Property and casualty carriers’ underwriting, risk, and technology leaders must decide how far agentic AI can act on a submission. Datos Insights research finds that 74% of large carriers and 24% of midsize carriers are piloting agentic capabilities. Technology choices made in the fourth quarter of 2026 and first quarter of 2027 will shape competitive position through 2030.
Agentic AI performs defined work within clear authority, controls, and escalation paths. The report argues that agentic AI is only as reliable as the product rules and context it can access. Governance comes first. Carriers should set decision limits and rehearse them with assisted underwriting before granting autonomy.
The report lays out a four-phase path: foundation, pilot, selective autonomy, and scaled deployment. Each phase has a success test. Underwriting leaders can use it to decide which submissions AI handles first and which stay with underwriters.
Clients of Datos Insights’ Property & Casualty practice can download this report.
This report responds to questions including:
- How can carriers adopt agentic AI in underwriting safely? Start with one source for product definitions and clear decision limits. Then move through pilot, selective autonomy, and scaled deployment.
- Who is accountable when agentic AI makes an underwriting decision? Carriers set decision authority, escalation paths, and traceability before AI gets any autonomy, so each decision can be shown and explained.
- Why must underwriting and policy systems use the same product rules? When they differ, a policy can be issued under different terms than the ones used to underwrite the risk.
About the Author
