More than four in five insurers report measurable premium growth from AI and data initiatives, but fewer than one-quarter have deployed the technology across their entire organization, according to Accenture. The findings show a wide gap between successful individual AI projects and adoption across insurance operations.
Accenture surveyed 263 senior insurance executives responsible for AI, data, technology or business transformation across the Americas, Europe and Asia-Pacific.
The research also included interviews with 15 executives from major global insurers. Accenture sells technology and transformation services to insurers, giving it a commercial interest in the subject examined by the report.
Eighty-one percent of surveyed insurers said data and AI initiatives had increased gross written premiums by at least 5%. Another 7% reported improvements above 20%, with better pricing, personalization and cross-selling cited among the sources of growth.
Despite those results, only 23% said they had achieved enterprise-wide AI adoption. In many companies, AI remains concentrated within individual functions or teams rather than operating across underwriting, claims, distribution and servicing.
Revenue growth also remains a secondary objective for many insurance AI programs. Only 32% of executives placed revenue growth and business expansion among their three main reasons for investing in AI and data.
Technical adoption is moving faster than insurers’ ability to connect AI with business outcomes. Eighty-three percent reported moderate or severe gaps in the skills needed to translate between AI capabilities and business requirements.
Insurers are investing in employee training, although those efforts remain limited in scale. Seventy percent operate targeted AI skills programs, while only 14% have extended such training throughout the organization.
Existing technology presents another obstacle. Half of respondents identified legacy system integration as their main difficulty when deploying AI and data at scale, while 45% cited data quality and accessibility.
Those issues matter because insurance AI depends heavily on information distributed across policy administration, claims and underwriting systems. Fragmented data limits how easily insurers move successful applications from one department into broader production use.
Agentic AI is attracting substantial attention as carriers consider more automated workflows. Sixty-eight percent of respondents expect AI agents to change roles and operating processes across insurance.
Accenture’s separate 2026 Pulse of Change research found 57% of insurance employees already work regularly with AI agents. That was 13 percentage points above the average across industries surveyed.
Underwriting is one area where insurers report measurable operational results. Eighty-three percent said AI had reduced underwriting turnaround times by at least 5%, while 81% reported better risk assessment and pricing accuracy.
Accenture argues that insurers should move beyond isolated applications and redesign larger workflows around AI. Its report focuses on connecting AI investment with business objectives while developing employee skills and improving underlying data systems.
Governance is already more established at many carriers. Fifty-six percent of surveyed insurers said they had implemented formal AI governance frameworks governing how the technology is developed and deployed.
The report identifies workforce preparation as an important part of further adoption. Underwriters, actuaries and claims professionals increasingly need enough AI knowledge to review outputs and supervise automated processes rather than leaving expertise solely with technical teams.
Insurers also face the longer-term task of modernizing legacy technology while supporting new AI applications. Accenture recommends pursuing shorter-term AI projects alongside larger changes to core data and technology infrastructure.
The findings indicate that insurers are already receiving financial and operational returns from individual AI initiatives. The larger challenge is moving those results beyond separate use cases, with only 23% of surveyed carriers reporting AI adoption throughout the enterprise.
