After several years of experimentation with generative AI and machine learning, many carriers have moved beyond asking whether AI has a role in insurance. Instead, they’re grappling with a different question: How do organizations know when an AI pilot is ready to scale?

Industry leaders say the answer has less to do with technological sophistication and more to do with organizational readiness.

“It’s how they talk about AI inside of their business,” said James Thom, chief product officer at Vertafore, during a panel discussion at Carrier Management’s May InsurTech Summit. “If they’re talking about the outcomes that they’re driving toward, the expectation of what the impact is going to be, the change on the processes inside of the business, that’s when you know that they’re ready to scale. If they’re still talking about it from a pure technology perspective or the models that they’re interested in or the theory and concepts behind AI, you know that they’re a long way off.”

Thom said that distinction is critical because insurers can easily become distracted by technically impressive solutions that deliver limited value. Instead, successful scaling means focusing on solutions for meaningful problems.

“A lot of times I see carriers, MGAs, agencies, anybody in insurance solving what I would call interesting problems rather than important problems,” he said. “They’re focused on something that seems like they could do it rather than they should do it.”

The insurers making this transition successfully tend to share several characteristics, he said. They focus on business outcomes rather than technology, embed AI into core workflows instead of layering it on top of existing processes, and establish trust through human oversight.

“It has to be the core of your process,” Thom said. “You can’t really bolt it on.”

However, he cautioned that building AI solutions internally means insurers have to be strategic about keeping up with the pace of change in the industry.

“I mean, you have to be able to shift your infrastructure rapidly to deal with the moving target that is AI,” he said. “You have to be careful about whether you are ready to take on that infrastructure for yourself in order to keep moving and keep up with that pace of where the market’s going.”

He added that technology adoption is as much a change-management exercise as it is a technical one.

“We have a lot of team members out there inside of carriers that are change resistant, and so you have to be thoughtful in terms of how you make it easy to use those tools by having it really as an integrated part of the process,” he said.

This emphasis on workflow integration was echoed by William Steenbergen, chief technology officer at Federato. He said insurers should begin by clearly defining the role AI will play in decision-making.

“It’s ultimately about just defining what are the guidelines that we want to use AI with,” Steenbergen said. “Are we willing and are we allowing AI to make actual decisions, or are we using AI as a tool that underwriters can use to get things done?”

For many carriers, it’s the latter. AI is primarily a tool used to augment current decision-making processes, he said.

“Underwriters are actually reviewing decisions AI agents make, and that’s where they’re applying their true actual underwriting skill versus trying to build from scratch,” Steenbergen said. “What you really need is a platform and a tool that allows underwriters to do a really good job at reviewing what AI agents are doing. And if they review less and less, and you have the data around that, you know you’re building that trust. Gathering that data is extremely powerful.”

The speed of acceleration is creating challenges for the industry. But Craig Weber, head of insurance strategy at Cognizant, emphasized that while AI adoption in insurance remains a work in progress, it is ultimately headed in a positive direction.

“It’s shocking if you think about where we’ve come from in the last couple of years,” he said. “I know even within our firm, with all of our clients, what we thought we knew about AI two years ago has just ramped up incredibly quickly. So, this is not a failure story. It’s really more of a work in progress story with some bumps along the way.”

The industry’s complexity is one of those bumps.

“It just turns out that we are a complicated ecosystem,” Weber said. “Every insurer has a set of systems that are trying desperately to work together, and putting AI in one of those and solving just a portion of that problem creates some probably unintended downstream impact. So, thinking that through is really an issue.”

Regulatory considerations add another layer of complexity.

“As a regulated industry, we have a lot of issues to sort out before we can turn those AI agents loose and let them do their thing,” he said.

That said, Weber warned that insurers face risks on both sides of the equation—moving too quickly without adequate governance or waiting so long that they miss opportunities competitors are already pursuing.

“I would say the bigger risk is simply not recognizing the moment that we are all in,” he said. “In my career—and I’ve been around insurance my whole career—I’ve never seen insurers do anything too quickly. The clock speed has shifted really dramatically faster in the past five years…so the head in the sand approach is not really going to suffice.”

Steenbergen agreed that insurers will need to keep up with the pace of technological change or risk falling behind.

“The world is changing very, very rapidly—extremely rapidly—around AI,” he said. “If you don’t adopt now and you don’t get on the train…you’re going to lose an ROI you didn’t even know existed. Are you getting yourself ready for the amount of innovation that is already happening and is going to continue happening over the next couple of years? Because if you’re not, you’re going to be very, very far behind.”