3 mins read

FT Insurance Innovation Summit 2018: Cytora CEO discusses future of AI underwriting

The summit brought together insurance executives and industry experts to discuss the best strategies to adapt and innovate. Moderated by insurance correspondent Oliver Ralph, Hartley was joined on the panel by Satadru Sengupta, Insurance General Manager at DataRobot, Cindy Forbes, Chief Analytics Officer at Manulife, and Gina Papush, Global Chief Data and Analytics Officer at QBE Insurance.

Hartley explained how AI is already being used by commercial insurers, including those working with Cytora, to deliver questionless underwriting – the ability to rate and price a risk purely with external data with no questions asked of the insured.

AI also improves risk selection and pricing decisions in terms of setting a more accurate price to cover a claim. This enables insurers to make risk selection with a very deep, high-dimensional view of the risk, and avoid premium leakage, which can often happen when you rely on information from a single source.

“Customers don’t want to spend a lot of time buying insurance. They want a fast, accurate quote. Using AI, we make insurance easier to price and easier to purchase by removing friction from the process.”

— Richard Hartley, CEO and Co-Founder, Cytora

In larger commercial lines, the benefits of AI currently centre more around augmenting the underwriter, providing risk scores and prices to make a better decision that reflects the combination of machine intelligence plus underwriting judgment.

Gina Papush said that QBE, which works with Cytora in the UK and Australia, has found parts of large and complex business that can be automated. In these cases, AI can inform accurate decision making, or eliminate the need for valuable underwriting resources to spend time investigating and reviewing every submission.

Satadru Sengupta added that the goal needn’t be 100% automation; instead, it is valuable for an insurer to quickly know which submissions can be automated, which submissions can be rejected outright, and which to investigate further.