The Mills Review could reshape AI governance in financial services

What happened?

The Financial Conduct Authority (FCA) has published the final findings of the Mills Review, assessing how advances in artificial intelligence (AI) could reshape financial services by 2030 and beyond. Commissioned by the FCA Board, the review draws on 140 written submissions and consumer research involving over 5,000 UK adults.

The review concludes that the industry is moving from a world where AI primarily supports human decision-making to one where it increasingly influences, recommends, and carries out actions autonomously on behalf of businesses and consumers.

Why does it matter?

This represents a critical shift from AI assistance to AI delegation. As firms move from using AI as a simple productivity tool to delegating complex tasks (such as automated rate switching or vulnerability assessments), the traditional lines of accountability under the Senior Managers Regime (SMR) and Consumer Duty will be tested.

Continuous monitoring and testing of model drift are vital. A model that performs accurately during testing may behave differently months later when customer behaviour, data inputs, or market conditions change, potentially producing unfair outcomes for different customer groups.

Who is affected?

This regulatory focus directly impacts all financial institutions utilizing AI technologies, particularly compliance, risk, and IT teams responsible for managing automated decision systems, customer support channels, and claims-handling tools.

Key risks

• Accountability Dilution: Difficulty in tracing responsibility across complex AI supply chains and third-party models.
• Model Drift: AI models producing unintended or unfair outcomes over time as real-world data and conditions deviate from testing baselines.
• Deepfake and Fraud Threats: Growing exposure to AI-enabled fraud, synthetic identities, and sophisticated social engineering attacks.

Actions to take

1. Map AI Supply Chains: Establish clear ownership and documentation of all third-party models and data integrations.
2. Implement Model Drift Monitoring: Conduct continuous testing and outcome-based audits on active AI models to ensure fair customer treatment.
3. Align with SMR and Consumer Duty: Integrate AI decision-making parameters within existing governance and senior management accountability frameworks.

Wider implications

The FCA is actively preparing for an AI-driven system by developing its own AI-enabled ‘Agentic Supervisory Model’. This will allow the regulator to identify cross-firm trends and emerging harms across the entire financial ecosystem using automated supervision.

Recommendations

Firms should not wait for fully autonomous AI systems to emerge. Compliance and risk teams must proactively assess their current AI governance, ensuring human oversight is meaningful rather than just a rubber-stamping exercise.

At TCC, we work with firms to navigate regulatory change, strengthen governance frameworks and ensure evolving technologies are implemented in ways that support both regulatory compliance and customer trust. Through our technology partner Recordsure, firms can also access AI-powered tools that help monitor customer interactions, oversee evidence, identify risks, and support better outcomes at scale. Whether reviewing AI governance, assessing Consumer Duty implications or building robust oversight frameworks, we can help firms prepare for an increasingly AI-enabled financial system with confidence.