A study of more than 300 respondents covering several industries finds a considerable disconnect between how organizations assess their progress on AI and where they actually are with real-world AI integration in the insurance sector.

The third annual EXL U.S. Enterprise AI Study found a significant shift from experimentation to enterprise-wide scale in AI.

Scaling AI is now a high priority for 96% of insurers, compared to 86% in 2025.

Agentic AI is advancing the fastest in risk management, actuarial, underwriting, and customer experience, orchestrating complex, end-to-end workflows across multiple systems and roles, the survey found.

Insurers are also moving beyond point solutions toward redesigning full workflows, connecting intake, conversational, and claims steps at FNOL in P&C and unifying data foundations across underwriting and onboarding in life/annuity.

Though 76% of insurers believe they are ahead of the competition, actual AI adoption in insurance is similar across the industry.

Only 6% qualify as leaders, and insurance holds the highest share of companies in the middle follower category (72%) of any industry surveyed.

According to the survey, leaders were two times more likely to improve operational efficiency with AI than laggards.

Insurers do move more AI pilots into production than any other industry, with 62% of pilots reaching production.

Pilots that advance share three traits: a clear business owner, a defined outcome, and agreement on what “production ready” means.

AI in the insurance industry is most mature in customer-facing and operational areas, the survey found.

The leading applications include fraud detection (54%), customer servicing (54%), financial crime compliance/AML/KYC (44%), risk management (44%), and claims (42%).

In agentic AI specifically, risk management leads at 54%, followed by actuarial, underwriting, and pricing (46%) and customer experience (45%).

Leaders were nearly three times more likely to adapt to market changes with agentic AI than laggards, and two times more likely to redefine the customer experience with agentic AI.

Of the insurers surveyed, 98% believe agentic AI has improved customer experience and that 45% of insurance agentic AI initiatives have reached success.

Adoption remains less mature in underwriting, actuarial, and risk decision-making, which EXL says signals substantial room for deeper AI transformation in insurance.

Measurable operational and financial gains were found in applying AI to insurance underwriting.

Nearly half of insurers (46%) have fully deployed AI in actuarial and underwriting, and leaders generated 40% more revenue growth and 37% more cost reduction than laggards in use cases where AI is applied.

Data is the leading obstacle to enterprise AI in insurance.

Nearly 92% of insurers say their data is a challenge to AI success, and insurance cites data silos as the top barrier more than any other industry.

Confidence in data quality nearly doubled since 2025, yet more than half of insurers now report problems with data efficiency, raising the cost of running AI.

Only 24% consider their edge on data management maturity, and just 38% completely agree they have sufficient governance for ethical and responsible AI use.

Closing these gaps in data readiness and governance is central to a durable AI strategy for insurance companies, said EXL, adding that insurers that want to make progress in stalled AI projects must move pilots into production, improve data, embed AI into workflows, and strengthen governance.