AI is reshaping insurance - from pricing and fraud detection to customer clarity - ushering in a new era where trust becomes the industry's most valuable asset

Transformation

By Chris Davis

Dec 03, 2025Share

AI is doing more than automating tasks in insurance - it’s changing how insurers calculate premiums, explain coverage, and detect fraud. And according to José Luis Bernal (pictured), chief digital, data and innovation officer at MAPFRE USA, the biggest impact may be trust. 

“AI is bringing much more predictive predictability into the premiums,” Bernal said. “It can be done much more accurately and, with a new technology, much more firm and being explainable to the end customer.” 

For years, customers struggled to understand how premiums were calculated. Now, AI offers more transparency. Bernal said this shift could finally give customers clarity on what they’re paying for - something they’ve long demanded. 

Alongside greater accuracy, automation continues to drive down operating costs. But Bernal pointed to another win: better fraud detection. “Right now, those third customers are paying the premium of the fraudulent customers,” he said. “Thanks to AI, now we can detect much more accurately.” 

He emphasized that effective fraud detection wouldn’t rely on a single algorithm, but rather a blend of machine learning, graph databases, and shared industry datasets. The result: a pricing model that no longer penalizes honest customers for fraudulent behavior elsewhere in the pool. 

Better understanding, smoother transactions 

Bernal also highlighted how AI is improving the way policies are explained. “As a normal customer, reading the policy is pretty hard, and knowing exactly if you are covered or not in certain areas is not easy,” he said. Now, with AI-powered tools, customers can ask specific questions and receive clearer answers about their coverage. 

Homeowners’ insurance is a prime example of how AI is simplifying the customer-agent interaction. Traditionally, determining the value of a home’s contents has been guesswork. That’s changing. “There are tools right now in which, taking a few pictures, you can have a quick estimate of the value of your house,” Bernal said. He added that many customers significantly underestimate the cost of replacing their belongings - AI can help close that gap and streamline the quoting process for both customers and agents. 

Risk models need to keep up with emerging threats 

The pace and complexity of risks - from climate to cyber - are increasing, and Bernal said AI can reshape how underwriters respond. 

“What AI is bringing is much more accuracy and better predictions of what's going to happen,” he said. That could lead to more price volatility, especially at renewal. But with improved data and modeling, insurers could justify sharper adjustments. That’s where regulatory tension could emerge. 

“In the US, compared to my background in Europe, regulators pay much more attention to whether you have enough money to cover your liabilities,” Bernal said. “The way you get to a price, they are much more free to take this to the market.” He expects US insurers to push for more pricing flexibility as AI improves underwriting accuracy. 

More volatility, he added, could also create demand for new insurance products that shield policyholders from rate swings. “Do I want to freeze rates or something like that for the next three years?” he said. “This is a product that can be built.” He said several firms were already exploring these ideas. 

Legacy systems still blocking transformation 

Despite the buzz around AI and digital innovation, Bernal said many insurers still aren’t meeting the basic tech requirements. “You need your system in the cloud,” he said. “The second, for me, is your system has to have good API connection.” Without these, access to new tools remains limited. 

Even for those with the right infrastructure, challenges remain. Data - especially unstructured data - must be reorganized to work with AI tools. And there’s another missing piece: the orchestration layer. “You have to acknowledge you need this piece of technology,” he said, referring to platforms that integrate different AI tools and processes. Without it, automation stays fragmented. 

Insurers must avoid falling for tech hype 

When it comes to digital transformation, Bernal said many companies make the same mistake: buying into technology for its own sake. “You realize your problem has not been solved,” he said. “Technology per se is not going to solve your problems.” 

He advocates a “bottom-up” innovation approach - giving teams controlled ways to test features and measure ROI without large upfront investments. “Nobody likes it, but the return on investment expected... we will need to have discipline in measuring and taking decisions based on ROI,” he said. 

That ROI model is evolving. “In the past, making an upfront investment was very high,” Bernal said. “Today, with these orchestration layers, the upfront investment is going to be very small.” That means insurers can launch smaller pilots, validate quickly, and scale what works. 

Modularity is the future of insurance IT 

Bernal sees insurers moving toward modular systems that can plug into specialized vendors. A cloud-first, API-enabled architecture allows legacy systems to tap into “best of breed” tools, such as advanced rating engines, without full replacement. 

“Leader by leader, our legacy systems are going to be more and more modular,” he said. This modularity opens the door to more personalized, lower-cost service at scale. It also allows companies to access external “marketplaces” of tech solutions that can rapidly improve the customer and employee experience.