AI Auto Damage Claim Triage Agent

How Cassidy automates auto claims document triage using AI

Step 1: Trigger on new claim submission

The Workflow activates when a new FNOL arrives—whether from email, web portal, mobile app, or broker feed—capturing policyholder details, incident narrative, vehicle information, and initial photos.

Step 2: Verify coverage and policy status

Cassidy pulls policy data from your claims management system to confirm coverage, deductibles, exclusions, and prior claims history, flagging any eligibility issues before triage begins.

Step 3: Classify and extract from all documents

The Workflow processes every attachment—FNOL forms, police reports, repair estimates, invoices, statements—using AI to classify document types and extract key fields into a normalized, claims-ready dataset.

Step 4: Analyze photos and detect damage

Cassidy evaluates submitted photos to identify damaged components, predict repairable vs. total loss scenarios, and generate preliminary line-level estimates for appraiser review.

Step 5: Run completeness checks and request missing items

The agent compares the claim file against your completeness checklist by loss type and jurisdiction, then auto-generates targeted requests via email, SMS, or portal with SLA-tracked reminders.

Step 6: Screen for fraud and anomalies

Cassidy flags inconsistencies—timestamp mismatches, duplicate patterns, suspicious providers, inflated line items—and routes high-risk claims to SIU with documented rationales.

Step 7: Score severity and route to the right handler

Based on damage severity, coverage complexity, liability factors, and fraud risk, the Workflow assigns claims to desk appraisers, field inspectors, DRP shops, total loss units, or casualty adjusters—with Human-in-the-Loop approval for high-stakes decisions.

Step 8: Update status and maintain audit trail

Cassidy sends omnichannel status updates to the policyholder and logs every decision, extraction, and routing action with source citations for compliance and QA.