In brief
On 27 January 2026 the Financial Conduct Authority (FCA) launched the Mills Review to examine the long-term impact of AI on financial services. Led by Sheldon Mills, this initiative invites industry feedback to help shape how AI might transform consumer experiences, market structures, and regulatory approaches in retail financial services. The call for input closes on 24 February, following which Mills will present recommendations to the FCA board in the summer, culminating in an external publication to foster informed debate.
In more detail
The Mills Review
The review focuses on four crucial areas:
- AI technology evolution: By 2030, AI systems are expected to become more autonomous and adaptive, with agentic AI being more capable of independent decision-making and continuous learning. Furthermore, multimodal AI, which integrates text, speech, images and other data, could radically alter customer journeys and service delivery. The FCA is keen to understand how, at present, firms expect to use AI systems within their business, and their plans for how their use of AI will evolve.
- Impact on markets and firms: AI is already generating efficiencies across core activities within financial services firms, and its influence is set to grow. As customers delegate decisions to AI, competition among financial services firms could potentially increase through the development of novel value propositions; however, new forms of market power may also emerge, for example if certain AI providers favour certain financial services firms. Financial services firms must adapt to maintain relevance and competitiveness. The FCA is keen to understand the potential impact on, among others, market structure and customer relationships, and to explore whether AI systems provide services functionally equivalent to regulated activities such as advice or intermediation while remaining outside the regulatory perimeter.
- Consumer trends: Consumers may increasingly interact with financial services via AI interfaces, with expectations likely to shift towards highly personalised, automation-driven financial journeys alongside reduced tolerance for friction, poor value or opaque outcomes. While this has the potential to improve financial outcomes, it also raises concerns about reliance on unregulated advice, vulnerability to mis-selling, model bias or hallucination risks, and AI-driven fraud. The FCA is keen to hear views on both the opportunities and challenges this presents.
- Regulatory approach: The FCA intends to remain an outcomes-focused regulator, leveraging existing frameworks such as the Consumer Duty and Senior Managers and Certification Regime (SMCR) to ensure flexibility and compliance. The regulator is seeking input on how the FCA can take advantage of the opportunities AI presents, but also, among others, on whether existing frameworks are suitable for the regulation of AI. For example, the FCA will assess how relevant senior managers under the SMCR can continue to discharge their responsibilities for the deployment and maintenance of AI systems, and whether Consumer Duty expectations should be revised to account for the impact of AI.
Key considerations for AI use in retail financial services
Use of AI to date
To date, AI has been deployed in the financial services sector for a variety of purposes, including assisting with creditworthiness assessments, enhancing customer due diligence and financial crime monitoring (such as fraud detection), and providing customer support interfaces such as online chatbots. Each of these use cases introduces unique risks, however, and requires a careful analysis of data protection issues. For example, AI-driven credit assessments can inadvertently perpetuate bias if historic data contains underlying prejudices. Similarly, AI chatbots may occasionally provide unclear or inaccurate responses, an issue that may be particularly acute for vulnerable customers. The FCA is aware of these risks, as highlighted by Sheldon Mills in his speech launching the review. As AI becomes more advanced and autonomous, these risks could increase.
There are also challenges for firms in applying their responsibilities and obligations under the SMCR to the adoption and use of AI, which may have resulted in a slower uptake than might otherwise have been the case. Given the FCA’s position that senior managers should maintain oversight of AI deployment, for example, the lack of “explainability” of many AI models could be seen to conflict with the SMCR’s requirement for senior managers to demonstrate they understand and control risks within their areas of responsibility.
The FCA’s evolving approach
At present, it appears that the FCA intends to maintain its current approach to regulating AI, refraining from introducing additional or bespoke regulations specifically for AI. Instead, the FCA will continue to leverage existing principles-based and outcomes-focused regulatory frameworks to guide firms’ adoption and use of AI, providing firms with a degree of flexibility in their approach to AI. However, this approach is not without challenges: the FCA’s supervisory approach has been reactive, leaving firms with little practical clarity and guidance on how to apply existing regulatory frameworks to their AI deployment, pushing the burden onto firms to take a position forward.
We may, nonetheless, see this approach evolve in the coming months, given the suggestion in a recent House of Commons Treasury Committee report (published a week before the launch of the Mills Review) that this lack of clarity from the FCA may have contributed to a chilling effect on AI adoption. In response, the Treasury Committee has recommended that the FCA should issue guidance on the application of its existing regulatory frameworks to AI usage, that HM Treasury should designate major AI and cloud providers as critical third parties under the Critical Third Parties Regime (giving the regulators greater supervisory control over those providers), and that the Bank of England and the FCA should conduct AI-specific stress testing. The Treasury Committee expects to see movement on the first two of those recommendations by the end of 2026.
Going forward
It is evident that AI will play an increasingly prominent role in financial services. In particular, well-designed and robust AI tools can drive efficiency and deliver cost savings, helping firms meet the Consumer Duty’s expectations around price and value. Nonetheless, firms must ensure that any AI tools or services they adopt are thoroughly evaluated against the FCA’s evolving regulatory expectations.
Firms that are especially reliant on AI, or that plan to increase their use of such technologies, should consider responding to the FCA’s Call for Input. The findings of the Mills Review will influence how the FCA interprets and applies regulatory requirements to AI tools and software, providing an opportunity for stakeholders to help shape the future regulatory landscape.
