Moody's finds AI in core work at two-thirds of North American insurers

Moody's Ratings has found that about two-thirds of 58 North American insurers, reinsurers and brokers use AI broadly or across several core operations, but fewer than half consistently track each use case's value.

The survey, reported by Insurance Business on 18 October, covered life, health, property and casualty, reinsurance and broking companies. About 34% of respondents described their deployment as broad and enterprise-wide, while the remaining third reported more limited use.

Only 41% said they quantified and tracked benefits for individual use cases. A further 34% quantified benefits without tracking them consistently, while 24% said the benefits were mainly qualitative.

Generative AI was the most widely used form of the technology, reported by 98% of respondents. Traditional machine learning or predictive tools were used by 79%, while 71% used agentic AI or workflow automation.

Among property and casualty insurers and reinsurers, 75% had AI in production in operations. More than half had put it into production for underwriting and pricing. Half had done so in claims and loss adjustment, and half in fraud detection. Catastrophe and exposure management stood at 19%.

Brokers reported a similar spread of maturity. Some 29% said deployment was broad, 43% had several production uses and 28% described their use as limited.

Productivity gains or avoided recruitment were cited as benefits by 76% of respondents, while 74% reported faster processing or cycle times. AI typically accounted for 2% to 10% of IT budgets. Some 13% said the share was higher than 10%.

Skills and talent shortages were the most common barrier, cited by 81%. Data quality followed at 71%, legacy system integration at 62% and regulatory uncertainty at 47%. Despite those concerns, 88% said their governance aligned with regulatory guidance.

Moody's expects AI's credit effect to range from neutral to positive, depending on execution, data quality and governance, according to Insurance Business. It also sees higher operational, regulatory and litigation risk, alongside greater dependence on important vendors.

The rating agency warned that advances in frontier AI could raise insurers' cyber exposure by accelerating the exploitation of software vulnerabilities. It cited an April 2026 example in which Anthropic's Mythos model found high-severity vulnerabilities in most major operating systems and web browsers.

Moody's said larger insurers are better placed, while smaller firms may struggle to keep pace.

Related coverage: Insurify opens its marketplace to personal AI agents with an MCP gateway and Opterrix launches AI tool for portfolio exposure and accumulation.

Sources

• Insurance Business: AI now mainstream at North American insurers, but few can prove the payoff
• PropertyCasualty360: Moody's, AI moving into core insurance operations

CoverWire is a news and information service. Nothing here is investment advice.