The FCA's Mills Review Sets Out Seven Recommendations for an AI-Enabled 2030, After Finding One in Five UK Adults Open to AI Acting on Their Behalf

The UK's Financial Conduct Authority has published a review that tries to describe what retail financial services will look like by 2030 once artificial intelligence is built into almost every part of them. The Mills Review, titled "AI and the future of retail financial services", was led by FCA executive director Sheldon Mills and commissioned by the regulator's Board. The FCA published it on July 6, 2026, although the page metadata gives July 3, so readers should check the date on the FCA site. The regulator describes it as the first work of its kind initiated by a regulator anywhere in the world. This article summarises the FCA's press release and a legal analysis from the law firm CMS published on July 10, and it does not replace the report itself.

How the review was done

The FCA launched the review in January 2026 with an engagement paper that closed for responses on February 24. According to CMS, it drew on more than 140 written submissions and on discussions with firms, trade bodies, consumer groups, technology providers and international stakeholders. In April 2026, Yonder Consulting surveyed more than 5,000 UK retail financial services consumers, with quotas set to make the sample representative. The review builds on earlier FCA work on AI, including its 2023 discussion paper, the AI Sprint, the AI Lab, the AI Live Testing programme, whose second cohort has been announced, and the Supercharged Sandbox, which is supported by NVIDIA.

The central finding

The review's main conclusion, as described by CMS, is that financial services are shifting from human-led, episodic activity toward services that are AI-enabled, continuous and delegated. It uses a five-level spectrum of autonomy, running from humans operating AI tools at one end to humans acting only as observers at the other. The review says few services will become fully autonomous, but that most leading firms are expected to embed AI across almost every function by 2030. It also concludes that the existing regulatory framework is broadly sound, though it may come under strain as AI becomes more agentic, meaning able to take actions towards goals with limited human involvement.

Four shifts

The first shift is the transformation of firms. Staff move from being operators of AI to approvers and observers. The review says that being "in the loop" is not enough, and firms must define what oversight involves, what reviewers see and how challenge is recorded. AI governance and model risk management become more important, especially where firms rely on third-party providers.

The second is the evolution of consumer journeys. The FCA's research found that about one in five consumers, around 11 million UK adults, would be likely to use AI that can act autonomously within pre-set goals. CMS reports that demand is strongest for debt advice, pensions and investments, and that about a quarter already trust general-purpose tools for financial guidance. The review sees AI as a way to close gaps: about 9% of people receive regulated advice, about 30% hold life or income protection, and about £300 billion sits in low-interest accounts. It also identifies risks, including bias through hyper-personalisation, opaque pricing and personalised deceptive design. Consumers surveyed raised concerns about data misuse (68%), lack of protection if things go wrong (67%) and power concentrating in large firms (65%).

The third is a reshaping of competition and market power. The review points to control of AI-mediated interfaces, such as agents that search, compare and buy products, to barriers to entry that may fall for digital-native firms while scale favours incumbents, and to concentration in the AI supply chain, which raises concerns about cost, lock-in and resilience.

The fourth is the amplification of fraud and cyber risk. AI will make fraud faster, cheaper, more scalable and more persuasive, with cloned voices, synthetic identities and deepfakes. AI can also strengthen detection, and firms will have to show that their controls improve detection and disruption while keeping meaningful human oversight.

The seven recommendations

The recommendations are addressed to the FCA Board and Executive. They are to secure and adapt the regulatory perimeter; strengthen system-wide coordination and oversight; monitor the transition to autonomous models and adapt regulatory frameworks; scale up the FCA's AI Lab; enable the foundations for agentic finance; build and adopt an AI-enabled agentic supervisory model; and develop a trusted public-interest AI-enabled financial capability service.

CMS adds some detail. On the perimeter, it says that in the short term the FCA should review the scale and impact of general-purpose large language models that sit outside the perimeter but influence consumer financial decisions, and in the long term should seek powers beyond the critical third party regime. On monitoring autonomy, the review is said to flag an open question about the Senior Managers "reasonable steps" obligation without answering it. On foundations, the review suggests a trusted framework for AI agent identity, authority, accountability and execution, ideally through standards developed via Open Finance. On the public-interest service, it envisages a free, inclusively designed AI service giving consumers reliable financial information and guidance.

What it means for firms

CMS notes that the review proposes no specific rule changes, but it is likely to shape supervisory priorities, perimeter policy and future guidance. Firms are advised to map their current and planned AI use against the autonomy spectrum, review governance and senior-manager responsibilities, strengthen auditability, explainability and complaints handling, test AI-driven customer journeys against Consumer Duty outcomes, review operational resilience and third-party dependencies, and strengthen AI-enabled fraud controls while retaining human oversight. The Consumer Duty, which requires firms to deliver good outcomes for retail customers, remains the main lens. FCA Chair Ashley Alder said the regulator's principles-based, outcomes-focused approach, relying on the Consumer Duty and the Senior Managers Regime, has been critical to keeping pace with change.

The FCA said it will publish an AI good and poor practice report later in 2026, based on direct engagement with firms. Insurers and banks using AI in pricing, underwriting, credit decisions and customer service will be watching that report for more specific expectations.

This article is a summary of public documents and commentary and is general information, not legal advice. Firms should read the review and take advice on its application to their business.