AI financial advice: FCA calls for disclaimers on agentic AI

What happened

The UK's Financial Conduct Authority (FCA) wants AI providers to attach disclaimers to services that give consumers financial advice through agentic AI, meaning systems that can carry out tasks and steer decisions with a degree of autonomy rather than only answering questions. The position was reported by Finextra.

The request is aimed at the providers building these services, not at the consumers using them. It places the responsibility for disclosure at the moment a customer receives guidance. The source material is brief and does not set out the wording, timing or enforcement route for any disclaimer, so those details should be checked against the FCA's own publications before anyone relies on them.

Why it matters

Agentic AI changes the role of a digital channel. A chatbot that answers a question is one thing. A system that recommends a product, moves money or sequences a set of decisions is another, and the customer may not register that a rule-driven or model-driven process is now shaping the outcome. A disclaimer is one way to make that shift visible, which makes this call relevant to any organisation that is deploying AI in services, not only to regulated firms.

The behavioural dimension deserves attention. People tend to place trust in fluent, confident and personalised output, and that tendency can outweigh a warning that appears elsewhere in the journey. Leaders in experience and digital transformation will need to think about where disclosure sits in the interaction, how it is worded, and whether the customer can act on it. Governance teams and experience designers will have to work from the same brief.

The Renascence take

Most of the debate will focus on the wording of the disclaimer. The more useful question is whether a customer would notice it, understand it and change what they do as a result.

A disclaimer that sits in a footer is compliance theatre, not customer protection. Operators should state the limits of the advice at the point of decision, show the customer what the system has and has not considered, and make it easy to reach a human before committing. Treat the disclaimer as a design brief, test it with real customers, and measure whether they understand it. If it does not change behaviour, it has not done its job.