Insurers leverage AI, tech for wildfire risk analysis

Ash and debris covers the site of a property destroyed by the Dixie Fire in Greenville, California, on Aug. 10, 2021. Fanned by gusting winds and low humidity, the Dixie Fire, now the second largest wildfire in California’s history, grew by nearly 6,000 acres overnight Monday, reported the San Francisco Chronicle.
Ash and debris covers the site of a property destroyed by the Dixie Fire in Greenville, California, on Aug. 10, 2021. Fanned by gusting winds and low humidity, the Dixie Fire, now the second largest wildfire in California’s history, grew by nearly 6,000 acres overnight Monday, reported the San Francisco Chronicle.
Maranie Staab/Bloomberg

According to the California Air Resources Board, wildfires have been increasing in severity and frequency since the 1950s due to climate change, which has led to greater hesitance by insurers to insure some properties in California.

While climate change leads to greater environmental hazards, technology could open some of those properties back up to be insured. Looking forward, property mitigation like clearing out brush and using different roof styles in construction, and artificial intelligence will be able to flag some properties as lower-risk.

The following insurance and tech companies have begun using AI and other technologies to better map what properties are high-risk and what properties could be eligible for insurance.

  • Delos: Home insurer, uses geospatial data and AI to account for different factors, like vegetation or wind, to better predict what properties will be hit the hardest by wildfires.
  • Verisk: Data analytics and risk assessment firm, uses its Fireline technology to assign risk scores to different locations. Data are updated frequently, as parts of a region that may have been high-risk years ago, could now have mitigation-- like cleared out growth-- that marks it a lower risk.
  • Farmers Insurance: Auto, home and small business insurance provider, partnered with Zesty.ai so that Farmers could better analyze the risks associated with individual properties and insure more homes.
  • Zesty.ai: Building analytics platform that uses AI to better understand the risk associated with a given property. The risk is determined based on materials, location and other factors. Zesty allows Farmers to measure factors like vegetation, roof type or construction materials.
  • Insurity: Software and data analytics provider, Insurity’s Wildfire Threat Analysis generates a burn perimeter 15 to 30 minutes after a wildfire is named through combining data from the National Interagency Fire Center and NASA Visible Infrared Imaging Radiometer Suite.

Kevin Stein, CEO and co-founder of Delos, said the 2018 Sierra Nevada wildfires, which used the dense forest as fuel and wind, are a good example of the need for modeling. For other parts of California, the wind would be a major factor in the spread of a fire, but had the model just relied on wind directions to predict the fire, it would have missed any spread in Sierra Nevada.

Models need to be location-specific with specific indices for accurate predictions, Stein said.

Tom Larsen, principal of insurance and spatial solution at CoreLogic, a global property data and solution provider, said farms and wineries, for example, are high-risk properties and difficult to insure. Other structures on a property and potential crop loss are both considered potential risk factors, Larsen said.

"Obviously, commercial businesses are facing a really difficult situation where not many [insurance] groups have a great understanding of wildfire exposure, so they can't really insure these consumers," Larsen said. "A farm is an occupancy, a very high-tech occupancy because of the processing equipment. …So, as a farmer, you have risks."

Other companies are working to predict those risks and potential loss. Insurity’s technology allows for insurers to more quickly respond to fires near higher-risk areas, like farms, which could lower the cost of claims, according to Kirstin Marr, head of Insurity Analytics.

Earlier this summer, Farmers Insurance and Zesty.ai began a partnership to analyze the risks associated with individual properties.

“Using Zesty in our decision tree will allow us to get down to a property level, that will open up the market for us,” said Keith Daly, president of personal lines for Farmers Insurance, to Digital Insurance in June 2021. “This is about adding [insurable areas] to the pie. Most carriers in the state of California, over the last decade, have done far more restricting, us included. And now we can open up the market."

The insurance industry is going through a tough time as it learns how to respond to wildfires, according to Stein. He cited people having trouble finding insurance after Hurricane Andrew and the Northridge Earthquake in the '90s as early examples of how the industry reacts to natural disasters.

"It doesn't help people right now, but there will be groups that will fill the need of consumers and allow life to go a little bit back to, not normal, given the catastrophes, but normal for insurance,” Stein said.

Correction
This article has been updated from its original version for clarity on the role of the wind in the 2018 Sierra Nevada fire.
August 19, 2021 10:08 AM EDT
For reprint and licensing requests for this article, click here.
Climate change Risk analysis Natural disasters California Insurtech Artificial intelligence Data Analytics
MORE FROM DIGITAL INSURANCE