My Device Configuration
Hospital CRM — AI Patient Management & Appointment Automation
A 102-node n8n system ("MediCare AI") that runs a hospital's patient communication and appointment lifecycle end-to-end, paired with an HTML operations dashboard.
At the core is a WhatsApp-native conversational agent (built on Meta's WhatsApp Business webhook with full verify-token handshake) with function-calling access to a defined tool set — look up hospital/doctor information, check a doctor's booked slots, book/reschedule/cancel appointments, check diagnostic report status, request an AI-initiated voice call, or hand off to a human at the front desk — backed by per-patient conversational memory and live data in Supabase, with an optional MCP connection layer for real-time hospital data access.
Around that core, the system runs several autonomous operational loops: a 15-minute reminder scanner that sends 24-hour WhatsApp reminders and 2-hour reminders paired with an outbound AI voice call; a daily missed-appointment sweep that triggers AI follow-up calls to no-shows; a voice-call outcome pipeline that parses call results (confirmed / reschedule requested / cancelled), updates records accordingly, and applies retry-count logic that escalates to the front desk by email after repeated failed contact attempts.
A report-readiness pipeline generates a secure, tokenized report-access link, produces a patient-friendly AI summary of clinical reports, and triggers an urgent AI voice call plus doctor/admin email alert for flagged critical results. An hourly feedback scanner requests post-visit ratings, escalates any rating of 3 or below directly to the front desk with an apology message, and nudges satisfied patients toward a Google review.
A dedicated error-trigger workflow catches failures anywhere in the system and alerts staff by email, giving it basic production-grade observability.
How long did it take you to build this?
It took me around 7 weeks to build the system end-to-end, including the n8n architecture, WhatsApp integration, AI agent, appointment automation, voice-call workflows, reporting pipeline, dashboard, and error-handling infrastructure.
How much would you charge for creating something like this?
For a system of this complexity, with conversational AI, WhatsApp integration, appointment management, voice automation, report processing, automated follow-ups, escalation logic, dashboard functionality, and production observability, I would charge somewhere in the range of $3,000–$3,800, depending on the hospital's requirements and the level of customization.
13 Jul 2026
Keywords
AI Usage
The conversational patient agent (appointment booking/rescheduling/cancellation, report-status lookup, human handoff) and the patient-friendly clinical report summarization are both AI-driven, running on Gemini with per-patient conversational memory and a defined function-calling tool set. I designed all of the surrounding business logic manually: the tool schema exposed to the agent, the multi-stage reminder-and-retry system for appointments, the escalation rules (feedback-rating thresholds, call-retry limits, critical-report alerting), the secure report-token generation, and the Supabase data model connecting patients, appointments, reports, and feedback.
