AI Subrogation Opportunity Flagger Agent

How Cassidy automates subrogation opportunity flagging using AI

Step 1: Trigger on claim file updates

The Workflow activates when new documents arrive in the claim file—police reports, adjuster notes, repair estimates, or settlement data—pulling structured and unstructured content from your claims system.

Step 2: Extract and classify key evidence

Cassidy processes documents using intelligent extraction to identify fault indicators, party information, vehicle details, citations, and repair line items, normalizing data for consistent analysis.

Step 3: Apply liability rules and score opportunities

The Agent applies your carrier's liability heuristics and jurisdictional rules—comparative negligence, no-fault limitations, waiver of subrogation clauses—to assign fault ratings, calculate expected recovery value, and generate confidence scores with explainable rationales.

Step 4: Assemble the evidence packet

Cassidy compiles a complete demand package including the liability narrative, police report, photos, repair estimates, storage and rental documentation, proof of payment, and deductible treatment—validated against E-Subro Hub and arbitration requirements.

Step 5: Route for specialist review

Flagged opportunities are routed to subrogation specialists with priority rankings, EV scores, and deadline alerts. Human-in-the-Loop approval ensures specialists review recommendations before demands are submitted.

Step 6: Track outcomes and refine

Cassidy logs recoveries, denials, and arbitration results, feeding outcomes back to improve scoring models and audit rules over time—reducing leakage and tightening cycle times with every claim.