Insurtech 2.0: Rooted in data management

This article is the second of a three-part series on aspects of “Insurtech 2.0.”

AI is likely to play a big part in “Insurtech 2.0,” the buzzword for the latest transformative insurtech advances, but experts think changes will be incremental and must be rooted in the right functions, like data management.

Ellen Carney, principal analyst at Forrester Research.

The best use of AI, new technology and new ideas is managing the large amounts of data in insurance industry operations, according to Ellen Carney, principal analyst at Forrester Research. Gen AI may get a lot of attention, but data issues soon will also, she added.

“You could see insurtechs that can tackle data quality, data optimization, normalization, synthetic data,” she said. “All these things used to train models, are going to get a lot of attention.”

The Insurtech 2.0 evolution will produce value and reduce costs in insurance business models, according to Bill Pieroni, president and CEO of ACORD, a standards body for the global insurance industry. Customer service functions, particularly chatbots, that insurtechs developed, are now commoditized and not where the true innovation is, Pieroni added. AI can and will transform insurtech for other functions, though, he said.

Bill Pieroni, president and CEO of ACORD.

“AI is going to transform this industry in such a powerful way that it will have one of the largest impacts we’ve had from anything that has ever occurred,” Pieroni said. “However, there are a couple of things. One, as in most situations, we will far overestimate its impact in the near term. And underestimate the transformative impact.”

Part of the hype of AI comes from misunderstanding its nature, which is simply applied mathematics, only with greater computing power, as Pieroni explained. The concept of AI has been around since at least the 1950s and AI systems were being built in the 1980s, he continued. The difference then was only organizations with millions of dollars to spend on the necessary processing power could do it. Now the same processing power can be had for $1,000 or less.

The deep learning subset, within the machine learning part of AI has practical applications for the insurance industry, according to Pieroni. “Deep learning has neural networks where you get back-propagation,” he said. “It’s learning back to those billions of parameters and zettabytes of data [that are now available].”

Darryl Siry, chief operating and technology officer at Cover Whale.

Typically, insurers only see one or two percentage points improvement in value from adopting new technology, according to Darryl Siry, chief operating and technology officer at Cover Whale, a commercial auto insurtech company. Applying AI, including Gen AI, to data management can increase value as much as 10%, Siry said. “That’s an extraordinary advantage,” he said.

With that potential, in five to seven years, every insurer will be using some form of Gen AI, including for claims and underwriting functions, according to Siry.

Aside from doing more with data, insurers can apply Gen AI to internal communication, according to Steve Anderson, CEO and co-founder of Catalyit, a technology services provider for independent agents. “Communication between the carrier, the distribution system, the agency, and the agency’s technology, then becomes a smoother path,” he said. “Now that would be around open access and APIs and things like that.”

Steve Anderson, co-founder and CEO of Catalyit.

For insurtechs that are part of Insurtech 2.0 to last, they must address pain points, said Anderson. “If the agents don’t perceive it as a big enough problem, then the solution might be fun, or it might be a new and shiny object, but it might not be worth their time and investment to actually get it, start using it or start figuring out how to maximize it,” he said.

Also, as Pieroni said, insurtechs should accommodate insurers’ preferences for incremental change. “The reality is, for major customers to use an insurtech’s product, they’re not going to throw away 20 to 30 years worth of committed legacy,” he said. “You’re going to have to incrementally tweak and bolt on to that legacy.”

Next: How experts think Insurtech 2.0 should be applied to claims and underwriting functions.