AI Could Finally Break Insurance’s 20-Year Profit Stagnation, McKinsey Warns
AI Insurance: AI could give the insurance industry a much-needed profitability boost after nearly two decades of stagnation, according to McKinsey.
AI Insurance: AI is reshaping insurance, from underwriting and claims to risk assessment and customer service. McKinsey says the technology could help insurers break a two-decade stretch of stagnant profitability.
AI Insurance: Global insurance premiums have grown steadily for two decades, but profits haven’t kept pace — and coverage gaps for climate and cyber risk have widened. A new McKinsey & Company report argues AI is the first technology capable of changing that, with direct implications for how consumers shop for, pay for, and experience insurance.
Anyone who has shopped for car insurance, filed a health claim, or been turned down for flood coverage has felt the friction built into the insurance industry. According to a new report from McKinsey’s Financial Services Practice and QuantumBlack, AI by McKinsey, published in July 2026, that friction is now squarely in artificial intelligence’s crosshairs — and the changes could reshape pricing, access to coverage, and the buying experience itself.
The report, “How AI will reshape the economics of insurance: A CEO’s guide to strategy,” was written by McKinsey partners Jason Ralph, Johannes-Tobias Lorenz, Nick Milinkovich, Sid Kamath, and Tanguy Catlin, with associate partner Gabriella Meijer. While it’s addressed to insurance executives, its findings point to real changes headed for everyday policyholders.
Also Read: Who Pays When the AI Gets It Wrong? Inside the Race to Insure Artificial Intelligence
AI Insurance: The Industry Has Been Charging More Without Delivering More
Global insurance premiums have grown at roughly 4.9 percent annually since 2005, reaching an estimated $8.3 trillion in 2025, according to McKinsey Global Insurance Pools data cited in the report. But profits before tax grew more slowly — about 4.3 percent annually — reaching approximately $580 billion. In other words, the industry has been collecting more in premiums without a matching improvement in profitability, a gap McKinsey partly attributes to rising capital requirements.
For consumers, that underperformance has shown up as coverage gaps rather than falling prices. The report cites an estimated $133 billion global protection gap for natural catastrophes in 2025 — with a separate Swiss Re Institute estimate placing the additional premium needed to cover currently uninsured losses at $424 billion. On top of that, less than 1 percent of global cyber costs are currently insured, leaving a gap of roughly $900 billion. Personal lines insurance actually shrank as a share of the global economy, falling from 1.2 percent of global GDP in 2019 to 1 percent in 2023.
AI Insurance: Why Your Premium Dollar Buys Less Than You’d Think
One reason coverage costs what it does: distribution. The report finds that commissions and acquisition costs consume between 10 and 25 cents of every premium dollar in property and casualty (P&C) insurance — and as much as 80 cents of a customer’s first-year premium in life insurance. That cost structure has barely budged in 20 years despite repeated waves of digital investment by insurers.
Layered on top of that, the report’s analysis shows insurance industry cost ratios (expenses relative to revenue) are actually 10 percent higher globally than they were in 2005 — up 22 percent in North America — even as other industries like telecommunications, automotive, and airlines meaningfully cut their own cost ratios over the same period. McKinsey notes this isn’t because technology failed to make workers more productive; labor productivity did improve, by 14 percent in P&C and 24 percent in life insurance. Those gains were simply offset by rising IT costs, compliance overhead, and the complexity of bolting new digital tools onto old systems — costs that, one way or another, get passed on to policyholders.
AI Insurance: Your Next Insurance Policy Might Not Come From an Agent at All
Today, about 85 percent of US P&C insurance and 95 percent of life insurance is still sold through agents, brokers, and managing general agents. The report says that’s now genuinely up for grabs. Nearly half of customers in North America already use AI somewhere in their personal insurance-buying journey — comparing coverage, tracking renewal dates, or getting recommendations — and McKinsey expects that to accelerate.
Two consumer-facing shifts stand out:
Insurance bought at the moment you need it, not before. The report points to the rise of “embedded insurance” — coverage offered automatically at the point of a relevant purchase, such as travel protection bundled into a flight booking or device coverage offered at checkout — as well as on-demand policies enabled by real-time data. For consumers, this could mean less time spent researching and comparing policies manually.
AI assistants doing the shopping for you. Agentic AI tools can already monitor renewal dates, pull live pricing from multiple carriers, and flag better deals before a customer even starts looking. The report frames this as a shift in the “front door” to insurance — away from an agent’s office or a carrier’s website, and toward whichever AI assistant or platform a customer trusts to act on their behalf.
The report is careful to note this won’t happen evenly. In straightforward, high-frequency products like auto or renters insurance, AI-driven platforms are likely to increasingly sit between customers and carriers. In more complex situations — specialty coverage, high-net-worth policies, or advice-heavy commercial insurance — agents and brokers are more likely to stick around, but working faster and cheaper thanks to AI support, rather than being replaced outright.
AI Insurance: Could AI Make Insurers Actually Prevent Losses, Not Just Pay for Them?
Perhaps the most consumer-relevant shift the report describes is a move away from insurance as a purely reactive product — one that pays out only after something has already gone wrong — toward continuous, AI-driven risk monitoring. Examples cited include telematics that adjusts a driver’s “premium per minute” in real time based on how they’re actually driving, satellite- and sensor-based monitoring for commercial property risk, and AI-enabled health coaching aimed at improving long-term health outcomes. McKinsey describes this as a shift from insurers sitting at the margin of a customer’s risk decisions to sitting at the center of them.
Also Read: Insurance AI Production Readiness Is an Operating Discipline
AI Insurance: Could AI Help People Get Coverage They’re Currently Denied?
The report also argues AI could open coverage to people and properties insurers currently avoid. It says the real barrier to covering many climate-exposed properties, cyber risk, and other hard-to-price risks isn’t a lack of demand — it’s that underwriters don’t have reliable enough data to price the risk with confidence, which results in high premiums, restricted coverage, or outright denials. Better real-time data and more accurate claims handling, the report argues, could let insurers price these risks more confidently and extend coverage to customers currently priced out of the market entirely.
That said, the report flags a real risk on the other side: digital and AI-related risks may turn out to be highly correlated — meaning a single AI system failure or cyber event could trigger losses across huge numbers of policyholders simultaneously, unlike traditional, more independent risks. That could make some new AI-related coverage harder to price responsibly, at least in the near term.
AI Insurance: Will AI Change Who You Trust With Your Claim?
The report breaks down insurance “trust” into three distinct pieces and argues AI will affect each differently:
- Trust in human connection during a difficult claim — something the report says AI can’t yet replicate, particularly in moments of genuine distress.
- Trust in credentials, like an insurer’s licensing and financial strength ratings — traditionally an advantage for established insurers, but one the report says is increasingly contestable as AI-native platforms build their own forms of credibility.
- Trust in clear explanation — helping a customer understand exactly what they’re covered for and why a claim was decided a certain way. Here, the report suggests AI may actually outperform the traditional experience, since it’s available at any hour, doesn’t get impatient, and doesn’t make people feel foolish for asking basic questions.
AI Insurance: What Comes Next
The report stops short of predicting an exact timeline, arguing that a rigid five-year forecast risks becoming outdated before it plays out. Instead, it frames the shift as already underway: carriers investing early in AI-driven pricing, claims handling, and customer-facing tools are reportedly already seeing measurably better business results, with McKinsey citing 20 to 40 percent reductions in customer onboarding costs and 10 to 20 percent improvements in agent productivity among early movers, along with total shareholder returns six times higher than industry laggards.
For consumers, the practical upshot described in the report is a slow but real shift already in motion: potentially faster and more personalized quotes, insurance offered at the moment of need rather than requiring active shopping, new coverage options for risks that are currently hard or impossible to insure, and claims processes that increasingly involve AI alongside — or in some cases instead of — a human adjuster or agent.
