n8n AI Voice Receptionist Automation

I need an n8n-based flow that turns every incoming call to my medical clinic into a smooth, AI-handled conversation. The system must pick up the phone 24/7, understand the patient’s intent in natural language, and then guide them through booking, rescheduling, or cancelling an appointment. During the call the agent has to pull real-time availability from both Google Calendar and our in-house scheduling database, confirm a suitable slot, collect the caller’s name, number, reason for visit, and preferred doctor or time, and then lock the appointment in while the patient is still on the line. Immediately after, it should send a confirmation through both SMS and WhatsApp, and alert the front desk team if anything sounds urgent or if the caller specifically asks to speak with a human. All appointment data needs to be saved to Google Sheets and mirrored into PostgreSQL (or directly into our clinic management API if you prefer to hit that endpoint). I also want a simple logic branch that escalates the call to a live receptionist whenever confidence drops below a set threshold or the caller requests it outright. Deliverables I will review against: • An n8n workflow file (JSON) with clearly labelled nodes for voice input, NLU processing, scheduling look-ups, confirmation messaging, data storage, and escalation. • Documentation or short Loom covering environment variables, external services (e.g. Twilio Voice/SMS, WhatsApp Business Cloud, OpenAI, Google Calendar API, PostgreSQL credentials), and how to redeploy or tweak the flow. • A quick test plan that shows at least three booking scenarios, one reschedule, one cancel, and one fail-over to a human, all passing end-to-end. Please respond with a detailed project proposal outlining the stack you would combine with n8n, any similar automations you have shipped, and estimated turnaround time for an MVP we can trial on real calls.

Project ID: 40589328