Gen AI can solve property cat risks: DGTAL

Generative artificial intelligence (Gen-AI) can be a powerful tool with the potential to solve some of the challenges the industry is facing in pricing and managing the risk associated with so-called secondary perils including severe convective storms (SCS).

That is the message from Vanda Giannara, the co-founder and chief client officer of AI insurtech DGTAL, who told APCIA Today that Gen-AI could be used to capture data around specific property-cat risks in a way that can help insurers get a much better handle on the peril.

Risks of this nature are challenging to insurers for a number of reasons. Losses are increasingly exponentially, potentially driven by climate change, but they tend to fall below reinsurance attachment points for a single loss, meaning they are largely absorbed by insurers.

Gen-AI—which, Giannara stresses, is very different from predictive AI—can enable the industry to classify and price the risks much more quickly and accurately, she said. 

“Insurers already have the relevant data to do this—but they cannot use it to make better-informed decisions,” she added. 

DGTAL has developed an AI application to identify and capture the emerging risks. Last December it launched a claims assessment application, DRILLER, which enables re/insurers to capture unstructured data related to claims. 

Speedy analysis 

In the case of property claims, Giannara offers a specific example. In the aftermath of an event such as Hurricane Helene, she argues, insurers already have access to very specific and key information that can tell them much about the potential loss.

“The scale of claims in the aftermath of an event come down to four simple things: the quality and specs of doors, windows, garage doors and roofs. Insurers have this information, but it exists in an unstructured format,” she explained. 

“Using data in this way can benefit everyone—insurers and their customers alike.”

“Gen-AI can help translate this data quickly into very specific information, which can determine the size of a claim and then be used for better, customer-specific underwriting in the future.

“It can drill down into exposures and give you much more detail of the true risks, a much better cohort. On that basis, it can incentivise homeowners to fortify their homes if their insurance premiums are customised to their risk profile,” she said. 

DRILLER is an AI-powered tool designed to develop and display a contextual understanding across a wide array of cases and documents. It allows users to sift through massive volumes of data to identify patterns, trends, and insights. 

It is particularly powerful at analysing unstructured documentation which could take weeks or months to be assessed manually, says DGTAL. Instead, it can do it in minutes.

Giannara argues that there is no reason insurers should not use such information to make much better-informed underwriting decisions. This already happens in lines such as motor insurance, she says, where better drivers are rewarded with lower premiums.

She added that linking such details to a purchase is completely normal in other areas such as retail. “You buy a phone or a holiday—you expect the marketing, the sale, to be based on your personal profile,” she said.

“The problem in insurance is we do not do that. We tend to focus on the risk as opposed to the client. But policyholders are also consumers. Using data in this way can benefit everyone—insurers and their customers alike.”

Identifying risks

DRILLER is able to analyse packages of up to 500 claims documents in just over 30 minutes. It would take a person a week or more to review the same claims packages. The company is now looking at how Gen-AI can be used to identify emerging risks, including property risks. 

“Gen-AI is a very competent tool for language understanding, for reading files, and processing and giving you this information,” Giannara said. 

Insurers often have the information they need to make better underwriting decisions, but it is often hard to access and process enough information. 

“Someone needs to read all of the claims files to get an understanding of what is going on, but a pricing actuary cannot do that alone,” she explained.

“AI is able to do things faster. This information already exists, it is just a question of having the tools to understand it.

“DRILLER can also quickly reveal the data needed to unlock insurance-linked securities trapped capital,” she concluded.

Vanda Giannara is the co-founder and chief client officer of DGTAL. She can be contacted at: v.giannara@dgtal.io 

For more news from the American Property Casualty Insurance Association (APCIA) click here.

Did you get value from this story?  Sign up to our free daily newsletters and get stories like this sent straight to your inbox.