AI Patient Communication Drafting Agent

How Cassidy automates this using AI

Step 1: Trigger on incoming patient request

The Workflow activates when a patient message arrives—whether through portal chat, SMS, email, or phone transcription—and immediately captures channel, language preference, and raw message content.

Step 2: Verify identity and match to patient record

Cassidy runs HIPAA-safe authentication checks (DOB, phone, one-time codes) and matches the patient to your MPI/EHR record, ensuring every response is tied to the correct individual.

Step 3: Detect intent and assess urgency

The agent classifies the request type—scheduling, refill, billing inquiry, test results, or clinical question—and flags urgency cues like distress language or symptoms requiring immediate attention.

Step 4: Draft empathetic, compliant response

Cassidy pulls from your Knowledge Base of approved templates, tone guides, and organizational policies to generate a patient-friendly message with appropriate reading level, bilingual options if needed, and structured fields for any required EHR data.

Step 5: Execute transaction or route to human

For routine requests, Cassidy completes the action (books appointment, submits refill, sends payment link) with real-time EHR writebacks. Clinical or complex matters route to the appropriate queue—nurses, schedulers, billing specialists—with a full context summary attached.

Step 6: Log interaction and update records

Every touchpoint is documented: transcripts, timestamps, identity verification results, routing decisions, and outcomes sync back to your EHR and CRM, creating an auditable trail and feeding KPI dashboards for ongoing optimization.