Insurers That Treat AI as Core Strategy Are Pulling Away
Most carriers have AI projects. Few are truly AI native. New research finds that insurers leading on AI integration delivered 6.1x the total shareholder return of lagging peers over the past five years - a gap that's bigger than in most other sectors.
The takeaway is simple: scattered pilots won't move the P&L. Insurers seeing real gains are transforming specific domains - underwriting, claims, pricing, and distribution - and embedding AI into daily work, not just standing up tools on the side.
What's Working: Domain-Level Transformation
Leaders pick a business domain, rebuild workflows end-to-end, and wire AI into the operating model. That means new processes, new roles, new data flows, and product-style delivery - not just models dropped into legacy steps.
With this approach, carriers are reporting clear, quantifiable outcomes:
- 10%-20% improvement in new-agent success and sales conversion
- 10%-15% lift in premium growth
- 20%-40% lower customer onboarding costs
- 3%-5% improvement in claims accuracy
These results come from focusing effort where value is highest, then scaling what works across similar lines and regions.
Six Moves Separating AI Leaders
According to the study, carriers winning with AI consistently execute six "signature moves":
- Align leadership on a business-led AI strategy: Start with growth, loss ratio, and expense targets - then choose AI plays that hit those numbers.
- Strengthen in-house digital talent: Data scientists, ML engineers, product managers, and domain experts working as one team.
- Adopt scalable operating models: Product pods, shared platforms, and governance that lets winning use cases scale fast.
- Use flexible, reusable AI stacks: Modular services, common data models, and tooling that avoids one-off builds.
- Embed data capabilities across the business: Clean, governed, and accessible data at the core of every workflow.
- Invest equally in change and adoption: Training, incentives, and process redesign so people actually use the tools.
A Practical Plan for the Next Two Quarters
- Pick two high-value domains: e.g., small commercial underwriting and FNOL triage.
- Define hard targets: Premium growth, hit ratio, claim cycle time, leakage - with baselines and weekly tracking.
- Stand up cross-functional pods: Underwriting/claims experts, product, data science, engineering, change leads.
- Map and redesign workflows: Decide where AI assists, automates, or augments decisions. Remove steps, not just add tools.
- Build for reuse: Shared data pipelines, feature stores, prompt libraries, and model monitoring.
- Industrialize adoption: Training, playbooks, refreshed KPIs, and incentive alignment for frontline teams.
- Govern for safety and compliance: Model risk controls, bias testing, explanations, audit trails, and privacy by design.
Why Culture Is the Hard Part
Technology isn't the blocker - behavior is. If employees see AI as a side system, usage stalls. If they see it as a daily copilot that makes quotas, loss ratios, and service levels easier to hit, adoption sticks.
Build that mindset by packaging AI into the tools people already use, giving fast feedback on outcomes, and celebrating wins publicly. Keep models accountable with transparent performance dashboards.
For Insurance Leaders: Where to Upskill Fast
If you're standing up domain pods or building internal enablement, focused training accelerates the curve. Explore role-based programs and tool guides that match underwriting, claims, and distribution workflows.
See AI courses by job role for quick paths to adoption inside carrier and broker teams.
Source
For the full analysis and benchmarks, see the report: McKinsey on AI in insurance.
Your membership also unlocks:
