Insurance Claims & Rating

What this covers

• —FNOL intake and claim triage decision tables
• —Multi-stage adjudication ruleflows
• —Fraud-screen / SIU routing by amount, injury, and policy age
• —Warehouse matched[] traces and content-hash provenance

Outcomes for this persona

What a successful evaluation looks like in production — not brochure metrics.

• —FNOL and adjudication changes ship under governance, not shadow spreadsheets
• —Every claim path is reproducible from matched[] traces and content-hash provenance
• —Business authors edit tables; engineering owns the model, not the claim policy

Insurance claims purpose story

Separate claim policy from FNOL and adjuster apps: authors own triage routes, severity gates and SLA bands as decision tables and ruleflows; intake and portals call one published service.

1. Model the claim contract — amount, injury flag, policy age, claim type, region — as decision-table inputs (BOM/JSON vocabulary, not a Java XOM JAR product).
2. Author FIRST-hit triage rules (injury → SIU, large+new policy → fraud-review, large → senior adjuster, else fast-track) with explicit reason codes and SLA hours.
3. Compose multi-stage ruleflows for adjudication packs when one table is not enough — intake → enrich → act — still single-pass Decision Manager, not Rete working memory.
4. Publish a version under governance and deploy to an environment; hot/canary and blue-green when you promote further.
5. Execute from Decision Server / HTDS-style apps; inspect matched[] and content-hash provenance in Decision Warehouse for every claim path.

Trial seed: Free trial signup still seeds Sample · Loan Pre-Approval (not a claims ruleset). In Manager, start from the Live template [Demo] Claims triage, or use the enterprise showcase Insurance Claims packs (fnol, adjudication, fraud-screen, settlement, recovery) with the claims BOM. The /demo claims scenario walks the same triage inputs illustratively — the trial and showcase run the real engine.