The UK's Financial Conduct Authority commissioned the Mills Review in January this year. It was published on 6 July 2026. As far as I know, this is the first regulator-initiated study of its kind globally, examining how AI, especially agentic AI, will reshape retail financial services by 2030 and beyond. It was led by Sheldon Mills, formerly Executive Director of the FCA.
The Review weighs in at 146 pages. This is a very short summary. Obviously, there’s much more detail behind it.
Downloadable at: https://www.fca.org.uk/publication/corporate/the-mills-review.pdf
Key Findings on Advances in AI Capabilities
AI capability is advancing rapidly and is likely to continue doing so to 2030 and beyond, fundamentally transforming what both firms and consumers can accomplish with AI.
The most significant development is agentic AI, which is driving a shift from mere assistance to actual delegation - with AI systems now capable of handling longer, more complex tasks and taking multiple actions within both firm operations and consumer workflows.
But even these capable models need robust controls to ensure reliability, consistency, explainability, and accountability. Human oversight remains essential as more responsibilities are delegated to AI systems.
Future AI architectures are expected to fundamentally change what's possible, making AI systems more efficient, explainable, adaptable, and contextually aware - potentially moving toward more powerful general capabilities.
AGI and quantum computing present “disruptive uncertainties” with as yet unknown timeframes, but massive potential consequences. Regulators and firms must prepare now.
The Review identifies this rapid advancement in AI capability as the systemic driver that will trigger four major system shifts reshaping financial services to 2030.
The Four Systemic Shifts
- AI will transform firms' operations.
- Consumer journeys will be led by agentic AI.
- AI will reshape competition and market power.
- AI is likely to amplify fraud and cyber risks.
The Seven Priority Recommendations for the FCA
Secure and adapt the regulatory perimeter
As consumers increasingly rely on AI models and agents to help make financial decisions, the FCA must consider whether certain AI-enabled services may begin to cross regulatory boundaries. A perimeter review would determine whether legislative reform, perimeter changes or specific guidance are needed. The Review recommends that the FCA should undertake this review within three to six months of its publication.
Strengthen system-wide co-ordination and oversight
The FCA should establish and enhance co-ordination across domestic and international authorities and regulators to address the risks and opportunities that AI presents. This will include focusing on concentration and ecosystem dependencies, e.g. key AI providers and cloud service providers.
Monitor the transition to autonomous models and adapt regulatory frameworks
The FCA should consider monitoring the transition to more autonomous AI and adapt regulatory frameworks accordingly.
Scale up the FCA's AI Lab to support AI models and system innovation in financial services
The FCA should consider creating a structured capability, centred on its AI Lab, to better understand and assess AI models and systems used in the sector.
Enable foundations for agentic finance
The Review recommends that the FCA should lead the development of a trusted framework for agentic participation in financial services, clarifying how AI agents can be authorised, identified and held accountable.
Build and adopt an AI-enabled agentic supervisory model
In essence, the Review recommends that the FCA should itself develop and deploy an AI-enabled supervisory model, so turning the technology and processes to its advantage, including using agentic AI in supervisory and system-wide oversight, including providing intelligence, early warnings, and supporting enforcement.
Develop a trusted, public-interest, AI-enabled, financial capability service
The FCA should consider taking a more proactive, ‘hands on’ approach to financial capability and access to future vital AI-enabled services, to provide consumers with access to reliable financial information, guidance and support from trusted sources without having to subscribe to the most capable AI models or services.
What is this likely to mean for AI-related financial services regulation in the UK?
The Mills Review finds that the overall regulatory framework “remains sound”, reiterating that the FCA’s principles and outcomes-based approach, including the Consumer Duty and the Senior Managers Regime (SMR), operational resilience regulation, and other key features, was “designed to flex across changing business models”. The Review notes that those respondents who contributed to its thinking “did not seek changes to this system”, but do want clarity on how to interpret and govern increasing deployment of AI within the existing regulatory framework. In our experience, this has been a long-running concern of many of the financial services firms we advise.
So, while saying that the current regulatory framework provides a credible foundation for an AI-enabled financial system, the Review also notes that the regulatory framework is coming under increasing strain, which will surely only worsen as AI and agentic finance are integrated in firms’ operations, the markets, and the whole financial ecosystem.
My personal view is that the existing regulatory framework needs far more than clearer, AI/agentic AI-specific guidance from the FCA: it will need a more radical adaptation to address the specific threats (and opportunities) of AI at firm, market and systemic levels. There are hints in the Review itself that the strain on current regulation makes this inevitable.
