The shift from AI as a helpful tool to AI as an autonomous decision-maker is reshaping how financial services will operate, according to a major regulatory review released in July 2026. The UK Financial Conduct Authority (FCA) published its Mills Review, which identifies a fundamental transformation underway in retail banking: AI is moving from supporting human decisions to making decisions independently on behalf of customers and firms. This transition, driven by agentic AI systems capable of taking action without constant human approval, represents the most significant challenge regulators and banks now face.

What Does the Shift From Assistance to Delegation Actually Mean?

The Mills Review introduces an "AI autonomy spectrum" that captures how the role of humans changes as AI becomes more autonomous. At one end, humans act as "Operators," using AI as an on-demand tool to support their decisions. As AI capabilities expand, humans become "Collaborators" who consult with the system, then "Approvers" who review AI recommendations before they execute, and finally "Observers" who simply monitor outcomes while the AI system acts continuously within pre-set boundaries.

This progression is not theoretical. The Review notes that initial AI deployments in retail financial services have mostly kept humans in the Operator or Collaborator roles, but current deployments are laying the foundation for future expansions toward greater autonomy. By 2030, leading firms in their respective retail markets are likely to have embedded AI into almost every aspect of their business, making it the principal method by which they serve customers, process information, and deliver outcomes.

Consumer appetite for this shift is already emerging. Research commissioned for the Review found that one in five UK adults are open to AI making decisions for them, suggesting that the move toward agent-led customer journeys is credible. Consumer demand for AI appears strongest in complex, high-stakes areas including debt advice, pensions, and investments, where navigating options is difficult and the stakes are high.

How Should Banks Prepare Their Governance for Autonomous AI Systems?

The Review emphasizes that governance will become an increasingly important differentiator and potentially a competitive advantage as firms deploy AI more widely and with greater autonomy. However, existing governance and model risk management frameworks will face new pressures. The FCA expects a key focus to be whether firms' governance evolves in step with their use of AI.

Several critical governance adaptations are necessary:

• Shift from validation to live monitoring: Unlike traditional models that remain static after deployment, AI models may update continuously, draw on third-party inputs, and produce probabilistic rather than deterministic outputs. This means firms must move beyond validating models at the point of deployment and instead monitor for issues such as model drift, model degradation, and statistical outliers throughout the system's operational life.
• Enhanced financial crime and cyber controls: As AI systems gain autonomy, firms will need clear permissions, effective monitoring, auditability, and robust escalation mechanisms to prevent misuse. The Review emphasizes that these are not additional safeguards but the conditions that enable AI to be deployed in regulated environments.
• Explainability across the entire AI lifecycle: A recurring theme throughout the Review is the need for explainability. AI models may produce different answers to similar questions, creating challenges for firms expected to provide consistent treatment, maintain clear audit trails, and deliver explainable outcomes. Firms must consider the quality of data inputs, pre-deployment testing, and ongoing monitoring.

The Review stresses a counterintuitive point: better models do not reduce the need for controls; they increase it. Capable models still require controls for reliability, consistency, explainability, and accountability, alongside human oversight as more is delegated to them.

Why Workforce Substitution Decisions Require Agentic Frameworks Too

Beyond financial services, the broader adoption of agentic AI is forcing organizations across sectors to rethink how they make automation decisions. A research paper published in July 2026 proposes an agentic AI-inspired framework specifically designed to help organizations predict workforce substitution and make informed automation decisions.

The framework combines regression models to predict workforce substitution timelines with classification models that recommend automation strategies, including Monitor, Assist, and Automate. A stacking-based ensemble mechanism, which combines multiple machine learning models, enhances predictive accuracy and robustness. Critically, the framework includes a risk-aware decision intelligence layer that transforms predictive outputs into actionable recommendations, enabling organizations to move beyond passive forecasting toward adaptive decision support.

When tested on the 2026 Intelligence Economy: Labor vs. AI Compute dataset, the hybrid model achieved an R-squared score of 0.857 for workforce substitution prediction and a classification accuracy of 95.4% for automation strategy selection, outperforming individual baseline models across multiple evaluation metrics. Cross-validation, robustness analysis, scalability testing, and explainability techniques confirmed the reliability, stability, and practical applicability of the framework.

The research demonstrates that integrating hybrid machine learning with agentic decision intelligence provides an effective, interpretable, and scalable solution for analyzing workforce transformation and supporting automation-related decisions in AI-driven economic environments. This is particularly important because existing approaches typically address workforce forecasting and automation policy selection as separate tasks, limiting their ability to provide integrated and actionable decision support.

What Are the Broader Implications for Regulated Industries?

The Mills Review's central insight applies far beyond banking: the real AI revolution is not about faster computation or better predictions. It is about systems that take action autonomously. As the Review observes, "AI becomes more significant for financial services when it moves from supporting action to taking action." This shift from assistance to delegation, driven by agentic AI, underpins almost every regulatory and operational challenge ahead.

The Review does not recommend AI-specific regulation, nor does it propose wholesale reform of the existing UK regulatory framework. Instead, it concludes that the current framework provides a strong foundation for AI-enabled finance. However, it recognizes that a number of existing regimes, including operational resilience, the regulatory perimeter, the advice guidance boundary, the Senior Managers Regime, and the Consumer Duty, will come under increasing pressure as AI evolves from an assistive tool to autonomous systems capable of acting on behalf of firms and consumers.

The Review therefore recommends a "disciplined and progressive" adaptation of the existing framework to ensure it can continue to deliver good outcomes for consumers while supporting innovation, competition, and growth. For firms, this means governance is no longer a compliance checkbox but a strategic capability that will differentiate winners from laggards in an AI-driven financial services landscape.