Historic approvals make Z-FIRE™ the first AI model to obtain regulatory approval for assessing and pricing wildfire risk at the individual property level
Jul 7, 2021 | ?min Read
Today, ZestyAI, the leader in climate risk analytics powered by Artificial Intelligence (AI), announced that the California Department of Insurance (CDI) has approved filings using its Z-FIRE™ model for both underwriting and rating, making it the first AI model for wildfire risk assessment to obtain regulatory approval in the state. The CDI first approved a Metropolitan Direct filing for underwriting and rating, and shortly thereafter approved a Farmers Insurance filing for underwriting. Together, the approvals make tens of thousands of homes eligible for coverage that legacy models deemed too high-risk.
Michael Soller, a deputy commissioner with California’s Department of Insurance told The Wall Street Journal that the state is seeking to ensure algorithms used for underwriting and rating don’t introduce or perpetuate bias and that he sees Farmers’ adoption of Z-FIRE™ as a positive development for the state’s property insurance market: “With Farmers’ promise to write at least 30,000 new policies for homeowners with higher wildfire risk using this new tool, we are glad to see the pendulum swinging in the other direction,” Soller said.
Z-FIRE™ is an AI model trained on more than 1,400 wildfire events across more than 20 years of historical loss data. It considers property-level features that influence risk, such as topography, historical climate data, and critically, factors extracted from high-resolution imagery like building materials and surrounding vegetation in multiple defensible spaces. This allows carriers to reward homeowners for mitigation efforts, which aligns with the proposed regulations announced in February by the CDI to create incentives for mitigation efforts and make coverage more affordable for homeowners that invest in prevention measures like vegetation clearance and fire-resistant roof materials.
“The unfortunate increase in the frequency and severity of wildfires in California has forced the insurance industry to explore new solutions for assessing risk, and the CDI’s approval of these filings is a profound step forward for carriers and property owners alike,” said Attila Toth, CEO of ZestyAI. “Now carriers can evaluate how something like overhanging vegetation impacts the overall fire risk to an individual property, and through that level of granular analysis they can offer coverage for buildings that were previously considered uninsurable. This enables carriers to continue doing business in California while providing more options for property owners for coverage and rewarding them for mitigation efforts, which is so important for reducing the overall fire risk to communities across the state.”
In California, where the California Association of Realtors lists the median home value at around $699,000, roughly $700Bn worth of homes are at risk of losing their mortgages driven, in part, by poor accuracy from legacy wildfire models. Z-FIRE™ enables carriers to reverse some of that risk by providing property specific risk scores with surgical precision. It has a proven ability to decrease loss exposure for insurance carriers with early adopters reporting a significant decrease in wildfire risk exposure while enabling them to offer insurance coverage even in regions considered high risk for wildfires while also offering a much greater level of transparency around scores and mitigation. Compared to legacy models that paint large regions equaling 15% of all properties in the state as high risk, Z-FIRE™ estimates 7% of properties in California to have a true high wildfire risk exposure. This difference enables higher availability of insurance and affordable premiums commensurate with the true risk of wildfire for property owners.
For more information about Z-FIRE™, click here.