The UK Financial Conduct Authority published its first AI practice paper. It found AI deployment outpacing governance in financial firms. Key failures include retrofitted oversight, diffused accountability and monitoring gaps. Human-in-the-loop controls are sometimes a tick-box exercise. Monitoring often focuses on model metrics rather than consumer outcomes. Some firms overstate AI accuracy or leave consumers unclear on decision-makers. Senior managers sometimes delegate AI development without adequate challenge. The paper expects firms to document audit trails for compliance. The FCA will not write AI-specific rules but rely on existing frameworks. Its oversight now covers AI use in underwriting and claims across the MGA sector.

• 5 Oct 2026 publication date
• 1,200+ businesses in Gallagher survey
• fewer than 50% with formal AI risk framework