Global insurers are increasing their use of artificial intelligence, but many are yet to make the deeper changes needed to turn the technology into new products, revenue streams and redesigned insurance operations, according to a new report by KPMG International.
The report, Unlocking AI Value in Insurance, found that 44 percent of insurance executives surveyed consider their organisations to be in the top quartile for AI transformation. However, the research also shows that the industry’s use of AI remains concentrated largely on routine automation and productivity rather than a broader redesign of how insurance businesses operate.
The gap is particularly visible in core insurance functions. KPMG said none of the organisations surveyed had fully redesigned sales and distribution or underwriting around AI, while only three percent had reached that stage in policy servicing and claims management.
The findings suggest that while insurers increasingly recognise the importance of AI, adoption is still taking place largely within existing business structures rather than fundamentally changing them.
KPMG found that 77 percent of insurance executives believe failure to redesign their enterprise architecture for AI could undermine competitiveness within the next five years. At the same time, 71 percent said their main use of AI remains content generation and routine task automation.
Only 29 percent said they were running front-to-back processes through AI agents or automation, while 68 percent considered moving too slowly on AI transformation a greater risk than moving too quickly.
The report shows that insurers are seeing immediate benefits from AI, particularly in reducing costs and improving productivity.
About 92 percent of respondents said AI is helping their organisations improve productivity and reduce operating costs. By comparison, only 25 percent said they were using AI to support growth through new products, services and AI-enabled offerings.
KPMG said nearly half of insurers’ AI budgets are being directed towards operational and back-office efficiency, while only 5 to 10 percent is allocated to new products and revenue models.
The difficulty for insurers is increasingly becoming how to determine whether those investments are producing measurable business value.
Only 11 percent of respondents said they have a very clear view of their return on investment from AI, while 23 percent reported having limited clarity or no clear view at all.
According to KPMG, this creates a risk that insurers measure AI progress by the number of tools deployed or tasks automated rather than by improvements in costs, processing times, customer outcomes and revenue.
Frank Pfaffenzeller, global head of insurance at KPMG International, said the industry understands that AI could change competition, customer expectations and business models, but many insurers remain focused on efficiency rather than considering how the technology could change the nature of their businesses.
Data remains a major constraint
For insurers seeking to move beyond AI pilots, the report identifies data readiness as one of the biggest obstacles.
Only 11 percent of respondents said their organisations have strong data foundations and governance capable of supporting AI at scale. Another 55 percent described their organisations as moderately ready, while 21 percent said they were only partially ready and 13 percent were not ready.
KPMG attributed the challenges to issues including fragmented data, poor data quality, unclear ownership and legacy technology systems.
The issue is particularly important for insurance because AI applications depend heavily on access to reliable historical and real-time data.
Underwriting, pricing, claims processing, fraud detection and customer service all rely on insurers being able to access and analyse large volumes of information. Weak data foundations could therefore limit how far insurers can move from basic automation to more sophisticated AI-driven decision-making.
The report also identified a skills gap. Only eight percent of insurers surveyed rated their workforce as highly proficient in AI tools, although 54 percent said they provide effective AI training.
The adoption of AI is also expected to alter the structure of insurance work.
KPMG’s research found that 72 percent of respondents expect underwriting to operate through a hybrid model by 2029, involving fewer people and redesigned roles.
The changes could extend to claims and policy servicing, with 36 percent of respondents anticipating significant role elimination in claims management and 33 percent expecting similar changes in policy servicing.
At the organisational level, technology executives such as chief digital, technology and information officers hold primary responsibility for AI in 45 percent of the insurance organisations surveyed.
However, 43 percent said AI ownership is centralised while understanding of the technology remains uneven outside senior leadership. Only 15 percent said AI governance is fully integrated into strategic planning.
KPMG said the next phase of AI adoption could therefore move beyond automating individual tasks towards redesigning customer journeys, operating models and decision-making processes.
Matthew Smith, global lead for insurance strategy and transformation and partner at KPMG in the UK, said insurers’ next challenge is to move beyond using AI to make existing processes more efficient and instead rethink how they operate and manage risk.
He said insurers that combine reliable data, clear accountability and workforce readiness with a long-term approach to AI transformation would be better positioned to create value from the technology.
