Among financial advisers, when you take a general consensus, most are already using AI, whether they think of it that way or not. A joint survey from the Bank of England and FCA in late 2024 found that 75% of UK financial services firms have adopted AI in some form or another. This could show up in tools that summarise meeting notes, draft client communications, or support research workflows. The technology has become ordinary enough that it often goes unnoticed.
The regulatory response has been measured. The FCA has confirmed it will not introduce AI-specific rules, choosing instead to apply existing frameworks to AI-assisted processes. Consumer Duty still applies to outcomes delivered through AI tools, and SM&CR accountability remains with the individual, regardless of whether technology played a role in a decision. Delegating to an algorithm does not dilute liability.
This means there is no new compliance burden to prepare for, but there is also no specific guidance yet on what good AI documentation looks like.
What’s happening in 2026
The FCA launched its AI Live Testing programme in October 2025, with the first cohort of firms now working through a roughly six-month engagement. A second cohort is expected in early 2026, and the FCA will publish an evaluation report in spring 2026.
The programme is not a certification or audit process. The FCA is using it to learn alongside participating firms, exploring questions that do not yet have settled answers. These include how to assess whether an AI model is robust, how to measure bias in outputs, what to do when an AI system produces results that cannot be easily explained, and how vulnerable consumers might be affected by AI-driven processes. The findings will be published so that all firms can benefit from what emerges, not just the participants.
The regulator is also still working out what good practice looks like. The spring 2026 evaluation is likely to inform expectations going forward, but until then, advisers are operating in a space where the rules are clear and the specifics are not.
What advisers can do now regarding AI use
Preparation is less about ticking new boxes and more about understanding what processes you are already doing and how you are keeping record. It is best practice to document which tools in your workflow involve AI, and what role they play in client-facing outputs. This will put you in a stronger position than needing to trace your steps to uncover it later. The existing regulatory framework already applies to those tools, so the question is whether you could explain your process if asked.
Watching what comes out of the AI Live Testing evaluation will also matter. When the FCA publishes its findings, that document is likely to signal where expectations are heading, even if formal guidance takes longer to arrive.
Automwrite helps advisers prepare for the future of AI letter writing and adviser workflows
Automwrite is built for financial adviser by financial advisers. The platform centralises client data, meeting recordings, transcripts, and documents in a single workspace, then generates CIP-aligned suitability reports from that complete context. Advisers review and approve every output before it goes anywhere.
The infrastructure supports accountability from the ground up. Automwrite is UK-hosted on AWS in London, registered with the ICO, and applies data minimisation to AI processing. Authentication events and privileged actions are logged, and role-based access controls are enforced across the platform. Data Protection Impact Assessments are conducted for AI-assisted processing as part of an ongoing compliance programme.
As regulatory expectations take shape, Automwrite is designed to keep pace.
