The guidance was announced on 19 October 2026 in a speech by an FCA official at the Bloomberg UK Policy Series. It draws on the regulator’s observations from firm testing during 2026 and covers three areas: consumer centricity, accountability and governance, and operational resilience. The FCA has said the material is not static guidance and will be updated as the technology evolves.
The publication comes as AI use in the sector moves from experimentation to deployment. According to a joint survey by the FCA and the Bank of England, due to be published the following week, 86% of firms now use AI.
Good and poor practice in three areas
On consumer outcomes, the FCA bases its expectations on the Consumer Duty, its outcomes-based regulatory framework. In one example of good practice, a firm kept a live list of vulnerability signals that its chatbot uses to flag consumers who may need additional support from human advisers. Poor practice included developing AI tools without a clearly defined need, excluding vulnerable consumers by default, overstating capabilities, and deploying automated customer journeys without sufficient testing.
The regulator identified governance as an area where many firms are still under-investing. It expects clear accountability, risk-based controls, human oversight, escalation triggers, and senior managers who are personally accountable for AI systems. It cited one firm that set up an independent ethics panel to review proposed use cases before deployment. Other firms, by contrast, added governance only after deployment or did not adapt monitoring as their systems changed.
On resilience, firms are expected to stress-test AI systems, maintain fallback options, and understand supplier concentration, including reliance on a single model.
Response to the Mills Review
The guidance forms part of the FCA’s response to the Mills Review. The review, published in July 2026, examined how AI could reshape retail financial services by 2030. The FCA is embedding six of its seven recommendations into live work, with a focus on testing, assurance, monitoring, and surveillance rather than further consultation. Its AI Lab gives the regulator insight into what firms are building, from agent-led payments to fraud detection and consumer support.
The regulator also pointed to the growing use of general-purpose large language models for financial information. Four in five less-experienced investors aged 18 to 40 have used AI to help with investing. However, only 40% of consumers recognise that there is no recourse after acting on advice from such tools. The FCA is calling on AI providers to introduce clear, proportionate, and consistent disclaimers. The review had suggested a new public-interest financial capability tool, but the FCA will instead build on existing work, including IOSCO’s first TechSprint.
Agentic finance and supervision
The FCA noted that AI agents can already search across products, manage payments and investments, and transact with other agents. It said firms’ incident reporting must extend to significant AI failures and near-misses, and that agentic systems require continuous, conditional assurance. The regulator plans to work with industry, other regulators, and frontier AI labs on a framework for agentic finance. That framework will cover agent identity, delegated authority, accountability, and execution.
The FCA is also applying AI to its own supervisory work. It is building tools to oversee 60.000 anti-money laundering firms due to enter its remit in 2028. It is also developing a market first responder agent to supervise around 9.000 small wholesale markets transaction reporting firms. In August 2026 it launched its Handbook API, which gives firms machine-readable access to its rules, and it is now working on machine-readable data collection APIs.
