On the 6th July 2026, the FCA published the final findings of the Mills Review, its assessment of how advances in artificial intelligence (AI) could reshape financial services by 2030 and beyond. Drawing on submissions from firms, technology providers, consumer groups, academics and regulators, alongside consumer research involving more than 5,000 UK adults, the review explores how financial services may evolve.
For most financial services firms, AI is no longer a future consideration. It is already being used to support customer service, identify fraud, analyse data and improve operational efficiency. The Mills Review recognises this reality and focuses on a more practical question: what happens as AI takes on a greater role in shaping decisions and customer outcomes?
Today, AI typically supports human decision-making. It helps employees process information, identify patterns and complete tasks more efficiently. Looking ahead, however, the review describes a future where AI increasingly influences decisions, recommends actions and, within agreed parameters, carries out activities on behalf of firms and consumers. As this shift occurs, the challenge moves beyond adoption and towards oversight. How can firms understand, monitor and evidence the outcomes being delivered by increasingly autonomous systems?
From AI assistance to AI delegation
One of the review’s central themes is the transition from AI as a productivity tool to AI as a delegated decision-maker. This does not mean removing people entirely from processes. Rather, it reflects a gradual increase in the responsibility’s organisations may choose to allocate to technology.
A simple example is a customer using AI to compare savings accounts. Initially, the technology may simply present options. Over time, it could recommend a more suitable product based on predetermined preferences. Eventually, consumers may permit AI systems to switch products automatically within agreed limits, reducing the need for direct intervention.
Financial services firms may follow a similar path. AI may move from identifying potentially vulnerable customers to recommending the most appropriate support pathway or intervention. It may move from highlighting risks to proposing actions designed to mitigate them. The more responsibility delegated to technology, the more important it becomes to understand how decisions are being made and whether outcomes remain fair and appropriate.
What an AI-enabled financial system could look like
The Mills Review outlines several ways AI could reshape financial services over the coming decade. One is the increasing integration of AI into firms’ core operational processes, including customer support, compliance monitoring, claims handling and risk management. Another is the emergence of more agent-led customer journeys, where consumers rely on AI systems to help manage aspects of their financial lives.
The FCA’s research found that, in certain circumstances, one in five consumers would already be open to AI systems making financial recommendations or taking actions on their behalf within pre-defined goals and parameters. The review suggests this could help address long-standing challenges such as low engagement with financial planning, limited access to financial guidance and low switching rates between products.
However, the review also highlights significant risks. Deepfakes, synthetic identities, AI-enabled fraud and increasingly sophisticated social engineering attacks are likely to become more prevalent as the technology evolves. At the same time, AI is expected to become an increasingly important tool in detecting and responding to those threats.
The role of humans is changing, not disappearing
The Mills Review does not predict a future where humans disappear from financial services. Instead, it suggests that human responsibilities are likely to evolve. As AI systems become more capable, people may spend less time executing individual tasks and more time setting parameters, monitoring outcomes, investigating exceptions and challenging decisions generated by technology.
That shift introduces important questions around oversight. Meaningful governance requires more than simply placing a person somewhere within a process. Firms need clarity around what information individuals receive, what decisions they are responsible for making, when intervention is expected and whether they have sufficient understanding of the technology they oversee.
As AI becomes increasingly embedded within decision-making processes, firms will need confidence that human oversight remains effective in practice rather than existing purely as a governance principle.
Why accountability becomes harder
One of the review’s most important observations is that accountability becomes more challenging as AI ecosystems become more complex. Firms may rely on multiple models, third-party providers and interconnected technology supply chains, with different systems influencing different stages of a customer journey. Understanding precisely how outcomes have been produced can become increasingly difficult.
This creates a growing need for evidence. It is no longer enough for firms to have governance frameworks documented on paper. Organisations increasingly need the ability to monitor outcomes, identify potential concerns and understand how decisions are being reached in practice. If issues occur, firms must be able to demonstrate what happened, why it happened and whether appropriate controls were operating at the time.
The review also highlights the challenge of model drift. A system that performs accurately during testing may behave differently over time as customer behaviour, data sources or market conditions evolve. An AI solution used to identify vulnerable customers, for example, may require ongoing monitoring to ensure it continues to perform consistently and fairly across different groups.
For many organisations, this means AI governance cannot be treated as a one-off approval exercise. It requires continuous monitoring, ongoing assurance and the ability to evidence outcomes over time.
Why the Consumer Duty remains central
One of the more reassuring conclusions from the Mills Review is that it does not call for an entirely new regulatory framework. Instead, it concludes that existing regimes, including the Consumer Duty and the Senior Managers Regime (SMR), provide a strong foundation for governing an increasingly AI-enabled financial system.
The challenge lies in demonstrating how those obligations continue to operate as technology becomes more influential. If AI is used to personalise customer journeys, firms need confidence that good outcomes are being delivered consistently. If AI helps identify vulnerable customers, firms need evidence that the process remains accurate and fair. These are not fundamentally new regulatory requirements, but they are likely to require new approaches to monitoring and assurance.
Regulation is becoming more data-driven too
The Mills Review is not solely concerned with how firms use AI. It also considers how supervision could evolve in response. One recommendation is the development of an AI-enabled Agentic Supervisory Model, which would allow the FCA to identify emerging harms, monitor trends and detect system-wide risks more effectively.
As firms become increasingly data-driven and AI-enabled, regulators are likely to do the same. Future supervision may place greater emphasis on outcomes, patterns and evidence, with regulators looking for clear indications that firms understand the impact of technology on customer experiences and decision-making.
Preparing for what comes next
The future described in the Mills Review may still be emerging, but many of the challenges it highlights are already becoming relevant. Firms do not need fully autonomous AI systems before questions around accountability, oversight and customer outcomes become important. As AI becomes more deeply embedded across financial services, organisations will need greater confidence that they can monitor decisions, identify risks and demonstrate that good outcomes are being delivered consistently over time.
At Recordsure, we help firms gain that confidence. Our AI-powered monitoring and quality assurance solutions enable organisations to review customer interactions at scale, identify conduct and compliance risks, monitor customer outcomes and provide evidence that controls are operating as intended. As financial services become increasingly AI-enabled, the ability to demonstrate effective oversight may become just as important as the technology itself. We help firms build the visibility, assurance and confidence needed to navigate that future.
