AI Certificate of Coverage Comparison Agent

How Cassidy automates Certificate of Coverage Comparison using AI

Step 1: Trigger on COI submission

The Workflow activates when a certificate arrives—via email, secure portal upload, or API integration from your procurement or project management system.

Step 2: Classify and extract document data

Cassidy identifies the document type (ACORD 25, endorsements, additional remarks) and extracts structured fields: named insured, certificate holder, carriers, policy numbers, coverage lines, limits, effective dates, and endorsement identifiers like CG 20 10 or CG 20 37.

Step 3: Pull requirements from Knowledge Base

The agent retrieves the applicable requirements matrix from your Knowledge Base—coverage minimums, mandatory endorsements, entity name requirements, and project-specific terms stored from your contract language.

Step 4: Validate coverage against requirements

Cassidy compares extracted data to the requirements checklist: confirms limits meet minimums, verifies date overlaps with project terms, checks that endorsements substantiate Additional Insured and Waiver of Subrogation status, and flags name mismatches or missing coverage lines.

Step 5: Generate diff report against prior submissions

For renewals or updated COIs, the Workflow produces a variance analysis highlighting changes in limits, carriers, endorsement scope, aggregate application, and retro dates—with risk impact notes for each change.

Step 6: Route for decisioning or remediation

Compliant submissions update vendor records automatically. Deficient COIs trigger correction requests to the vendor or broker with specific asks—missing endorsement form numbers, required limit increases, or entity name corrections—keeping humans in the loop for exceptions.