Property and casualty insurance organizations throughout the world are leaving money on the table by narrowly using AI to create “efficiencies” without fundamentally changing the human-centric processes that prevents them from scaling AI, a new global report says.
“AI strategies that mainly focus on efficiency create natural pressure for near‑term returns,” says the CapGemini report, The intelligence era in P&C, released this week.
“According to our analysis of the top 20 [global] P&C insurers, ranked by gross written premiums earned (S&P Global, 2025), only 35% explicitly link their AI strategy to business outcomes beyond efficiency.”
The report highlights a small group of intelligence trailblazers, roughly 10% of insurers, that are using AI to the best competitive advantage. Compared to their P&C insurance industry peers, these organizations have increased revenue growth by 21% and have bumped up their share prices by 51% over a three‑year period.
“These organizations outperform peers on revenue growth and share price performance by treating AI as a core operating capability, not just a technology initiative,” the report states.
P&C organizations need to change in three ways simultaneously to make the most out of AI, the report notes. They need to make technological changes, address talent changes, and also make operational changes.
Changing tech
Most organizations are focussing simply on the tech side, the report says.
“Insurers’ spending patterns make the challenge clear,” the report states. “On average, 72% of AI investments go toward technology and infrastructure, and only 28% to change management.
“That imbalance leaves many programs short of the organizational support required to move from pilots to full‑scale implementation.”
Changing talent and roles
It’s not just a matter of installing the AI, accessing and capturing unstructured data, and developing the data sets required for AI analysis, CapGemini notes. To make AI a core operating capability, companies must address talent and operational gaps as well.
For example, many agree only human experts can make judgment calls about an appropriate use of AI. This is what CapGemini describes as an “expert‑centric P&C insurer.” And they include “orchestration managers,” who translate business strategy into AI principles and govern how intelligence scales across the organization.
Without these types of experts, AI isn’t scaled. “But with them, [AI systems] become coherent, well‑governed systems.”
In addition to orchestration managers, three other types of leadership are required, says the report:
- Executive leadership sets guardrails
- Human subject matter experts such as underwriters, claims specialists, and distribution specialists define outcomes and establish the accreditation frameworks AI must meet before they’re trusted to act
- Experts who handle situations when high‑volume work gets escalated, and complexity exceeds defined thresholds.
Collaboration between these experts across various departments in an organization remains a work in progress, CapGemini says.
“Changing how work gets done remains the central challenge, even for those already ahead on strategy, technology, and adoption,” the report says. “Forty-nine percent of employee time is spent on cross‑team collaboration, yet most AI tools operate at the individual task level, automating work after decisions are made rather than shaping those decisions.”
Ideally, AI would give executive and team members real-time access to data they need to shape strategic decisions, the report says.
To do this, a company will need to shift its focus from using only structured data to unstructured data. But only 12% of insurers reported high maturity in data readiness.
Changing operations and process design
Finally, insurers need to change their operations to incorporate the introduction of AI.
“AI has been added to workflows built for humans, including sequencing, handoffs, and decision points, none of which were originally intended to incorporate AI,” the report states.
On top of that, insurers need to address their workers’ suspicions about AI.
“Forty-three percent of employees cite job security as one of their top concerns about AI, and 25% worry that the transition to AI will increase rather than reduce their workload,” the report says. “Employees navigating genuine uncertainty about their future are unlikely to lean into a technology they associate with displacement.
“Until insurers address process design and the trust deficit together, transformation will remain out of reach.”