AI Referral Management Agent

How Cassidy automates this using AI

Step 1: Ingest and classify incoming referrals

The Workflow triggers when a new referral arrives—whether via fax, EHR order, or provider portal. Cassidy automatically classifies the document type, separates multi-page faxes, and routes referral documents into a centralized queue.

Step 2: Extract and validate referral data

Cassidy's AI extracts key fields including patient demographics, referring provider, diagnosis codes, requested specialty, urgency, payer information, and authorization status. Low-confidence extractions are flagged for Human-in-the-Loop review before proceeding.

Step 3: Match patient and check eligibility

The Workflow matches the referral against existing patient records, resolves duplicates, and runs real-time eligibility verification to confirm coverage and benefits for the requested services.

Step 4: Apply triage rules and route to the right provider

Cassidy applies your organization's triage protocols—factoring in service line, clinical urgency, geography, language preferences, and provider capacity—to route the referral to the appropriate specialist and location.

Step 5: Coordinate scheduling and patient outreach

The Workflow initiates patient contact via SMS, email, or phone with self-scheduling links and pre-visit instructions. When self-scheduling isn't available, Cassidy routes appointment requests to centralized schedulers with all relevant context pre-filled.

Step 6: Update referrers and close the loop

Cassidy sends status updates to referring providers at each milestone—received, scheduled, completed—via EHR integration, portal, or fax. The Workflow tracks aging referrals, escalates stalled items, and captures outcomes to maintain a complete audit trail.