Instant, narrow-scope insurance claims
The problem
Traditional claims handling is slow and labour-intensive, hurting customer experience for simple, low-value claims.
The AI approach
A claims bot (“AI Jim”) handles first-notice-of-loss conversationally, cross-references the policy, runs anti-fraud checks, and can approve and pay straightforward claims with no human in the loop; a separate bot (“AI Maya”) handles onboarding.
Evidence it works
Lemonade’s claims bot “AI Jim” set a world record in 2016 by settling a claim in ~3 seconds end-to-end (reviewing it, checking the policy, running anti-fraud algorithms and paying), later beaten by a ~2-second UK claim in 2023. The share of claims handled end-to-end without any human has grown from ~30% (2021) to roughly 40–55% (2025), with complex or questionable claims routed to people.
What “good” looks like
Near-instant resolution of genuinely simple claims, with complex or suspicious claims cleanly escalated to humans.
Feasibility & cost shape
Most achievable where products and claims are standardised; the design challenge is the escalation boundary — knowing exactly what AI should not decide.
A clean positive counterpoint to the Klarna case below: automation succeeds when scope is tight and the hand-off to humans is deliberate. That scoping discipline is the whole game.
Based on publicly reported information about the Lemonade work.
This is an industry example included for illustration. It is not a Leia Intelligence project, and no client of ours is implied. Figures are as publicly reported by the original parties.
Sources: Lemonade (company blog / shareholder communications) · contemporaneous press coverage