5.pdf
- Policy statementFinancial Conduct AuthorityJul 1, 2022PS22/9: A new Consumer Duty ↗
https://www.fca.org.uk/publication/policy/ps22-9.pdf
- Discussion paperFinancial Conduct Authority / Bank of England / Prudential Regulation AuthorityOct 1, 2022DP5/22: Artificial Intelligence and Machine Learning ↗
https://www.fca.org.uk/publication/discussion/dp5-22.pdf
- Feedback statementFinancial Conduct Authority / Bank of England / Prudential Regulation AuthorityOct 1, 2023FS2/23: Artificial Intelligence and Machine Learning ↗
https://www.fca.org.uk/publication/feedback/fs2-23.pdf
Related Evidence
Similar official source documents found with semantic search.
Receive updates like this by email
Get AI-generated analysis for the regulators and topics you care about.