AI adoption divide widens in P&C insurance as 10% of trailblazers outpace peers by 21%

A sharp divide is opening in the property and casualty insurance industry, and it comes down to who actually knows how to use artificial intelligence. A new report from Capgemini's research institute, now in its nineteenth year of tracking the sector, found that only one in ten insurers have moved beyond experimentation and built AI into the core of how they operate. Those ten percent are pulling away from everyone else at a striking pace: they're growing revenue 21 percent faster and seeing their share prices climb roughly 51 percent more over a three-year stretch.

The gap exists not because the lagging insurers lack access to the same technology. It exists because they're spending money on the wrong things. Across the industry, carriers are pouring 72 cents of every AI dollar into technology and infrastructure—servers, software, platforms. Only 28 cents goes to the human side: training people, managing organizational change, building the skills that actually make the tools work. The trailblazers, by contrast, treat AI as a business transformation, not a software purchase. They invest heavily in change management, nearly four times more than their peers. They build explainable AI systems that help employees understand what the algorithm is doing, making them three times more likely to have that infrastructure in place. And they embed AI responsibilities directly into job descriptions, creating clear accountability where most insurers have none.

The cost of this misalignment shows up everywhere. Fifty-five percent of insurers can't even say whether their AI investments are paying off—they have no clear return on investment to point to. Forty-two percent track no AI metrics at all, which means they're flying blind. Without measurement, without clarity on who owns what, AI initiatives end up in the hands of scattered individuals or small teams. Two-thirds of insurers report a shortage of AI skills. Nearly half of employees who have been given access to AI tools for eighteen months say their workday hasn't changed at all. The technology sits there, underused, its potential locked away.

The human dimension of this struggle runs deeper than spreadsheets. Forty-three percent of employees worry that AI will cost them their jobs. Only 14 percent feel they have a clear sense of how AI actually fits into their work. Nearly half of employee time goes to cross-team collaboration, yet most AI tools are built to handle individual tasks in isolation. The infrastructure and the workflow don't match. Data readiness lags too—only 12 percent of insurers say their data is in very good shape, even though they're increasingly relying on messy, unstructured information.

The research drew from interviews with 344 senior executives at major P&C insurers, surveys of 809 insurance workers across underwriting, claims, and customer service, and feedback from 1,113 policyholders. The respondents spanned the Americas, Europe, and Asia-Pacific. What emerged is a portrait of an industry at a crossroads. The trailblazers have figured out that AI only works when leadership sets clear strategic direction, when skilled employees use real-time insights to handle complex decisions, and when AI agents handle the routine, repetitive work. They've redesigned their workflows around this collaboration. They've strengthened their data foundations. They've created what the report calls orchestration managers—people whose job is to align business strategy with AI principles so intelligence can scale across the whole enterprise.

For the rest of the industry, the path forward is visible but steep. It requires embedding AI into everyday collaboration and decision-making, not treating it as a separate initiative. It requires redesigning how work actually happens. It requires accepting that the insurer of the future won't be defined by who has the fanciest algorithm, but by who has built an organization capable of using it.