Skip to content

The on-ramp / critical path

The concrete, ordered sequence from “Joe has no license” to “Glacis is underwriting Tech E&O for healthcare AI.” This is the project plan the whole academy supports.

The strategy has evolved. The path below is the Volume 1 framing, where Glacis itself operates the MGA and holds the pen (steps 7–8). Volume 2 moves Glacis off the paper: a syndicate writes and holds the book, and Glacis stays the neutral evidence layer that prices it — because the moment Glacis holds risk, its attestations read as a competitor’s format and the standard collapses. Read steps 7–8 as “an MGA/syndicate binds, on Glacis evidence.” See the independence thesis.

1. Joe gets a RESIDENT producer license in the CASUALTY (P&C) line [pre-licensing → exam → background → application] (sections 02, 04-06) 2. Joe EXPANDS to WA / CA / NY (and beyond) as NON-RESIDENT by reciprocity (CA: lock a compliant entity NAME first) (sections 02-06, 10) 3. Form the ENTITY (+ corporate prereqs: SecState, registered agent) and license it with JOE AS DRLP — only in states where Joe is licensed (section 03) 4. Carrier APPOINTMENT(s) (section 02-05) 5. Secure a CAPACITY / FRONTING carrier (+ reinsurance) (section 13-04) 6. Sign the BINDING-AUTHORITY / PROGRAM agreement + UNDERWRITING GUIDELINES (likely SURPLUS LINES / E&S for a novel AI risk) (section 03-05, 03-06) 7. OPERATE AS AN MGA: risk-select, rate, bind Tech E&O — Glacis evidence drives selection + schedule-rating credits → low loss ratio (sections 13, 15-01) 8. The FLYWHEEL turns: profit share → capacity → data → better models → repeat

Each step is a prerequisite for the next — see the delegated-authority stack for the layered view, and the flywheel for what step 7→8 produces.

  • Resident first. Joe must be licensed in his resident state before expanding. GROUNDED · Westmont
  • DRLP gates the entity. The entity can only be licensed where Joe (the DRLP) is licensed. GROUNDED · Westmont
  • License + appointment before binding. No risk can be bound until both exist and authority is granted. STABLE
  • Casualty line throughout. Tech E&O is casualty, so every license (Joe’s and the entity’s) must include Casualty. STABLE
  • Surplus lines likely. A novel AI risk probably writes non-admitted — add a surplus-lines broker license and budget roughly double. GROUNDED · Westmont

Cost & timing (Westmont estimates — subject to change)

Section titled “Cost & timing (Westmont estimates — subject to change)”

Westmont’s figures for a one-person + one-entity countrywide build-out, inclusive of Westmont fees + state application fees (and explicitly labeled estimates, subject to change without notice): GROUNDED · Westmont

Scope Estimated cost Notes
Producer (one person + one entity, countrywide) ~$40,000 – $50,000 Westmont fees ≈ 40% of total
With Surplus Lines added ~$80,000 – $90,000 “roughly double”
Typical timeline ~4 months on average “all licensing efforts are unique”

These figures exclude the cost of a registered agent and other licenses. They are a vendor estimate, not a quote — get a current written quote scoped to the actual states/lines, and cross-check live state fees via NIPR. GROUNDED · Westmont

  • Do it yourself via NIPR/Sircon — cheaper, more hands-on, slower for multi-state.
  • Engage a consultancy (Westmont-type) — faster countrywide, ~40% premium for the service. GROUNDED · Westmont

Glacis is itself a “third party” under the NAIC’s emerging vendor-oversight definition (see 01 third-party WG). So the same documentation discipline that powers the MGA’s underwriting also positions Glacis ahead of its own future regulation — the on-ramp and the moat are the same road. GROUNDED · Law360 GROUNDED · Glacis

  • Keep the 16 VERIFY hub open as your action list.
  • Drill the whole arc in the app: ../../app/run.sh --mode quiz and --mode puzzle (scenario puzzles like “you want to sell healthcare Tech E&O in CA — which licenses/steps?”).

16 VERIFY hub: the checklist

Drill this page →4 bank questions stand behind what you just read. Check it while it’s warm.