Financial Services AI Adoption Plan

Last week as part of Mansion House former Chancellor Rachel Reeves published Government’s Financial Services AI Adoption Plan. Developed with the government-appointed AI in FS Champions Harriet Rees (Chief Information Officer, Starling Bank) and Dr Rohit Dhawan (Head of AI and Advanced Analytics, Lloyds Banking Group), the plan is the result of extensive engagement with industry and regulators. techUK was delighted to feed into this process on behalf of our members. The plan’s ten recommendations span five themes covering regulatory clarity, the regulatory perimeter, resilience, skills & talent, and agentic payments. It is worth noting that whilst Reeves has left office following the change of government, the AI in FS adoption plan represents an independent set of recommendations developed with industry and the plan's workstreams are expected to continue under the Burnham government.

The opening headline is that the UK’s current principles-based, outcomes focused regulatory approach to AI in financial services is working well and should not be changed. techUK members support this position. The key regulatory frameworks and supervisory guidance which underpin this approach are and will continue to be: Consumer Duty, SM&CR, the Operational Resilience Regime, Model Risk Management Principles, and more recently the Critical Third Parties Regime (CTP Regime).

Considering the support for this regulatory architecture, the core industry challenge remains accessibility and practical application of said architecture rather than wholesale reform. The challenge of practical application and navigability is particularly true for smaller firms and new market entrants who can find the existing regulatory pathways difficult to navigate. The proposed Financial Services AI Adoption Support Hub as part of the plan is a practical answer to that challenge. This hub will bring together cross-regulatory guidance in one single, accessible place and is a sensible response to what techUK members have been asking for.

The plan's resilience recommendations are where things get more structurally significant. The CTP designations of AWS, Microsoft, Google Cloud and Oracle on 10 July were a landmark moment for the technology sector. This was the first time non-financial services firms have been brought within the supervisory perimeter of UK financial regulators. The week before the announcement as part of Agents of Change techUK called for progression on the CTP Regime to provide industry much needed clarity so we were pleased to see the designations progress.

The plan also calls for the assessment of 'critical AI and cloud providers' under the CTP regime. At a minimum this language covers AI-enabled cloud services and potentially extends to frontier model providers further down the line although does not imply this directly. The direction of regulatory travel seems to be clear with the plan signalling the regulatory gaze is starting to move up the AI stack with the CTP Regime the likely framework that will be used if this happens.

The proposed voluntary AI third-party assurance scheme is genuinely new ground and one of the most commercially significant recommendation in the plan. This idea came up in our industry engagement when writing techUK’s Agents of Change report. The concept is a standardised audit protocol where a qualified assessor evaluates AI model and application providers against agreed standards, so that a certification can be accepted by regulators as evidence of baseline due diligence. If it works, it could dramatically reduce the duplicative, firm-by-firm model risk assessment burden that currently creates friction in AI procurement. For technology companies supplying AI to financial services, it creates a clear route to market credibility. For regulated firms, it lowers the cost of responsible adoption. It is one of those recommendations that, if delivered, would quietly reshape how the sector procures and governs AI.

The plan also touches on aspects of the wider regulatory perimeter. According to FCA data, 26% of UK adults already use general-purpose tools for financial advice, which does not fall under the regulatory purview. Actual regulated advice reaches only around 9% of UK adults. The asymmetry makes for uncomfortable reading as regulated firms face strict obligations and liability whilst unregulated advice exists without similar challenges. The government plan calls for an FCA review of the consumer and competition impacts of advice-like outputs from general-purpose LLMs. There was industry discussion about potentially deregulating financial services in this area to match the less regulated technology sector, although it is now clear this will not happen. Expect this area to continue to evolve at pace. Particularly considering the Mills Review which recommended the FCA run a 3–6-month review of the regulatory perimeter to ensure consumers do not suffer harm when taking financial advice from unregulated LLMs. techUK looks forward to engaging with the technology sector and feeding into this review.

On agentic payments, the plan's trust framework recommendation covers legal and liability, Know Your Agent protocols, and machine-to-machine authentication standards. These areas represent the desired governance architecture for the sector as agentic payments redefine the sector at pace. During the Mansion House speech itself, Reeves announced that techUK member Mastercard has chosen the UK as the first place in Europe to launch its new agentic payment tools.

techUK’s work in this area

techUK launched Agents of Change: Generative and Agentic AI in Financial Services 2026, a report and accompanying strategic summit on 24^th June 2026. This was two weeks before the Mansion House speech and the accompanying AI adoption plan for financial services, and one week before the findings of the FCA’s Mills Review. The Agents of Change report was launched following extensive engagement with the financial technology ecosystem, including with the government appointed AI in FS sector champions ahead of their recommendations to the Chancellor which formulated the AI in FS sector adoption plan.

The Agents of Change report documented live advanced AI deployments across fraud, complaints, capital markets, insurance and enterprise AI. It provided a comprehensive look on how generative & agentic AI are being used in the financial services sector today and featured use cases from techUK members including Microsoft, Cohere, DXC, Google, Nvidia, Visa and IBM to name a few. The report sets out three shared directions for the ecosystem which were: establishing a dedicated agentic AI governance workstream, expand shared AI deployment infrastructure, and increase AI supply chain resilience including the advancement of the CTP Regime. techUK were pleased to see progression on the first designations of the CTP Regime shortly after the report was published. We were also delighted to see an acknowledgment of the utility of the FCA’s AI Lab, and commentary on the accessibility of some of the lab’s services. techUK believes continued investment in the FCA’s AI Lab and shared deployment infrastructure more broadly is needed, particularly around the sandbox services and the size of the cohorts, which could be expanded further to support the ecosystem and we look forward to seeing how this develops.

The Agents of Change report launch was keynoted by FCA CEO Nikhil Rathi. Nikhil’s speech outlined the latest thinking from the FCA on AI in financial services. You can read the full speech here. He noted the report was already informing how the FCA were working internally. Some points of note from his speech include the reframing of the FCA’s role towards stewardship rather than supervision, that the facilitating of intelligence between firms under competition law may happen to enhance collaboration on frontier AI risk, and that the FCA expected tech firms to ‘step up’ in the fight against fraud.

The wider Mansion House picture

To touch briefly on the wider Mansion House agenda, which may be subject to change with the new Government - Reeves’ Mansion House speech contained several announcements that will matter to techUK members.Beyond the AI Adoption Plan, Reeves’ Mansion House speech contained several announcements that will matter to techUK members. The UK is set to become the first G7 country to issue a Digital Sovereign Bond by early next year, the Great British Tokenised Deposit initiative is moving to pilot transactions, and the FCA-approved natively tokenised authorised fund, announced by Rathi at the Agents of Change launch. For those following the digital assets agenda, these are connected moves toward the programmable, tokenised wholesale market infrastructure that agentic finance demands. techUK's Digital Assets Working Group will track these developments closely.

The former Chancellor also reaffirmed the UK's Sovereign AI ambitions which includes a Sovereign AI unit, an AI hardware plan, an AI Economics Institute. Her speech framed domestic AI capability explicitly as a national security priority as well as an economic one. techUK’s Financial Services Council will explore the theme of sovereignty relating to both AI and cloud later this year.

Reeves also placed improving access to growth capital for smaller businesses at the centre of the government's growth agenda, announcing a significant expansion of the British Business Bank's Growth Guarantee Scheme backed by a £6.5 billion uplift in capacity. The changes extend the maximum loan term from six to ten years, increase the turnover eligibility threshold from £45 million to £54 million, and are expected to unlock an additional £2 billion of SME lending annually by 2028/29, supporting around 12,000 more businesses each year.

Alongside this, Reeves announced a £500 million package to back innovative SMEs and scale-ups, including ringfenced guarantee capacity for IP-rich businesses, enhanced support for community development finance institutions (CDFIs), a new ambition to expand community lending, and a joint British Business Bank and UK Export Finance initiative to improve exporters' access to finance. Together, the measures are intended to address the long-standing funding gap facing UK SMEs and scale-ups by widening access to debt finance, particularly for businesses that have traditionally struggled to secure bank lending.

What comes next

The AI in FS adoption plan creates a clear forward programme for the second half of 2026. The regulatory perimeter review, the third-party assurance framework, and the agentic payments consultation are the three workstreams where techUK members will want to engage most actively. The FCA's good and poor AI practice publication, the joint FCA and Bank of England AI survey due in the autumn, and the forthcoming HMT payments consultation will shape the detailed regulatory landscape as it develops.

If you would like to learn more about techUK’s financial services programme of work please get in touch with James Challinor ([email protected]) and Lourdes de Miguel ([email protected]).

Financial Services Programme activities

The techUK Financial Services programme connects tech firms, the FS industry, and regulators to ensure innovation and technology can be fully embraced. Through market engagement activities and events, we help to empower decision makers and aid collaboration.