FCA launches long-term review of AI in insurance and retail finance
The UK's Financial Conduct Authority has opened a review into how AI will impact insurance markets and retail financial services over the long run. The regulator says AI already influences pricing and claims operations, and the momentum is building. Continuous risk assessment and automated claims handling are now common. Deeper shifts are coming.
What's under review
- AI-driven agents: Tools that compare policies, recommend alternatives, and help customers switch with fewer steps. Good for consumers, but it challenges pricing strategies that depend on inertia.
- Fraud and manipulation: More advanced scams as tools get stronger and cheaper. Firms and regulators will need faster countermeasures.
- Three core questions: How AI develops next; what that means for markets, firms, and consumers; and whether current regulation can keep up without breaking.
Wholesale markets sit outside the scope. Broader social effects are mostly out as well, though indirect links will be considered where they matter.
Regulatory stance and timeline
The FCA is not writing new prescriptive rules through this exercise. Instead, it will test whether current frameworks stretch far enough and adjust guidance if needed. Findings will go to the FCA board this summer. Industry and other stakeholders have until Feb. 24 to submit views.
The authority will also examine its own supervisory methods, acknowledging that AI could change how oversight works. For updates and submission details, see the FCA website.
Why this matters for insurers
For sector-specific examples and resources, see AI for Insurance.
Distribution and retention: If AI agents reduce friction, switching costs fall and churn rises. Customer value strategies built on stickiness lose ground. Expect sharper price competition and faster product comparisons at the point of need.
Underwriting and pricing: Continuous risk signals tighten risk selection and micro-segmentation. That can improve technical pricing, but it raises fairness and explainability questions that won't go away.
Claims operations: Automation shortens cycle times and lowers handling costs. The flip side is model risk, error propagation at scale, and the need for audit trails that stand up to scrutiny.
Fraud: Generative tools lift the ceiling on social engineering, document tampering, and synthetic identities. Counter-fraud will lean more on real-time signals, device intelligence, and cross-carrier collaboration.
What to do now
- Pressure-test pricing strategy: Model higher switching and lower inertia. Rework renewal playbooks, broker levers, and incentives for retention without hidden friction.
- Stand up model governance: Inventory every AI model across pricing, claims, marketing, and fraud. Set owners, thresholds, bias tests, and rollback plans. Document decisions in plain language.
- Reinforce claims controls: Keep a human in the loop for complex or high-exposure cases. Track error rates and appeal paths. Log prompts and outputs where generative tools touch customer data.
- Upgrade fraud defenses: Add image and document forensics, voice risk analytics, and device intelligence. Share signals where lawful. Run red-team tests using the same tools fraudsters use.
- Vendor due diligence: Ask for training data sources, monitoring metrics, update cadence, and security posture. Bake service levels and audit rights into contracts.
- Consumer outcomes: Check pricing and claims automations against duty requirements on fairness, clarity, and redress. Make opt-outs and support easy to find.
- Scenario planning: Simulate a market where AI agents control a chunk of distribution. Stress test capacity, rates, and lapse impacts under aggressive quote-shopping.
PRA priorities alongside the FCA review
The Prudential Regulation Authority set separate priorities for 2026. These include oversight of high-demand pension risk transfers, a semi-live crisis exercise in spring, and a consultation on a new captive insurance regime. For more, visit the PRA website.
Upskill your teams
Front-line literacy reduces risk and speeds adoption. If you need a quick overview of practical tools used across finance and insurance, this list can help: AI tools for finance.
Bottom line
AI is already affecting pricing, claims, and fraud. The FCA's review signals a long-term shift: competition will intensify, oversight will tighten, and defensible execution will matter more than bold promises. Build the controls now, and treat distribution, pricing, and fraud as one connected system.
Your membership also unlocks:
