Directions Brain

Broker Jobs-to-Be-Done Map

7 min read

Broker Jobs-to-Be-Done Map

Definition

A nine-job functional decomposition of the end-to-end SMB commercial-insurance brokerage workflow, drafted by saar-arbel and nizan-shifman as the spine for both the product’s V1 scope and the investor deck’s “Solution” narrative (see 2026-08-15-directions-acquisition-wedge-and-deck-structure’s Solution-chapter action item). It sits underneath the “AI Native Insurance Broker for SMBs” vision in smb-insurance-portfolio-brokerage and the three-axis (customer / carrier / office) framing raised there.

Key points

  • Job 1 — Qualification & Exposure Map: first-meeting discovery that turns a business conversation into an exposure map (who the business is, what it’s exposed to, risk appetite/budget), using website scraping, short KYC, and pulling existing policies (via canopy-connect) to run scenarios toward a holistic coverage solution.
  • Job 2 — Intake & Data Gathering: treated as legal-adjacent — structured evidence/claims-data collection ahead of possible litigation. Current lean: gather the information and refer out to lawyers for a revenue share rather than represent clients in court (a suggestion attributed to alon-huri).
  • Job 3 — Market Selection & Submission: a judgment job — deciding which carriers (of e.g. ~100 relationships) are eligible for a given risk based on KYC/context, then submitting; failure mode is declination from submitting to an ineligible carrier.
  • Job 4 — Quoting: distinct from market selection — mechanical retrieval of actual price numbers via self-service portals once submissions are in, to build the client-facing quote.
  • Job 5 — Binding, Onboarding, Billing & Certificates: post-bind QA of policy documents; routing carrier invoices (agency-bill vs. direct-bill) into the platform so clients see billing in one place; and same-day automated Certificates of Insurance (COI) generation — COIs run ~400M/year in the US and are otherwise produced manually on request.
  • Job 6 — Endorsements & Mid-term Changes: framed as “the living policy” — ongoing policy changes, AMS updates, and continuous policy checking; ~30% of inbound tickets are change requests, with under 50% after-hours answer rates today, an explicit service-quality opening.
  • Job 7 — Claims Advocacy & Decoupling: the moment clients learn a broker’s value (first notice of loss, coverage positioning, expectation-setting, carrier chase) — but some clients file directly with the carrier (e.g. Menorah) bypassing the broker entirely, so V1 may just index/route to the right claims channel rather than handle claims in-platform; full ownership is a later decision.
  • Job 8 — Renewal Prep & Retention: an annual per-account 90-day-out cycle (reminders, remarketing, client updates, change reconciliation) across hundreds of accounts on staggered dates; anchored by 85–95% retention and 10–12% renewal commission economics — consistent with the more rigorously sourced figures in us-insurance-distribution-economics (~90% retention, 10–20% commission).
  • Job 9 — Commission Reconciliation: named “the silent revenue leak” — agencies frequently fail to collect commission they’ve actually earned because matching carrier statements against the agency’s books is a painful manual process; pain is “invisible until quantified,” making automated statement-to-book matching a candidate operational-alpha feature.
  • Investor narrative structure: opening statement → full jobs-to-be-done map presented at once (to avoid vertigo) → grouped into clusters (e.g. customer / carrier / office, Venn-style, with many jobs touching multiple groups) → zoom-in comparison on select jobs (old world: red, wasteful, unpleasant; Actually: efficient) → value summary spanning growth, margins, efficiency, speed, and better coverage (the team referenced “six” value points but only five were named in-session — sixth (unknown — needs source)). Explicit guardrail (Saar): stay at the jobs/value level for investors, do not drop into microservices/technical detail.
  • Naming note: the product is referred to consistently as “Actually” throughout this session, superseding the “Ainsure” working name used as recently as 2026-08-10-directions-huri-pitch-deck-build and 2026-08-15-directions-acquisition-wedge-and-deck-structure. No session in the wiki yet documents the rename decision itself — flagged for the team to confirm/log explicitly.
  • Open terminology gaps that risk mis-scoping jobs 3–4: “binder,” agency-bill vs. direct-bill, and portal-raking — flagged for expert clarification (Guy or an external broker) before locking V1 scope. Operator validation, 2026-08-17. The map was drafted from first principles on 2026-08-16; the next day it met two real brokers (greg-ehly, jim-coronado). What the interviews changed:
  • Jobs 2 and 3 (market selection, quoting) are one knowledge problem, not two workflow problems. Both brokers select market by tacit carrier-appetite knowledge before any data entry, because carrier portals accept a full submission and decline only on the final screen. Neither runs a commercial comparative rater. Promoted to its own hypothesis: carrier-appetite-knowledge-is-the-broker-bottleneck.
  • The map understates servicing. greg-ehly named service and billing — payments, invoicing, COIs, insurance cards, cancellation notices — as the dominant consumer of his day, above quoting and selling, and estimated he could take 15–20 more customers/month with no extra headcount if it were removed. Any jobs map weighted toward acquisition mis-weights where broker hours actually go.
  • Binding splits in two, and the map treats it as one job. Direct-appointment risks quote and bind same-day and auto-download into the AMS; wholesaler risks take 3–4 days to quote, 1–2 weeks to bind, and are re-keyed by hand. Roughly 50/50 of one broker’s SMB commercial book. See broker-operational-tech-stack.
  • Onboarding starts with a manual document handoff. Neither broker can see a prospect’s existing coverage — it arrives as a texted or emailed dec page, and a broker cannot see a policy written by another broker even at a carrier they sell.
  • Renewal is more defended than assumed. greg-ehly’s agency touches commercial accounts 60 days out, issues renewal ~45 days out, and proactively remarkets claim-free double-digit increases, absorbing lower commission for retention. Consistent with H-INS-3’s >90% incumbent save rates.
  • A human escalation channel sits inside the quoting job. The decisive appetite call goes to a carrier marketing rep, who then introduces an underwriter. Not represented in the map as drafted.

Evidence

Open questions

  • Has “Actually” formally replaced “Ainsure” as the venture name, and if so, when/why was that decided? No logged decision found.
  • Where exactly does claims handling get decoupled vs. owned in V1 — deferred in-session.
  • What is the sixth investor-narrative value point (only five of six were named)?
  • What do “binder,” agency-bill/direct-bill, and portal-raking precisely mean in this team’s target workflow — pending expert clarification.
  • Should the map be re-weighted by broker hours rather than by sequence? On the 2026-08-17 evidence, servicing dominates and acquisition does not.
  • Does the map need to fork jobs 4–5 by channel (direct-appointment vs wholesaler)? The two paths differ by roughly an order of magnitude in cycle time and by 100% in manual data entry.
  • Where does the carrier marketing-rep relationship live in a jobs map for an AI-native broker that has no such relationships yet?
  • Count discrepancy to reconcile: Nizan presented 21 jobs to moshe-tamir on 2026-08-18, while this page holds nine and insurance-technology-vision describes a 20-job machine. Three different numbers are now in circulation, one of them in front of an external advisor. Settle on one before the investor round.
  • Cross-sell asymmetry is not represented in the map. jasmyne-mcdonald can quote personal lines while still on the phone with a commercial client because the commercial intake already collected addresses and driver’s licences — which makes commercial intake a data asset for a second sale, not just a step in one.

Graph View

Backlinks