Seventy-seven percent of insurance executives now believe that failing to respond to artificial intelligence will erode their competitiveness within five years. More than two-thirds go further, telling researchers that moving too slowly on AI transformation is the bigger danger, not moving too fast. Yet despite that urgency, most insurers remain focused on using AI to make existing processes faster rather than rethinking what their businesses actually sell and how.
A gap between ambition and outcome
These findings, drawn from global research commissioned by KPMG involving insurance leaders across 20 countries, point to a widening disconnect. Investment in AI is climbing. Executive attention is high. New use cases, from underwriting automation to claims triage, keep surfacing across the industry. But the bulk of that effort is being channeled into optimizing operations that already exist: faster claims processing, leaner policy servicing, more automated back-office work. Few organizations are using AI to redesign how value gets created, delivered or monetized in the first place.
That distinction matters. Efficiency gains are real and valuable, but they are also finite and increasingly table stakes. Competitors can replicate a faster claims workflow. It is much harder to replicate a reimagined product, a new distribution model or a fundamentally different customer relationship built around continuous risk insight rather than an annual policy renewal.
Why insurers default to incremental change
The pull toward optimization over reinvention is not surprising. Insurance is a heavily regulated, risk-averse industry built on actuarial precision and long-tail liabilities. Deploying AI inside a known process, with defined inputs and outputs, is a contained and measurable undertaking. Reworking a business model touches pricing, distribution, regulatory approval, legacy systems and organizational incentives all at once. It is slower, riskier and harder to justify in a single budget cycle.
There is also a structural issue: many insurers still measure AI initiatives using the same metrics applied to traditional IT projects, namely cost reduction and process speed. Those metrics reward automation of the familiar. They do not capture the value of a new revenue stream that does not yet exist, which helps explain why transformation efforts cluster around what is already measurable rather than what might be possible.
A framework for moving beyond experimentation
The research introduces a three-horizon approach to help insurers sequence their efforts rather than treat AI as a single undifferentiated initiative. The first horizon covers operational efficiency, where most current investment sits. The second involves redesigning customer experience and the operating model around AI, rather than layering it on top of unchanged workflows. The third is the most demanding: using AI to build genuinely new business models, products or value propositions that were not previously possible.
- Horizon one: automate and optimize existing underwriting, claims and servicing workflows.
- Horizon two: redesign customer journeys and internal operating models around AI-native processes.
- Horizon three: create new products, pricing structures or revenue streams enabled by AI capability.
Few insurers operate meaningfully in the third horizon today. The report frames this not as a failure of ambition but as a natural sequencing problem: organizations need the discipline to treat horizon one as a foundation rather than a destination.
Implications for leaders and the wider market
For insurance leaders, the practical challenge is less about acquiring more AI tools and more about building a clear point of view on where the business needs to be in several years, then working backward. That requires governance structures capable of overseeing AI decisions that touch pricing, underwriting fairness and customer data in ways traditional model-risk frameworks were not designed for. It also requires honest internal conversation about which existing revenue lines AI-enabled competitors or new entrants might threaten.
The stakes extend beyond individual companies. Insurance underpins consumer financial protection, from health coverage to property and casualty risk. How insurers use AI to assess risk, price policies and handle claims has direct consequences for fairness, transparency and access to coverage. Regulators in multiple markets are already scrutinizing AI-driven underwriting and claims decisions for bias and explainability. Insurers that treat AI purely as a cost-cutting tool risk building systems that are efficient but difficult to justify to regulators or customers when decisions are challenged. Those that invest in transparent, well-governed AI redesign are more likely to sustain both competitive advantage and public trust as adoption deepens across the sector.