AI Subrogation Recovery Agent

How Cassidy automates this using AI

Step 1: Trigger on new claims data

The Workflow activates when medical claims, 835/837 transactions, or ADT feeds arrive scanning ICD-10 codes, provider types, and place-of-service indicators to flag potential third-party liability cases in near real-time.

Step 2: Enrich and validate the case

Cassidy pulls from the Knowledge Base to verify plan language (first-dollar priority, made-whole disclaimers, equitable lien provisions), confirms self-funded ERISA status, and enriches the file with public records like police reports or workers’ comp databases.

Step 3: Build the episode and calculate the lien

The agent aggregates related claims across the injury episode, flags pre-existing conditions, and maintains a dynamic, auditable lien ledger that updates automatically as new claims post.

Step 4: Generate and send notice letters

Cassidy drafts ERISA-compliant reimbursement assertions including plan name, date of loss, and source of rights then routes them to adverse carriers, plaintiff’s counsel, and lien-resolution groups with certified delivery tracking.

Step 5: Support negotiation and settlement

The Workflow surfaces plan language, policy limit disclosures, and fee allocation calculators to guide negotiations. Human-in-the-Loop approval ensures final settlement terms align with fiduciary duties before fund disbursement.

Step 6: Reconcile and report

Once funds are received, Cassidy posts recoveries against paid claims, coordinates with stop-loss where applicable, and updates dashboards with PMPY, cycle time, and recovery-per-file metrics for continuous improvement.