Directions Brain

Directions offsite — SMB insurance value-chain mapping (כיוונים)

5 min read

Directions offsite — SMB insurance value-chain mapping (כיוונים)

Source: raw/meetings/notetaker/2026-08-03-meeting-11-30.json

Summary

Opening session of the 2026-08-03 insurance offsite at Guy’s house, driven by a research deck saar-arbel prepared (insurtech “three waves,” top-10 glossary, positioning quadrant). The team mapped the broker→MGA→carrier value chain and BOR economics, and converged on a strategic thesis: the money ultimately sits with whoever takes the risk (carrier), but the only realistic entry is smart brokerage → MGA → carrier over time — the “Robinhood of insurance” playbook. A preliminary wedge was defined: Continuous Portfolio Management for multi-carrier SMBs with 5+ policies. Deliberate prep discipline for the upcoming alon-huri conversation ran through the whole session.

Key takeaways

  • Value-chain roles and economics. Broker = distribution only, no risk, ~10–20% of premium at every renewal. MGA = “half carrier”: underwriting authority delegated from carriers under an expanding underwriting-credit line (start ~100 premium, ~$50–70 returns as claims.
  • Broker economics compound like SaaS. A 4.5K/year recurring at ~90% retention plus upsell — “you’re selling like SaaS with 90% retention; no such models exist.” Micro-accounts (400 → ~4-year payback), which is why agents under-serve them.
  • BOR is the activation event. A one-page Broker-of-Record letter moves the whole account: old agent gets a 5–10 business-day window, then the new broker collects renewal commissions at zero new cost to the client. “The audit is the pitch; the BOR is the activation; renewal commission is the revenue.” Renewal outreach happens 60–90 days pre-renewal.
  • Path precedent. next-insurance went MGA → carrier; lemonade is the outlier that was carrier from day one (requires enormous capital). Rule of thumb: broker → carrier is possible, carrier → broker is not. Becoming a meaningful carrier is a 10–15 year marathon.
  • Insurtech three waves. Wave 1 (digital carriers/MGAs — Lemonade, NEXT) proved SMBs buy online but “a better buying experience is not a business model.” Wave 2 (comparison/wallets — CoverWallet, Bold Penguin, Marble) proved point-of-purchase has zero post-sale retention. Wave 3 (now — “replace the broker”: harper et al., ~$1.6B VC in Q1) — nobody has survived a public renewal cycle. “AI-native is the entry ticket, not the differentiator”; the open question is who owns the relationship after placement.
  • Target segment. 87% of US SMB policies sold through agents/brokers; 77–97% of SMBs underinsured while simultaneously overpaying (duplicate coverage). Sweet spot: 10–20-employee businesses with 5+ policies across multiple carriers (the “Maya” persona: BOP → workers’ comp → cyber → 4 carriers, 3 agents, no idea who’s who).
  • Robinhood lifecycle mapped in phases: introducing broker on rented rails (Apex Clearing) → self-clearing (the MGA-shaped move) → bank-charter withdrawn in favor of partner banks → fee layer on the business of record. Caveat: in finance the client-facing layer keeps the economics; in insurance the risk-taker usually does, because catastrophe risk sits on the insurer.
  • Personal-agent data thesis (nizan-shifman) vs. data reality check (guy-barkat). Nizan: if telemetry/personal data sits with the client’s agent rather than the carrier, the agent negotiates only in the client’s favor; in ~10 years “there’s no such thing as non-personalized insurance.” Guy’s pushback: “an insurance company is worth the quality of its data — that’s the basis of the industry, not a discovery”; generic extra data points don’t build a company — you need vertical depth (“bring me 19 data points per vertical, not breadth”).
  • Don’t sell savings. Research showed nobody switches for $5 cheaper — “sell coverage, confidence, and continuous service.”

Decisions

  • Entry via brokerage/MGA, not carrier — promoted to 2026-08-03-enter-insurance-via-brokerage-mga-not-carrier.
  • Strategic north star: “Robinhood of insurance” — own the customer relationship as a smart multi-carrier broker, then integrate up the risk stack.
  • Wedge: Continuous Portfolio Management (not point-of-purchase) for SMBs with 5+ policies — the direction already captured in smb-insurance-portfolio-brokerage.
  • Accepted regulatory constraint: no new “pay-per-exposure” category; stay within existing lines with dynamic pricing.

Action items

  • Team — continue structured research sessions and master the terminology (BOP, P&C, BOR, MGA economics, loss ratios, reinsurance) before the alon-huri conversation.
  • Team — go through the last two years of YC insurance investments one by one; dissect coverwatch’s “we’re alone” claim.
  • saar-arbel — dig into Robinhood’s lifecycle: broker vs. MGA-equivalent at each phase.
  • Team — follow-up session mapping the concrete jobs of each value-chain block (broker, MGA, carrier, wholesaler) to pick the exact entry point.

Open questions

  • Timing of the personal-AI-agent world: 1–2 years (Nizan) vs 3–5 (Guy) — and what to build in the delta period?
  • Is “Robinhood of insurance” a broker or an MGA at entry, and what convinces a carrier to extend underwriting credit?
  • Which vertical to go deep on, and which ~19 data points per vertical actually move pricing?
  • Adverse selection: won’t carriers eventually punish an agent that routes them only bad risks?
  • What does the wholesaler block actually do? (flagged as not understood)
  • Is there venture-scale money in pure brokerage?

Graph View

Backlinks