Content
The UK’s Financial Conduct Authority has planted a flag. On July 6, the regulator published The Mills Review, a study it bills as the first of its kind commissioned by any financial regulator anywhere, laying out how artificial intelligence could rewire retail financial services by 2030.
Led by FCA executive director Sheldon Mills and commissioned by the Board, the report reads less like a cautious regulatory memo and more like a strategy document for an industry the FCA expects to look very different in a few years.
And the word doing the heavy lifting throughout is “agentic.”
What The Mills Review Actually Is
The FCA launched the review back in January 2026 to figure out what advanced AI means for consumers, retail markets, and regulators. Mills ran it, drawing on input from across the industry.
The core finding is not subtle. AI is set to become a defining force in retail finance, changing how firms run, how consumers make money decisions, and how markets function. The upside the report flags is better access, personalisation, and efficiency. The downside is a sharper edge on fraud, cyber risk, consumer harm, and market concentration.
In Mills’s own framing, artificial intelligence will transform financial services by 2030. The report is pitched as a roadmap for industry, regulators, and government to get ready for it.
Four Shifts The FCA Sees Coming
The review groups the disruption into four AI-driven shifts likely to hit retail finance:
- Firm operations get transformed as AI moves deeper into back-office and decision-making functions.
- Consumer journeys evolve, with AI reshaping how people find, choose, and manage financial products.
- Competition and market power get reshaped, raising the risk that a few dominant AI providers concentrate influence.
- Fraud and cyber risks get amplified, as the same tools that help firms also arm bad actors.
That last point is the one worth watching. Regulators rarely publish a “here’s the opportunity” report without flagging exactly where it can blow up.
Consumers Already Want Autonomous AI
The demand signal is already there. Research commissioned for the review, run by Yonder Consulting in April 2026 across more than 5,000 UK retail consumers, found that 20% of people would likely use AI capable of acting autonomously within pre-set goals.
That fifth of the population works out to roughly 11 million UK adults ready to hand a bot the keys to act on their behalf. The survey was weighted to be representative across age, gender, ethnicity, region, housing tenure, and internet ability.
But willingness came with a caveat. Respondents flagged real concerns about trust and control once AI starts acting on its own. Appetite is not the same as blind faith.
Seven Recommendations For The Regulator
The Mills Review hands the FCA Board and Executive seven recommendations to weigh:
- Secure and adapt the regulatory perimeter.
- Strengthen system-wide coordination and oversight.
- Monitor the shift to autonomous models and adapt frameworks accordingly.
- Scale up the FCA’s AI Lab to support model and system innovation.
- Enable the foundations for agentic finance.
- Build and adopt an AI-enabled agentic supervisory model.
- Develop a trusted public-interest AI-enabled financial capability service.
Read them together and a theme emerges: the FCA wants to regulate agentic AI while using agentic AI to do the regulating. Recommendation six effectively proposes an AI supervisor watching an AI-driven market.
The Agentic Finance Angle
For anyone in digital assets, recommendations one and five are the ones to circle. “Securing the regulatory perimeter” is regulator-speak for deciding which new activities fall inside the rules, and autonomous agents transacting on a user’s behalf sit in exactly that grey zone.
“Enabling the foundations for agentic finance” is a striking phrase from a regulator. It signals the FCA doesn’t plan to fight autonomous financial agents so much as build guardrails around them. That framing matters for the crypto and DeFi crowd, where agent-driven wallets and onchain automation are already live experiments rather than 2030 projections.
The FCA’s approach here leans on tools it already has, the Consumer Duty and the Senior Managers Regime, rather than a bespoke AI rulebook. Chair Ashley Alder framed the principles-based, outcomes-focused method as the reason the regulator has kept pace so far. Whether that flexible stance survives contact with fully autonomous agents is the open question.
What Happens Next
The FCA has been building toward this for a while. Its AI Lab, AI Live Testing, and the NVIDIA-backed Supercharged Sandbox already let firms trial AI systems under the regulator’s eye.
Later this year, the FCA plans to publish an AI good and poor practice guide, drawn from direct engagement with firms on what’s working and where they want clarity. So the Mills Review isn’t the finish line. It’s the FCA setting the terms for the next round, and betting that agentic AI is coming to finance whether the rulebook is ready or not.