Hail damage – how technology and data insights can improve property risk

“The first episode [to hail losses] and climate change. “We’re seeing a lot of storms that are very dangerous,” said Randy Ishikawa, director of home insurance at LexisNexis Risk Solutions. Rising commodity prices and limited availability due to disruptions in insurance coverage add to the headlines for insurance.

Read more: Catastrophic weather accounted for 39% of all home insurance claims in 2020 – report

The so-called “snow highway” states, such as Wyoming, South Dakota, Texas, and Oklahoma, see the most snowfall each year, so property in those areas is considered the most dangerous. But location is only one factor: the location of a building also affects its risk.

“New houses and roofs will not be damaged depending on the size of the snow. Pitched roofs perform better than flat roofs, for example, although they are more expensive to replace,” Ishikawa told Insurance Business.

“Most of the time, hail is less than an inch long and shouldn’t damage a new roof. But you will probably see damage with an old roof with shingles that may be curled. “

Good roof inspection

Accurate inspections are critical to finding hazard information, helping carriers avoid damaged roofs on their books. But the roof inspection itself, called a drive-and-look, gives a limited idea of ​​roofs and potential damage.

“The inspectors drive a car to check the condition of the roof. But they only see the front side of the roof, not the back. They cannot see the top of the building or the roof itself. “That’s where aerial photography comes in,” said Ishikawa.

In order to improve rooftop inspections, LexisNexis developed a tool called Rooftop that creates a risk assessment based on data, including image analysis, historical data, and weather data.

“By taking a panoramic or ground-level image of a building and roof and extracting information from it through artificial intelligence and machine learning, [Rooftop] they give the insurer a percentage of the loss.

“You take the geospatial data, and then you analyze the climate: how many storms have occurred in this area? When did they occur? How big was the storm?” Ishikawa asked.

Using data information

Photographic researchers consider the design of a building, including the shape of the roof and siding, the materials used, and other features. But other types of data can provide clues about pre-existing roof damage.

Ishikawa explained: “If a storm happened in a certain area and the house did not come, but all the neighboring houses refused, how is it possible that the house has a damaged roof that the owner of the house does not know about?” When another storm comes, it destroys [could] to be very old. Technology can help insurance agents reduce risk. “

Claim methods show information about snow exposure. Ishikawa said they focus on automotive product categories, for example, because they tend to grow faster than predicted during severe weather.

“We take all of this into account to get an interest rate from one to 100. The higher the number, the higher the interest rate at that location,” Ishikawa summarized. Each carrier will have a different method and result, depending on their risk appetite and writing method. But with accurate data, carriers can better prepare themselves to protect against the dangers of snow.

“Most carriers look at something and say: ‘We can afford that.’ [risk]. We can pull the underwriting lever and give them cheap money to help reduce risk, or we can give the real value of the money through the roof and not all the extra money.’ It’s enough for an insurance agent,” Ishikawa continued.

Snow season in the US starts in May and ends in late August. For Ishikawa, it’s time for the carriers to start preparing for next season.

“Right now, most carriers can’t do much about the risks in their books. But they need to look at their books, especially in areas that have seen a lot of snowfall, so they can reduce the risk before the snow season,” Ishikawa said. .