Technology Transactions Soundbite: HM Treasury's Financial Services AI Adoption Plan
HM Treasury's Financial Services AI Adoption Plan - Providing Answers to the FCA/Mills Review Call-to-Arms on Agentic Finance
UK: HM Treasury’s Financial Services AI Adoption Plan, published on 14 July 2026, sets out 10 recommendations for AI covering regulatory framework, regulatory perimeter, resilience, skills, and agentic payments.
Produced by the government's independent AI champions—Harriet Rees (Group Chief Information Officer at Starling Bank) and Dr Rohit Dhawan (Head of AI and Advanced Analytics at Lloyds Banking Group)—the Plan has been warmly welcomed by the government, which has accepted all recommendations directed at it. Ministers now intend to work closely with regulators and industry on implementation of next steps.
Key messages
No requirement for standalone AI regulation - but a mandate for clear, "authoritative single source" cross-regulator guidance
Recommendation 1 asks regulators to work together so that expectations on firms are clear – with cohesive and easily understood (rather than piecemeal) regulatory guidance a key focus. This is an answer to the fragmentation problem outlined in the Mills Review (see my previous article on Mills) – not unexpectedly, the government is going to seek to address with coordination and clarity, rather than new standalone legislation.
AI initiatives and support to be clearly promoted and accessible
Recommendation 1 also highlights the need for services that support AI adoption to be accessible – suggesting the creation of a "Financial Services AI adoption Support Hub", with two potential functions: (i) an Information Portal with details of current initiatives – e.g. FCA AI lab updates; and (ii) a means of accessing supervisory / SMEs support.
The risks of LLM financial guidance and advice to be subject to formal FCA review
Recommendations 2 and 3 instructs the FCA to undertake a comprehensive review of the consumer, competition and wider impacts of financial guidance and advice-like outputs generated by general-purpose LLMs – with the Plan noting disclosures and education for consumers as possible answers to the associated consumer risk – this very much lines up with the EU AI Act's transparency and AI literacy remit.
Resilience – addressed by three key mechanisms
Recommendations 4 to suggest (1) that critical AI and cloud providers should be assessed under the "Critical Third Parties" regime to ensure systemic risks are considered and mitigated against; (2) that an industry-wide "AI Incident & Near-Miss" repository, should be created to foster a culture of collective intelligence – potentially supported by the Cross Market Operational Resilience Group; (3) the establishment of a voluntary, industry-led third-party assurance scheme to be explored to enable consistent assessment of general-purpose or third-party AI models, systems, applications and other tools. These mechanisms should go some way to addressing the Mills Review's concern that firm-by-firm supervision is insufficient.
Bolstering AI skills and attracting global AI talent to the UK are key
Recommendations 7 to 9 push industry participation in the Financial Services Skills Compact, a sector-wide AI skills plan built on the Financial Services Skills Commission's research, and pragmatic visa adjustments to attract global AI talent. The framing is that regulated AI deployment needs deep capability in model risk management and governance, not just engineers, extending to risk managers, legal experts and frontline staff.
Agentic payments gets the trust framework Mills asked for
Recommendation 10 is the Plan's centrepiece and its most direct answer to the Mills Review's agentic finance framework: it calls on the Treasury to use its upcoming consultation on modernising payment services regulation to establish a trust framework built on three pillars: (1) legal and liability frameworks to "unambiguously" assign accountability when autonomous agents transact; (2) "Know Your Agent" protocols for standardised identity and verification of AI and autonomous software agents; and (3) interoperable technical standards to ensure safe, dependable and seamless authentication.
Conclusion - from diagnosis to delivery plan
Where the Mills Review called out the issues and sketched out potential solutions, the AI Adoption Plan provides recommendations to assist in fully addressing real life concerns with greater Agentic AI adoption by firms.
Want to know more?
Technology Transactions Soundbite: The FCA Mills Review & its Agentic AI Finance Framework - The real message for UK Financial Services, 13 July 2026
The FCA’s Mills Review published on 06 July 2026 is not just another paper on AI; it's a milestone signpost that the architecture of UK retail finance is about to change.
Other author: Aimi Gold, Senior Associate
