Apolifina

Financial regulation

FCA sets out AI good practice for financial firms

The FCA has published examples of good AI practice and areas for improvement in financial services, covering consumer outcomes, governance and resilience.

· · 1-minute read

Catch-up coverage · Source released

Verified facts

The examples include meaningful human oversight and escalation, rather than assuming a person's presence makes an automated process safe.

In a published speech, Jessica Rusu also called on AI providers to explain the limits of financial information and advice produced by their models. The FCA notes that the drafted speech may differ from the delivered version.

Apolifina analysis

A fluent answer and a regulated financial service are not interchangeable. Before acting, an investor needs to know who provides the service and which protections apply, rather than relying on an interface's confidence.

For shareholders in financial firms, AI promises should be assessed alongside testing, responsibility for decisions and service continuity. The publication does not quantify any company's cost savings or prove that a particular AI investment will improve profits.

Uncertainty and risk

This is supervisory practice and a policy speech, not a newly announced universal compensation right or a guarantee of reliable AI advice. No unpublished survey statistic is used. Financial decisions and investments involve risk; this is general reporting, not personal advice.

Primary sources

Sources rechecked on 20 October 2026.

Continue reading: CMA software proposal; editorial policy