Data we need
Read-only access to your book's asset-level identifiers and underwriting criteria (or a sample subset for the pilot). Every quote is tied to a specific verified module record — defensible on audit.
Keystone for insurers + MGAs · Callisto Bridge
MGA data infrastructure that prices the policy off the verified signal. Not the class rate.
What is MGA data infrastructure? The underlying data + pricing layer a managing general agent runs on top of carrier paper — a per-asset verified record the MGA quotes against, the carrier binds against, and the auditor reconstructs from, without re-keying bordereaux or forensically re-establishing the risk profile after a claim.
Strong industrialized modules quote at 89bps. Weak ones quote at 140bps. The difference isn't art — it's seismic class, KeyScore, and net-zero status, computed off the verified registry record. Tighter pricing, better risk selection, defensible underwriting on an insurance underwriting data platform that licensed MGA partners bind against, with verified asset risk scoring for insurers at the per-unit level rather than the class-rate roll-up.
What's different for you
Three things every carrier conversation opens with.
What stays with your partners
Licensed MGA partners bind on carrier paper. Reinsurance treaty placement stays with licensed reinsurance intermediaries. Keystone is the pricing engine underneath — never a carrier, never binds a policy.
What a pilot looks like
A 50-quote sample against a real modular book. See the premium spread between KeyScore tiers (KS-92 ≈ 89bps · KS-65 ≈ 140bps) before you write a line.
Named pain points MGAs actually surface
Bordereaux data quality, delegated authority compliance data, parametric verification — one substrate.
The current live MGA conversation is not "we need better pricing models." It's bordereaux data quality (carriers rejecting or manually re-keying bordereaux because the underlying records don't tie back to a verifiable source), bordereaux audit cycles that eat weeks of delegated-authority bandwidth, and delegated authority compliance data that can't be reconstructed on demand. Keystone answers all three from the same signed per-asset record: the record the MGA quotes against is the record the carrier receives, and the record the auditor reconstructs — no re-keying, no reconciliation gap, no forensic reconstruction of what was bound and why.
The same primitive supports parametric insurance data verification: when a parametric trigger depends on a verifiable asset condition, the trigger reads the same signed record the MGA priced from and the carrier bound against, closing the trigger-dispute loop before it opens.
What's broken today
Modular construction is priced like generic commercial construction. One class rate. Same premium for a steel-frame seismic-class-A net-zero module as for an aging panelized unit with no QA record. The result: good risks are over-priced, bad risks are under-priced, and adverse selection runs the book.
What Keystone gives you
- Per-module risk record — seismic class, KeyScore, net-zero status, resilience class, audit trail
- Pricing API — submit module + coverage parameters, get rate_bps + premium + expected_loss_ratio + underwriting_margin
- Partner-MGA book stats live — policies in force, premium float, partner commission earned, blended rate, expected loss ratio (all from your licensed MGA's book)
- Defensible audit — every quote tied to a specific verified record, immutable
How the regulated work happens. Keystone is a pricing-engine infrastructure layer, not an MGA, surplus lines broker, or producer. Policies are bound by state-licensed MGA partners on carrier paper. Reinsurance treaty structures, when needed, are placed by licensed reinsurance intermediaries. See the partner-model architecture.
Who this is for
- Reinsurers — Swiss Re, Munich Re, Hannover Re — writing capacity on modular books
- MGAs writing modular construction lines on behalf of carrier capacity
- Captive insurers for large modular owner portfolios
- Parametric players writing resilience-tied parametric coverage
- Specialty carriers entering modular housing as a class
The economics
For the licensed MGA partner: MGA commission on every bound policy (the partner's regular MGA economics), plus underwriting margin from pricing discipline (premium − expected losses − commissions). Keystone receives either a per-policy SaaS subscription or a shared share of the partner MGA's commission for use of the pricing engine. Recurring per-policy revenue for both parties as the modular book grows.
The same KeyScore the lender priced the loan on becomes the same number the insurer prices the policy on. Both signals point at the same verified asset.
Related · Applied AI for insurance MGAs
Submission triage, claims analytics, portfolio-risk monitoring, renewal intelligence.
Same partner-model architecture as Keystone Insurance Rail, extended to any line an MGA writes. Every AI-influenced decision carries hash-chained audit trail — the compliance posture NAIC AI-governance bulletins and state DOIs expect.
GUIDED DEMO WALKTHROUGH FOR YOUR ROLE
Walk the live demo through your lens. A 4-step script tells you which tabs and numbers to look at first.
Open the Specialty MGA / carrier walkthrough on the demo →
No signup, no auth wall. The guided script lives at the top of /demo/; the live demo runs below it.
Next step
Quote a strong vs weak module side-by-side.
The sample workspace ships with three live policies. KeyScore-92 strong module: 89bps. KeyScore-65 weak module: 140bps. Same API, same math, opposite risk profile.