This Week in Brief

The conversation around AI in financial advice has moved from the general to the specific this week. The FCA has published its first batch of sector-specific Regulatory Priorities reports, explicitly naming AI governance as a priority for firms across insurance, wholesale markets, and consumer investments. At the same time, the regulator’s second cohort of live AI testing is getting under way this month, the FCA has published its open finance vision — with a direct warning for advisers who aren’t paying attention — and fresh research puts some concrete numbers on what AI adoption could look like in practice. The Bank of England, meanwhile, has confirmed it will begin stress-testing AI agent behaviour in the financial system.

Key Developments

FCA Formalises AI Governance as a Named Priority

The FCA has this week launched a new format for communicating regulatory expectations, replacing more than 40 individual portfolio letters with a series of sector-specific Regulatory Priorities reports. Two themes run consistently through every report published so far: the responsible adoption of AI, and the management of off-channel communications. For IFAs, the significance is that AI governance is no longer a background concern the regulator is watching from a distance — it is explicitly on the list of things firms are expected to get right in 2026/27. The message across every sector report is the same: innovation is welcome, but it does not reduce firms’ accountability for the outcomes it produces. Consumer Duty still applies, and senior managers are still personally on the hook if AI causes harm.

FCA’s Second AI Live Testing Cohort Begins This Month

The application window for the FCA’s second cohort of AI Live Testing closed on 24 March, and testing is set to begin in late April. The programme — the first of its kind in UK financial services — gives selected firms hands-on regulatory support and technical guidance as they develop and deploy AI tools in a live environment. Use cases being tested include financial guidance, personalised recommendations, and complaint handling, all of which map directly onto the advice process. The FCA has also confirmed it will publish an evaluation report from the programme by the end of 2026, which will give the wider industry its first detailed, regulator-backed picture of how AI is actually performing in practice. That report will be worth reading carefully when it arrives.

The 70% Figure: What AI Could Cover in Your Practice by 2030

Research published this week by WealthTech platform Fincite puts a headline number on AI’s potential in wealth management: 70% of advisory tasks could be AI-supported by 2030. That is a striking figure, but the more immediately useful finding is about where the friction lies today. Fincite’s analysis found that 40% of every advisory meeting is currently spent on data capture rather than actual client conversation — and that many firms are held back not by a lack of ambition but by fragmented systems and siloed data that make meaningful AI integration difficult. The company’s platform addresses four specific problems: real-time meeting transcription, an AI compliance and portfolio dashboard, guided MiFID II profiling, and automated personalised client reports. None of these are theoretical — they are live tools in use today. The 70% figure is a long-term projection, but the four use cases it is built around are worth understanding now.

FCA Opens the Door to Open Finance — and Issues a Warning to Advisers

The FCA published its open finance vision document on 14 April, setting out a roadmap to extend data sharing beyond open banking to a much wider range of financial products and services. The document explicitly names AI as one of the main beneficiaries: “automated fraud detection, back-office operations and personalised customer service” are all cited as areas where open data and AI together create new possibilities. The roadmap runs to 2030, with Treasury framework options expected by end 2027. Ian McKenna of the Financial Technology Research Centre was quoted in trade press with a pointed warning for advisers: firms that are not paying attention risk being “caught on the back foot” when other parties — banks, platforms, fintech providers — start showing up with better client data than the adviser has. The FCA’s vision also explicitly references “tailored financial advice” as something open finance should support. Whether that translates into a narrowing or an expansion of the regulated advice perimeter is a question the industry will be debating for some time.

Update: Parliament, the FCA and the Bank of England on AI Accountability

Building on last week’s story about the FCA holding its principles-based approach to AI, this week brings concrete developments from both the FCA and the Bank of England. The FCA has confirmed it will share best-practice examples of responsible AI use with firms, and the Bank of England has announced plans to stress-test the behaviour of AI agents operating in financial markets — specifically looking at the risk of “herding,” where multiple AI systems make the same decisions at the same time and amplify market stress. At the Parliamentary level, Treasury Select Committee chair Dame Meg Hillier remained publicly critical, stating she was “perplexed at the apparent inertia shown by the Treasury” on its failure to designate major AI and cloud providers as critical third parties. The Committee’s specific demands remain outstanding: practical guidance on how Consumer Duty applies to AI, clarity on SM&CR accountability when AI causes harm, and that critical-third-party designation. The FCA’s deadline for its practical guidance is end 2026 — less than nine months away.

The Human Advantage: Why Trust Is AI-Proof

An FT Adviser analysis published on 16 April by Crawford Hollingworth makes a case that is worth reading carefully before writing off the human adviser. The argument is not that AI is inaccurate or unreliable — it is that the things clients value most in an advisory relationship cannot, by their nature, be replicated by a machine. What AI cannot deliver, the analysis argues, is “a sense of personal accountability, shared risk and human presence.” In a relationship built on trust — where clients are making decisions about their financial futures and their families’ security — those qualities are not peripheral. They are the product. The practical implication for advisers is that the right response to AI is not to compete with it on efficiency, but to lean into the things that make the human relationship irreplaceable. Firms that do this explicitly — and can articulate it to clients — are likely to find AI makes them more valuable, not less.

From the Trade Press

  • “FCA and BoE commit to action on AI” — The Bank of England confirms AI agent stress-testing plans and will investigate “herding” risk; the FCA commits to sharing best-practice examples with firms; HM Treasury faces renewed criticism for failing to designate AI and cloud providers as critical third parties. FT Adviser, 17 April 2026

  • “Human capital is premium in AI repricing of advice” — Crawford Hollingworth argues that trust, accountability, and human presence are the adviser’s irreplaceable advantages — and that firms that lean into these qualities will benefit as AI reshapes the market. FT Adviser, 16 April 2026

  • “FCA unveils open finance vision including ‘tailored financial advice’ help” — The FCA’s open finance roadmap extends data sharing beyond open banking and explicitly names AI as a key enabler. Industry commentators warn advisers to act before other parties gain the data advantage. Professional Adviser, 14 April 2026

  • “FCA targets AI governance and off-channel messaging” — The FCA’s new sector-specific Regulatory Priorities reports put AI governance and off-channel communications at the centre of firms’ 2026/27 compliance agenda. Boards and senior management are the intended audience. FinTech Global, 16 April 2026

  • “How AI is reshaping wealth management advisory” — A detailed look at how WealthTech platform Fincite is embedding AI into the advisory workflow, covering the four use cases firms are deploying today and the infrastructure challenges that remain the biggest obstacle to wider adoption. FinTech Global, 14 April 2026

  • “UK Financial Services Regulators’ Approach to AI in 2026” — A useful summary of where the FCA, PRA, and Bank of England currently stand on AI regulation, covering the principles-based approach, the Mills Review timeline, and what firms can expect from guidance later this year. Inside Global Tech, 9 April 2026

What to Watch

The FCA’s Mills Review — a long-term look at how AI could reshape retail financial services — is due to publish its final policy recommendations in summer 2026. That is now only a few months away. For advice firms, the review’s scope includes how Consumer Duty applies to AI-generated advice, what accountability looks like for senior managers when AI causes harm, and whether the current regulatory perimeter is adequate for AI tools operating at scale. Whatever the review concludes is likely to define the shape of AI governance in financial advice for years to come — and firms that have already started thinking about their AI governance frameworks will be in a much stronger position to respond.

Sources: FT Adviser, Professional Adviser, FinTech Global, Inside Global Tech, FCA.org.uk, Parliament.uk, web search. Compiled 17 April 2026.