ZestyAI, a San Francisco-based InsurTech offering a risk decision platform to insurers, has secured regulatory acceptance in Minnesota and Maryland for Z-WATER, its AI model for non-weather water risk, bringing the number of states where it can be used for rating and underwriting to 22.

With the latest approvals, carriers in both states can now draw on Z-WATER within their rate and rule filings. The model’s acceptance now extends past 20 states, widening the scope for insurers to build property-specific risk data into how they set prices.

Non-weather water now ranks as the fourth most expensive peril in homeowners cover, driving more than $15bn in yearly losses spread across over one million claims, ZestyAI said. Rather than stemming from extreme weather, these claims typically arise from routine problems such as burst pipes, ageing plumbing and faulty appliances. Conventional rating methods, which lean largely on location-based territories and the age of a property, struggle to pinpoint which specific homes carry the greatest exposure.

To close that gap, Z-WATER draws together computer vision, aerial imagery, property attributes, permit records, local climate data and infrastructure information to forecast how often non-weather water claims will occur and how costly they are likely to be. The model has been trained and tested against insurer loss data, and ZestyAI says it achieves 18 times greater risk segmentation than territory- and age-based methods, helping carriers separate higher and lower risks within their books more effectively.

The new state approvals add to wider regulatory progress for the company.

Source link