Use Cases¶
How ElevateNow products solve specific insurance operations problems. Each use case follows the same structure: situation → what we did → outcome → artifacts.
Workers' Compensation FNOL¶
Situation: A WC claim arrives. The adjuster needs jurisdiction rules (TTD rates, benefit caps, medical fee schedules, subrogation rules) to assess compensability and set reserves — but jurisdiction rules change annually and vary across 50 states. Manual lookup takes 20–40 minutes per claim and introduces error.
What Adjust360 does:
1. Triage: flag for litigation indicators, attorney representation, catastrophic injury
2. Jurisdiction resolution: state code → jurisdiction_rule chunks in enCODE → benefit rates, deadlines, compliance requirements for that state and year
3. Compensability assessment: injury type, body part, mechanism of injury → compensability verdict with citable basis
4. Reserve calculation: AWW → TTD/PPD computation using jurisdiction-specific rates
5. Authority routing: reserve amount → authority matrix → bind / refer / escalate
Outcome: Deterministic, jurisdiction-aware FNOL packet in under 60 seconds. Every figure cites the governing rule version. Adjuster reviews, approves, moves to next claim.
Covered states: WC — 50 states (Phase 2 expansion via P2-17). Auto — 51 states/DC.
Commercial Underwriting Intake¶
Situation: A GL submission arrives via email. The underwriter needs to assess eligibility, exposure, and completeness before deciding whether to quote. Manual assessment takes 45–90 minutes. High-volume periods create backlogs.
What Clear360 does: 1. Parse and normalize submission (EML → structured payload) 2. Three-gate deterministic eligibility: SIC code lookup → exclusion matrix → vertical appetite → operations check 3. Parallel: exposure analysis (UW-CP-001 cognitive pattern) + loss analysis (UW-CP-002) + completeness check 4. Each agent output cites the governing enCODE chunk version 5. Workbench Hub displays verdict with full decision trace — adjuster sees Trust Mode (auto-process / guided / exception)
Outcome: 93-second eligibility-to-verdict pipeline. Underwriter handles exceptions only; clean submissions auto-process.
Binding Authority Automation¶
Situation: A producing broker submits a risk for binding. The MGA needs to check class appetite, geographic eligibility, scoring gates, and coverage availability before binding — all against program-specific rules that change per-cedant.
What Bind360 does:
1. Class appetite check: SIC/NAICS → ba_class_appetite chunks
2. Geographic restriction: state → ba_geographic_restriction chunks
3. Score gate: premium/exposure metrics → ba_score_gate chunks
4. Coverage gate: requested coverages → ba_coverage_gate chunks
5. Authority routing: all gates → BIND / REFER / DECLINE with per-gate rationale
Outcome: Bind decisions in seconds, inside authority, with a full audit trail per cedant program.
Data Quality — AI/ML Readiness¶
Situation: A carrier wants to train a loss prediction model. The training data contains PII (SSNs, claimant names), has completeness gaps in key fields, and may have identity duplicates across claim and policy systems.
What Assure + Shield + Resolve do: 1. Assure: score training dataset across 6 DQ dimensions; flag fields below threshold 2. Shield: detect and tokenize PII/PHI before data reaches the model training pipeline 3. Resolve: deduplicate party records across systems; produce golden party references
Outcome: A PII-free, quality-scored, identity-resolved training dataset. The CDO can certify to the board that the model was trained on clean data — with the Witness audit trail to prove it.
Adding a use case¶
Use the use-case template from the repo. Follow the situation → what we did → outcome → artifacts structure. Link to demoboards and briefing decks from the Asset Library — do not embed or copy them.