AI Insurance Pre-Authorization Agent

How Cassidy automates this using AI

Step 1: Trigger on order entry

The Workflow activates automatically when a provider places an order in the EHR—whether for imaging, procedures, DME, or referrals—capturing patient demographics, coverage details, diagnosis codes, and procedure codes.

Step 2: Query coverage requirements

Cassidy calls the payer's Coverage Requirements Discovery (CRD) service, interpreting the response to determine if prior authorization is required and what documentation the payer needs for the specific member, plan, and service.

Step 3: Assemble clinical documentation

Using Documentation Templates and Rules (DTR) logic, Cassidy searches the Knowledge Base and connected EHR systems to auto-populate questionnaires with existing clinical evidence—vitals, diagnoses, imaging results, clinical notes—flagging only missing elements for staff review.

Step 4: Submit authorization request

Cassidy constructs a compliant PAS bundle containing patient, provider, service, and clinical attachment data, then submits it to the payer's Prior Authorization API for real-time adjudication.

Step 5: Monitor and route decisions

The Workflow tracks payer responses—approved, pended, or denied—and routes outcomes accordingly: approvals push PA numbers to scheduling and claims, pended requests trigger automated retrieval of additional documentation, and denials surface for staff review with denial reason codes.

Step 6: Update systems and notify stakeholders

Cassidy writes the authorization status and PA number back to the EHR, notifies relevant team members via Slack or Teams, and logs a complete audit trail for compliance and reporting.