ElevenLabs Tutorial: Building Outbound Appointment Reminder Voice Agents with ElevenAgents
·2026.11.20 21:00
Key point
A technical guide to building a voice agent that verifies patients via date of birth and updates EHR appointments to reduce the 23% average no-show rate.
Details
- Tutorial for building an outbound appointment reminder agent using ElevenAgents, targeting the 23% average no-show rate and 3% no-show rate achieved via real-time phone calls.
- Architecture supports two trigger paths: a nightly batch for mass reminders and a real-time trigger for same-day changes, both integrating with FHIR R4 resources for identity verification and scheduling.
- Workflow uses 12 nodes to verify patient identity (DOB match) before disclosing details, with tools for rescheduling, canceling, and handling voicemail or opt-outs.
- Testing revealed that switching the LLM from Haiku to Sonnet and moving appointment details from the base prompt to specific nodes resolved identity leaks and failed hand-offs.
- Compliance requires adhering to TCPA healthcare carve-outs (max 1 call/day, 3/week) and HIPAA, with Zero Retention Mode available only for the real-time path.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.