When a Londoner stays up past midnight wondering whether to consolidate their pension or when to open an ISA, they are increasingly not reaching for an IFA’s phone number. They are typing into ChatGPT or Claude. And according to the Financial Conduct Authority, when they do, they are on their own.
The FCA’s Perimeter Report 2026, published on 26 March, drew a stark boundary. General-purpose AI tools offering financial guidance or recommendations sit outside the regulated perimeter. Consumers who rely on them have no access to the Financial Ombudsman Service and are not covered by the Financial Services Compensation Scheme. If the chatbot gets it wrong, there is no redress.
The advice gap, made visible
The scale of the problem gives the FCA’s concern its urgency. Fewer than one in ten UK adults currently receive regulated financial advice. Yet nearly one in five investors now turn to social media for help with financial decisions — a proportion that has risen sharply over the past three years. The rapid adoption of conversational AI has added a new and largely invisible layer to that informal ecosystem.
The FCA stopped short of declaring these tools unlawful. Its position is more nuanced: the line between general information and regulated advice is increasingly blurred when an AI can answer highly personalised questions about specific pension schemes or investment strategies. A tool that tells a user to switch from one particular fund to another may be crossing that line whether or not it carries a financial promotions disclaimer.
Regulatory pressure builds
The Perimeter Report’s findings have not gone unnoticed at Westminster. The Treasury Select Committee has recommended that the FCA publish comprehensive guidance on AI in financial services before the end of 2026 — giving the regulator fewer than nine months to produce what amounts to a new framework for a technology that did not exist in its current form when the existing financial promotion rules were written.
Meanwhile, the Sheldon Mills Review is examining how AI could reshape retail financial services by 2030, with particular focus on whether new forms of AI-assisted guidance could help close the advice gap without requiring full transition to regulated advice models. The FCA is also working closely with the Information Commissioner’s Office and the Competition and Markets Authority through the Digital Regulation Cooperation Forum. The CMA published its own warning in March, flagging that algorithmic tools can facilitate anti-competitive pricing collusion — a concern that extends directly into financial markets.
What the City is watching
For the City of London’s financial services sector, the regulatory direction of travel is clear even if the destination remains uncertain. Firms supplying AI tools that touch personal finance — from embedded robo-advisers to automated investment platforms — are on notice that existing rules apply regardless of the underlying technology. The FCA’s message is unambiguous: delegating a decision to software does not delegate legal accountability.
What remains unresolved is whether the FCA will draw explicit new perimeter lines to capture general-purpose AI tools, or rely on existing financial promotion and advice rules to pursue cases where consumer harm can be demonstrated. The Perimeter Report hints at the former; a timeline for consultation has not been set.
For the estimated nine in ten UK adults without access to regulated financial advice, the question is immediate. The tools they are using to make consequential decisions about their money are operating in a regulatory grey zone. The FCA has noticed. Whether it moves quickly enough to matter is another question entirely.
Discussion
Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.