Insurance executives are feeling good about artificial intelligence. Their businesses, according to new KPMG research, have yet to catch up.
KPMG International's report, Unlocking AI value in insurance, found that 44% of respondents place their organisations in the top quartile for AI transformation, and none consider themselves significantly behind. Yet not a single organisation surveyed has fully redesigned sales and distribution or underwriting around AI, and only 3% have done so in policy servicing and claims management.
The survey was carried out between May 20 and 29, 2026 among insurance leaders at organisations with 500 or more employees across 20 countries and six regions. KPMG has not published the number of respondents. The report also drew on 53 insurance respondents from its separate Transforming the Enterprise 2026 study.
The gap between concern and action is stark. Some 77% believed that failing to redesign their enterprise architecture for AI will damage competitiveness within five years, and 68% said moving too slowly is a bigger risk than moving too fast. But 71% still use AI mainly for content generation and routine task automation, and only 29% run end-to-end processes through AI agents or automation.
"The gap between activity and transformation is where the real opportunity and risk now sit," said Dr Frank Pfaffenzeller, global head of insurance at KPMG International.
He added that many insurers remain focused on efficiency gains rather than asking how AI might change the kind of insurer they could become.
Spending patterns bear that out. Some 92% of respondents said AI is helping them improve productivity and cut operating costs, compared with 25% using it to drive growth through new products and services. Nearly half of AI budgets go to operational and back-office efficiency, while new products and revenue models receive 5% to 10%.
Few can say what that spending returns. Only 11% described their view of AI return on investment as very clear, while 23% report limited clarity or none at all. KPMG warned this risks progress being judged by activity and adoption rather than by changes in cost, cycle times, customer outcomes or growth.
The finding echoes other recent research. Capgemini's 2026 World Property and Casualty Insurance Report found around 60% of P&C insurers still at the exploration or proof-of-concept stage, and a separate industry study found that only 10% of P&C carriers have successfully scaled AI.
Data remains the main obstacle. Just 11% said they have the data foundations and governance needed to scale AI beyond pilots. A further 55% call themselves moderately ready, 21% partially ready and 13% not ready, citing fragmented and poor-quality data, unclear ownership and legacy systems.
Skills lag too. Only 8% rated their workforce as highly proficient in AI tools, even though 54% said they provide effective AI training. Accountability is also narrow: technology leaders hold primary responsibility for AI in 45% of organisations, and only 15% have fully integrated AI governance into strategic planning.
Executives nonetheless expect significant change to jobs. By 2029, 72% expected underwriting to run on a hybrid model with fewer people and redesigned roles, while 36% anticipate significant role elimination in claims management and 33% in policy servicing.
Matthew Smith, global lead for insurance strategy and transformation at KPMG in the UK, said the insurers likely to make the most progress would be those combining trusted data, clear accountability and workforce readiness with a long-term view.
For intermediaries, the most telling figure is the zero. No carrier in the survey has rebuilt sales and distribution around AI, and budgets are overwhelmingly aimed at internal cost. That means the broker-facing parts of the value chain, such as submissions, quoting and placement, are likely to stay largely manual on the carrier side for some time.
The underwriting forecast cuts the other way. If most carriers expect leaner, hybrid underwriting teams within three years, brokers may find fewer people at the other end of the phone and more decisions made or triaged by automated systems. The firms whose own data and processes are ready to plug into that change are likely to benefit first.
