In the insurance world, adopting artificial intelligence isn’t a sprint to the finish—it’s a scramble just to get to the starting line.
“This is more of a race to start than a race to the end,” said Tom Wilde, CEO of Indico Data. “Getting your data ready so AI can actually do something useful with it—that’s going to be the big differentiator in the next decade.”
AI in insurance—whether we’re talking agentic systems or generative tech—is still in its early stages. But Wilde sees it fundamentally changing how insurance decisions are made, from underwriting risks to handling claims.
“At the core, insurance is all about decision-making,” Wilde explained. “Should we take on this risk? How do we resolve this claim? We’re calling this the beginning of the ‘decision era.’ With the rise of cloud computing, smarter data strategies, and AI, we now have the chance to make faster, more consistent decisions across the board—especially in insurance.”
One of the major shifts Wilde is seeing? The move from relying on individual experts to building institutional expertise with AI.
“Carriers are realizing that depending solely on human expertise can only take them so far,” he said. “AI lets us scale that knowledge. It brings repeatability, transparency, and consistency to decisions. That’s a major leap forward.”
A big part of that leap comes from being able to extract value from unstructured data—something insurers have always had in abundance.
“Say you’ve got a 25-page underwriting guideline document,” Wilde said. “Now, instead of someone having to read it and figure out what to do manually, we can turn it into a live, usable data source. It becomes part of an automated workflow. That’s a huge time-saver.”
Still, even with AI becoming more embedded in underwriting and claims, humans aren’t out of the picture.
Michael Parcelli, SVP at Xceedance, emphasized that real transformation comes with balance—letting AI handle the bulk of tasks while humans step in for quality control.
“We’re not replacing humans entirely,” Parcelli said. “There will always be exceptions, auditing, and oversight. But where a person used to do 80% of the work and 20% reviewing, now it can flip—let the tools handle 80% and humans step in for the final 20%.”
This shift allows insurers to become more efficient while keeping fairness and accuracy in check.
“AI still needs human review,” he added. “It’s not infallible. Sometimes the information it gives is off or misrepresented. You need a process to refine and correct it. That’s why you’ll always need human involvement.”
Wilde echoed the need for transparency as insurers adopt more autonomous or “agentic” AI systems that make and learn from decisions on their own.
“There’s a lot of complexity behind an AI-driven decision,” he said. “Which model was used? What version? Who wrote the prompt? Who edited it? That all needs to be recorded in an audit trail so you can trace, evaluate, and improve decisions over time.”
As AI becomes more active in making decisions—and regulators start asking tough questions—insurers will need clear answers.
“You need to be ready to show exactly how decisions were made,” Wilde warned. “That’s going to be the standard moving forward.”