How AI forces underwriters to rethink data handling and staffing

Takeaways:

• AI elevates junior underwriting personnel
• Insurer AXA XL looks for where human review is needed
• Evaluating data using AI enhances risk insights

Applying AI to underwriting enhances staff capabilities and improves data evaluation for underwriting decisions, according to experts in insurance risk management.

Junior underwriters can gain confidence by using AI platforms to analyze risk data and make recommendations accordingly, said Juan de Castro, president of Cytora, a risk digitization and automation platform provider. De Castro spoke in a recent webcast hosted by Insurtech Insights.

"It allows them to grow and develop much faster. We're going to see accelerated careers in underwriting," he said.

These underwriters' careers will shift toward portfolio analysis, according to de Castro. 

"Some of those people are not going to be looking at individual pieces of work, individual risks," he said. "They're going to move into the control room, like a control room of a factory. They're going to do the portfolio analysis, thinking about how to optimize the book for reinsurance."

When considering how to implement AI in underwriting, AXA XL, a P&C and specialty risk insurer, asks its clients about how employees are trained, what data is needed and where human review is necessary, according to Caitlin Alpern, team lead for large cyber and technology accounts at AXA XL. Alpern spoke in a recent webcast hosted by Newfront, an insurance brokerage.

"There's not a lot of consistency yet to how different carriers are approaching this," she said. AXA XL also asks its clients about their AI use policies, compliance, choices of AI models and data, according to Alpern. 

AI can enable closer scrutiny of data used in underwriting, according to Jennilee Foo-Kune, head of underwriting strategy and performance for business solutions at Scor, an insurance risk management company. Foo-Kune spoke in the Insurtech Insights webcast.

"Skepticism is super essential, still," she said. "Certain elements will find a hard time automating or replacing. That's the human and the relationship elements."

However, underwriters can get more time understanding risk as AI evaluates more of the data, Foo-Kune added. "That's how we're going to evolve, and I don't necessarily think that people are going to get replaced," she said.

Insurance underwriters can no longer afford to take five days to produce a quote, said de Castro. 

"The rules of the game have changed, and we are already seeing those who have made progress using AI are accelerating growth, and those who haven't are actually stagnating," he said. "In a soft market, you have to grow through policy count, not through rates. In the financial results in a year's time, the difference is going to be very clear."

While using AI to automate handling of cyber risks can be challenging, AI helps insurance underwriters access the best risks from distribution partners, according to de Castro. 

"When a broker is working on a risk, how do you signal to the broker, 'Hey, that is a risk I would like to quote, it's in my appetite,'" he said. "That is already happening. Those who are not engaged in that, somebody else is going to take the risks for them."