MGAs should tread lightly with AI, underwriter says

Diverse business team in modern office discussing AI with whiteboard.
HudHudPro - stock.adobe.com

Takeaways:

Processing Content
  • Insurers and underwriters must communicate about risk data
  • Insurtechs should understand underwriting challenges
  • AI is still an 'unknown' for insurers

Managing general agents contend with high volumes of data as they work on underwriting risks, according to an underwriting executive at a global MGA firm.

Lisa Rowe of Specialty MGA
Lisa Rowe, senior underwriter for financial lines and cyber at Specialty MGA.
LinkedIn

Lisa Rowe, senior underwriter for financial lines and cyber at Specialty MGA, described the work necessary to verify and evaluate risk data, in a webcast hosted by InsTech on February 19. Specialty MGA is a London-based insurance and reinsurance company with offices in Austin, Texas; Milan and Casablanca.

Clear communication between insurers and underwriters is necessary to explain and verify risk data, Rowe said. Insurers, reinsurers and MGAs work hard to get data current and correct, and without their efforts, evaluation output would not be reliable and consistent, she added. Clean data with the correct metrics is also necessary, Rowe said.

Specialty MGA looks for insurtech service providers who understand underwriting challenges, according to Rowe. 

"From a system use case standpoint, we're looking for something that's intuitive, that's easy to use and fast," she said. "Speed, intuition, and not being overly complicated to use. Something that makes my life easier, simply put."

Whatever the system, underwriting documents that Specialty MGA reviews can be as long as 2,300 pages, according to Rowe. "It's a data problem, in the sense of the sheer volume of data that we have to ingest in order to make an underwriting assessment on the risk," she said.

AI can help, but has to be administered carefully, as Rowe explained. APIs and low-code platforms can sometimes be sufficient to manage data and information for underwriting risk, she added. 

"AI is still somewhat of an unknown entity," Rowe said. "We need to move with the times and get abreast of the challenges that AI poses from an insurance perspective, in terms of providing coverage for clients across the market. These are all new vectors than we have previously had to deal with in the insurance market. Getting ourselves ready for these and learning how we can still support our clients whilst managing risk is going to be a focus for us."

For reprint and licensing requests for this article, click here.
Artificial intelligence Underwriting Agents Property and casualty insurance
MORE FROM DIGITAL INSURANCE