Adaptive Insurance adopts ZestyAI storm model for hail and wind pricing

Adaptive Insurance has put ZestyAI’s Z-STORM model directly into its wind and hail underwriting and pricing workflow, a clear sign that peril-specific AI is moving from analytics talk into the mechanics of core underwriting. The move matters because severe convective storms have produced more than $50 billion in annual insured losses for three straight years, and broad ZIP code assumptions are no longer enough to separate good accounts from the ones most likely to blow up loss ratios.

Adaptive, founded in 2024 and based in Austin, Texas, said the integration is aimed at sharpening how it prices, selects, and issues coverage in storm-prone business. The company’s platform already uses historical and real-time weather and climate data to identify risk and guide pricing, and it says agents can quote, bind, and issue a policy online in less than five minutes. Adding Z-STORM gives the carrier a property-level view of hail and wind exposure, which is the kind of segmentation underwriters have been asking for as storm losses become more volatile from one neighborhood to the next.

That is where the software questions get interesting. If Z-STORM is influencing eligibility as well as rating, Adaptive is no longer just using a model to inform a renewal conversation. It is embedding third-party risk scores into the quote path itself, which raises the stakes for governance, auditability, and how a carrier explains a decline or a higher rate to regulators and agents. ZestyAI says that is exactly the point: the company argues its models help insurers “build smarter, more resilient systems” using real-world data, and it has pushed explainable, mitigation-aware scoring as a way to reflect verified roof upgrades or corrected property data.

The regulatory backdrop gives the deal more weight. ZestyAI said its severe convective storm models have now been accepted in 32 states, up from 29 before additional reviews in Maryland and Nevada. The company also said its broader portfolio has more than 200 regulatory approvals, a signal that carriers are looking for model support they can defend in rate and rule filings, not just attractive loss maps. For Adaptive, the bet is straightforward: if property-level intelligence can improve where it writes, how it segments, and how fast it quotes, then AI stops being a dashboard feature and starts acting like underwriting infrastructure.