Lemonade

Overview

AI-first, Israeli-founded US insurance carrier (NYSE: LMND) spanning renters, pet, home, life, and car insurance; end-to-end carrier in >90% of its business. Matters to this KB as the team’s richest carrier case study: two Lemonade operators (tom-granit-ramu, ex; shai-slobodov, current) were interviewed on 2026-08-03 about what running an AI-native carrier actually takes — regulation, pricing filings, telematics, claims, and distribution — to calibrate the Insurance 2.0 direction.

Relationships

  • tom-granit-ramu — former car-insurance product lead.
  • shai-slobodov — current platform/claims group lead.
  • next-insurance — the SMB-commercial counterpart; NEXT went MGA → carrier while Lemonade was carrier from day one.
  • Founders: Shai Wininger (previously founded fiverr) and Daniel Schreiber.

Notes

  • Car-insurance book built telematics-first (CMT phone data; Metromile device acquisition; Tesla partnership for historical driving data at signup) — an underwriting edge incumbents can’t replicate on legacy books.
  • Started militantly B2C; after two years in car insurance went all-in on a broker-facing sales system — evidence that broker distribution is unavoidable in US insurance.
  • AI-first claims experience backfires on complex car claims (customers want a human); works for simple renters/pet claims.
  • Built but never shipped crypto insurance; internally explored trials/ships/workplace lines.
  • Grew partly by acquiring companies with regulatory filings already done (e.g. Metromile).
  • First product (NY renters) took ~1.5 years of regulatory approval with near-zero revenue; carrier margins run ~8–10% of premium vs ~90% at real SaaS.
  • Under 1% of the US auto market after ~4–5 years in the product — the cautionary tale on car-insurance CAC and competition.
  • ~10M policies of proprietary data; the team’s data-lock-in example (your telemetry inside Lemonade can be used against you) and routing example (its Tesla/autonomous product with a claimed ~50% price cut).

Mentioned in