Survey: 83% of insurers would let AI handle repeatable work

Only 6% of respondents said they would trust a general-purpose AI model alone for high-stakes underwriting or claims decisions.

A study from ISG, commissioned by mea Platform, found 83% of insurance respondents would let AI handle repeatable operational tasks, and 75% would only allow that through an insurance-specific or governed model. Just 6% said they would trust a general-purpose AI model alone for high-consequence underwriting or claims decisions.

The source's own statement, as linked above. Nothing beyond it has been independently checked by us.

Sample size not disclosed. This describes stated attitudes, not measured production deployments.

Expect carrier AI rollouts to concentrate on document intake, triage and simple rating rather than final underwriting or claims calls. Agencies should ask which specific tasks a carrier's AI touches before assuming a submission goes untouched by a human.

AIUC raises $40M to build audits and insurance for AI systems