When insurance companies offer quotes on homeowner’s insurance, they naturally want to know as much as they can about the property involved.
Traditionally, that’s meant sending insurance applicants sometimes lengthy questionnaires and dispatching inspectors to fill in any missing details. But recently, a Bay Area startup called Cape Analytics has been using machine learning and geospatial imagery to automatically pull out data points—like building geometry, roof condition information, and nearby tree cover—that insurers can use to evaluate risk.
“They’re seeing a much more streamlined process where they’re able to offer a quote more quickly and more accurately to the customer,” says cofounder and CEO Ryan Kottenstette.
The roughly three-year-old company, which just announced a $17 million Series B round of funding, said earlier this spring that it can provide data nationwide on about 70 million homes. It got its start in the Southeast, where wind and hurricane risk make insurers particularly concerned about risk factors like what roofs are made of and what condition they’re in. Cape Analytics takes in satellite and aerial imagery from third-party providers, then uses neural networks to parse out features of interest to insurers, sometimes with more accuracy than homeowners can describe their own properties.
“Most consumers don’t know what a mansard roof might be, for example,” says Kottenstette.
Insurers already regularly use aerial imagery to understand damage from natural disasters, and other companies have found innovative ways to pair such photos with artificial intelligence, from spotting houses that could benefit from solar panels to tracking oil storage for a glimpse at the global market. But Kottenstette says his company helps insurance companies learn about features specifically useful to them, without having to devote human effort to spotting roof tiles and potentially damaging tree limbs in individual images.
“You can’t sit there and have humans consistently reviewing pictures of 70 million homes in the event that somebody wants to get an insurance quote,” he says.
At a time when big tech companies like Google have come under fire for helping the government parse drone imagery, Cape Analytics is careful to take privacy considerations into account, Kottenstette says, emphasizing the company doesn’t generate any images of its own that weren’t already available on the market.
Cape Analytics is considering expanding into other markets around the world and expanding the range of information it can provide to insurance companies. Right now, the existing data could also be useful to insurers looking to suggest potential improvements to their customers’ properties, like repairing roofs or trimming overhanging trees.