A July 2026 review by Sheldon Mills for the UK Financial Conduct Authority.
Structure
The review organizes its analysis around an AI autonomy spectrum and four system shifts: the transformation of firms; new consumer journeys; a reshaped competition landscape; and amplified financial crime and cyber risk. It carries priority recommendations for AI-enabled finance and annexes covering engagement-paper responses, consumer survey findings, focus groups, economic growth impacts, international approaches, and infrastructure for agentic finance.
The autonomy trajectory
The review's central projection concerns where firms sit on that spectrum. Today, firms "have begun deploying AI for specific tasks, mostly keeping humans in the loop. This is changing. Firms are already piloting and rolling out use cases with greater autonomy, and by 2030 many could have moved significantly further along the autonomy spectrum, embedding AI into almost every function from customer support and underwriting to compliance, claims and product design."
For leading firms, "AI may become the main method by which they process information, serve customers, and evidence outcomes" — the last item mattering most for a conduct regulator, since evidencing outcomes is how firms demonstrate compliance.
What happens to human oversight
The review's sharpest observation concerns the degradation of oversight as a category rather than its removal: "As autonomy grows, the role of people within firms changes, from operators close to each decision, towards collaborators, approvers and, eventually, observers who monitor outcomes and step in when systems move outside agreed parameters."
Its assessment of what that requires: "This is a substantial organisational shift, requiring new skills and a clearer account of what human oversight actually involves." That is a regulator stating that the human-in-the-loop requirement common across AI regulation does not currently have a definition adequate to the deployments it is meant to govern.
Relationships
- supports: Financial Services — AI Deployment — the fullest regulatory assessment of AI in UK retail finance
- related: Agent Autonomy Spectrum (5 Levels) — the review's organizing frame
- supports: Human Oversight — states that the requirement lacks a definition adequate to the deployments it governs
- related: Scalable Oversight, Agentic AI, Financial Conduct Authority (FCA)