AI Could Break Insurance's Two-Decade Growth Stalemate, McKinsey Says

Two Decades of Stagnant Efficiency in Global Insurance

Global insurers have grown premiums steadily for twenty years yet failed to translate that expansion into meaningful efficiency or profit margin improvement, according to new research by McKinsey & Company. Annual gross written premiums expanded roughly 4.9 percent since 2005 to an estimated $8.3 trillion in 2025, with profits before tax edging up just 4.3 percent to roughly $580 billion. Over the same period, insurance cost ratios rose 17 percent globally even as telecommunications, automotive and airline sectors reduced theirs.

The industry has also struggled to stay relevant relative to the size of the economy. Personal lines insurance slipped from 1.2 percent of global GDP in 2019 to 1.0 percent in 2023 despite an intensifying risk environment. Meanwhile, protection gaps have widened dramatically: the natural catastrophe protection gap reached $133 billion in 2025, and less than 1 percent of global cyber costs — representing a gap of about $900 billion — are currently covered by insurance, the report notes, citing data from Aon and the Financial Stability Institute.

McKinsey argues that prior waves of digitization tested insurance’s edges but never altered its underlying economic structure. Distribution remains overwhelmingly intermediated, with agents, brokers and managing general agents accounting for roughly 85 percent of U.S. property & casualty premiums and 95 percent of life premiums. Commission levels have barely changed since 2005, and distributors have consistently outperformed carriers in total shareholder returns.

How AI Threatens – and Could Reshape – the Insurance Value Chain

The Stagnation That Digitisation Never Broke

While insurers did improve labour productivity — 14 percent in property & casualty and 24 percent in life insurance across claims, servicing and policy issuance — those gains were offset by rising IT costs, compliance overhead and the complexity of grafting digital tools onto legacy systems. That experience distinguishes insurance from industries where digitisation fundamentally reshaped customer relationships and cost structures, and it explains why McKinsey believes AI could be different.

Why AI May Outperform Decades of IT Investment

Generative and analytical AI can tackle activities that sit at the heart of insurance economics: underwriting, pricing, claims management and customer acquisition. McKinsey identifies four dynamics that have defined the sector’s stagnation — fading relevance, high distribution costs, flat productivity and a slow pace of change — and asserts that AI can address all four simultaneously. Already, early implementations are producing 20 to 40 percent reductions in customer onboarding costs and 10 to 20 percent improvements in agent productivity.

Where Distribution Faces Real Disruption

With nearly half of North American customers already using AI in their personal insurance-buying journeys, the report warns that agentic AI tools capable of monitoring renewals, comparing coverage and recommending switches could redirect the “front door” away from agents and carrier websites. Disintermediation is expected to move fastest in commoditised personal lines. In more complex segments such as midmarket commercial and specialty risk, AI is likely to compress costs and boost broker productivity rather than replace advisers outright — but even that would challenge a distribution structure that has kept commissions almost static for two decades.

Two Paths — and a Warning

McKinsey lays out two viable competitive paths: deep specialisation, where firms excel in narrow domains and access other capabilities through partners, and platform orchestration, where carriers build AI-powered ecosystems that retain customer ownership. The firm also cautions that digital risks such as AI liability and systemic cyber exposure may be highly correlated and difficult to diversify, urging carriers not to chase premium into new lines without first understanding the underlying risk structure.

Strategic Imperatives for Insurers and Intermediaries

• Carriers should audit which activities AI can realistically retool. The 20–40 percent onboarding cost reductions cited by McKinsey are based on existing deployments; firms not yet testing AI in quote-and-bind workflows or claims triage risk falling behind competitors who are.
• Personal lines insurers must prepare for the erosion of the agent front door. With customer AI usage near 50 percent in North America, investment in direct-to-consumer AI interfaces and product-comparison tools is no longer optional — it is a retention necessity.
• Brokers and agents should reposition around advice, not access. In complex lines, AI will augment rather than replace advisory relationships, but productivity gains mean fewer intermediaries will be needed for the same volume of business.
• Companies eyeing new risk categories such as cyber and AI liability need to build granular underwriting models first. McKinsey’s correlation warning implies that a rush for market share in poorly understood lines could create concentrated exposure that traditional diversification cannot mitigate.