McKinsey: AI Could Reshape the Global Insurance Industry

McKinsey’s recent analysis argues that artificial intelligence (AI) may trigger a rapid, fundamental transformation in the insurance sector—one stronger than prior waves of globalisation, digitalisation or platformisation. The consultancy notes that while the industry’s global economic role has been relatively stagnant over the past two decades, risk environments have become more complex, and AI can both create new risks and provide tools to insure previously inaccessible exposures.

Where the sector has fallen behind

The report identifies four structural shortfalls that have left the insurance industry at a relative disadvantage:

• Growth and relevance: global insurance premiums grew on average 4.9% per year since 2005, reaching about $8.3 trillion in 2025. Pre-tax profits rose at only 4.3% annually, which—combined with rising capital requirements—signals deteriorating operating efficiency. The share of personal lines relative to global GDP fell from 1.2% in 2019 to 1.0% in 2023.
• Large protection gaps: uninsured losses from natural catastrophes were estimated at $133 billion in 2025, while less than 1% of global cyber losses are covered by insurance. Another estimate puts the additional premium needed to cover uninsured natural catastrophe losses at $424 billion.
• High distribution costs: in property & casualty (P&C), commissions and acquisition costs consume 10–25% of premium income; in life insurance, first-year acquisition costs can consume up to 80% of the premium.
• Persistent productivity stagnation: P&C operating expense ratios have remained between 27% and 32% for two decades—leaders sit below 22% while laggards exceed 32%.

These structural issues help explain why the industry is losing relative relevance even as exposures such as cyber and AI-related risk expand.

AI’s double-edged impact: new risks and new markets

McKinsey outlines three pillars of an optimistic AI-driven scenario:

1. New risk categories: AI liability, non-physical business interruption, and workforce transformation driven by AI are expanding faster than current products can cover.
2. Shift from transfer to partnership: AI-enabled continuous, data-driven risk monitoring allows proactive risk reduction rather than purely reactive claims payment (examples include telematics-based driving coaching, satellite/IoT-informed commercial risk prevention, and AI health coaching).
3. Improved underwriting and claims: real-time data and continuous model recalibration can sharpen pricing and reserving; better claims decisions reduce the cost uncertainty that forces conservative pricing.

At the same time, McKinsey warns of valid downside scenarios: digital dependencies and shared cloud or model infrastructures can create systemic, highly correlated losses that hinder diversification and make accurate pricing difficult. Alternative capital that expects independent risks may be reluctant to back strongly correlated digital exposures.

Pressure on traditional distribution

The consultancy highlights that agent and broker channels have dominated insurance distribution for decades—around 85% of P&C premiums and 95% of life premiums in the United States still flow through intermediaries. AI threatens that structure for the first time by planting AI-driven platforms at critical points of customer decision-making: owning customer data, gaining privileged access to demand, and orchestrating complex ecosystems. Nearly half of North American customers already use some AI tool in their insurance purchase journey.

For standardised retail products, AI agents and digital platforms can rapidly interpose themselves between customer and insurer. In more complex segments—mid-market commercial, specialty risks or high-net-worth personal lines—AI currently tends to lower costs and augment broker capabilities rather than fully replace human intermediaries. The strategic question becomes not whether AI will eliminate all distribution roles, but who will capture the value and own the customer relationship in the reconfigured value chain.

Cost efficiency: an end to two decades of stagnation?

Historically the industry has not materially improved cost efficiency. While telecoms, automotive and airlines have reduced their cost ratios since 2005, insurers’ cost ratios have globally risen by about 10 percentage points despite heavy investments in automation and digital tools. AI could change this equation in two ways:

• Broader impact on cost structure: unlike narrow automation, AI can reduce marginal costs across underwriting, claims and service simultaneously; early evidence shows 20–40% savings in customer acquisition costs and 10–20% productivity improvements for agents.
• Changing economies of scale: although compute, models and integration carry real costs, AI can lower unit costs for many underwriting, service and claims tasks below current human- and rules-based levels. Firms that scale AI successfully can grow volume much faster while achieving efficiency levels that laggards will struggle to match.

McKinsey outlines two viable strategic routes: scale—where large insurers and brokers amortise infrastructure and modelling investments across a broad base—and deep specialisation—where focused players combine niche underwriting skills, distribution platforms, capital providers and claims capabilities through partnerships. The firm cautions that delay and half-measures are risky.

Organisational speed becomes a strategic factor

The report stresses that the competitive edge is not simply access to AI tools but organisational ability to deploy them at market speed. Three main bottlenecks are identified:

1. Decision speed: unclear governance and slow approval chains delay AI strategic decisions.
2. Data architecture: insurer data is often organised around historical claims and product groups, which hampers real-time signal detection, cross-sell and customer-level pricing.
3. Operating model: technology and business teams frequently work in silos rather than embedded, multidisciplinary squads that iterate quickly.

These limitations are organisational rather than purely technical. McKinsey notes that insurers leading in AI adoption addressed strategy, talent, operating model, technology infrastructure, data architecture and change management simultaneously.

Insurers that have advanced beyond pilots to industrial-scale AI deployments have delivered materially higher total shareholder returns—the report cites a roughly sixfold TSR advantage for leaders vs. laggards in AI adoption.

Three immediate steps for leaders

McKinsey recommends three immediate actions for executives:

1. Define strategic ambition and assign clear accountability: decide which business problems to prioritise and name a senior leader with full authority and resources to execute the AI transformation.
2. Build execution capability: focus on the two or three highest-value use cases, create a modular architecture for rapid integration and iteration, and redesign operating processes to match technological speed.
3. Invest in talent and change management: embed AI skills within business functions rather than isolating them in a centre of excellence, and deploy incentives and feedback mechanisms that drive behavioural change at scale.

The consultancy warns that treating AI as an incremental modernisation risks leaving insurers with entrenched competitive gaps, whereas leaders who rewire their operating model around new capabilities can generate sustained advantage.

Conclusion

McKinsey concludes that AI is both an opportunity and a risk for the insurance industry: it can help close protection gaps and improve cost efficiency, but it also creates new correlated exposures and rewards organisational speed and strategic clarity. The ultimate outcome will depend on which insurers can reconfigure strategy, data and operating models to scale AI-driven advantages.

The topic will be discussed in a dedicated session at the Portfolio Future of Finance 2026 conference on 23 September 2026.