AI-Powered Voice Receptionist for Multi-Location Medical Clinics

Booking • Support •SMS • Email

Case Study · AI Voice Receptionist
Write to our Agent & Get personal pricing
24/7 Availability
3 Clinic Locations
100% Bilingual EN/ES
0 Hold Time
The Client

CMC WeCare — Primary Care & Neurology, Central Florida

CMC WeCare operates three outpatient clinics across Kissimmee, Tampa, and Lakeland, providing primary care and neurology services including general follow-up care, Dr. Vernier neurology consultations, telemedicine visits, and specialized EEG diagnostics. Their patient population is diverse, with a significant Spanish-speaking segment.

Each location runs on its own schedule with different providers and availability windows — making it hard to maintain consistent, error-free phone scheduling through a human-only front desk.


The Challenge

A Single Point of Failure for Every Patient Call

Before the AI agent, every scheduling request, cancellation, and prescription inquiry had to flow through a human staff member — during business hours only. The result was a daily cascade of operational problems.

Missed Calls & Lost Patients

When the desk was busy, calls went to voicemail. Evenings, weekends, and holidays were dead zones — patients simply called a competitor.

Staff Overload on Routine Tasks

The majority of calls were routine: a follow-up, a cancellation, a refill. High-value work was constantly interrupted by tasks that follow the same script every time.

Scheduling Complexity & Errors

Three locations, multiple providers, different hours, new vs. returning patient rules — errors led to double-bookings and frustrated patients.

Bilingual Demand

Not all staff are fully bilingual. Spanish-speaking callers faced longer calls and felt underserved, pushing some toward competitors.

No After-Hours Service

Patients needing to cancel after hours had no option but to call back the next day — often forgetting, creating no-shows that left slots empty.


The Solution

Meet Alex — A Receptionist Powered by AI

"The goal wasn't to build a chatbot. It was to build a receptionist that happens to be powered by AI — one that patients can't distinguish from a well-trained human on the phone."

Alex answers every inbound call instantly, identifies the caller's intent, follows the same structured workflows a trained human receptionist would, and executes real actions — booking appointments, sending confirmations, handling cancellations, and routing prescription requests to clinical staff.

1

Immediate, Natural Answer

Alex picks up on the first ring with a natural greeting. No phone tree, no hold music — just a conversation.

2

Automatic Intent Detection

The caller states their need in plain language. Alex classifies it — schedule, cancel, reschedule, prescription refill, or general inquiry — and proceeds accordingly.

3

One Question at a Time

Alex collects only what it needs, confirms each answer before moving forward, and reads back details phonetically to catch errors before they happen.

4

Automatic Language Switch

If the caller's first words are in Spanish, Alex responds in Spanish for the entire call — including email confirmations and date formatting — using formal usted address.

5

Instant Confirmation

Every completed action — booking, cancellation, reschedule, or refill — triggers an email and SMS confirmation to the patient immediately.


Technical Implementation

Built for Production. Designed for Healthcare.

The solution integrates six best-in-class platforms into a single, HIPAA-aware voice experience.

Voice Layer

ElevenLabs Conversational AI

Real-time speech-to-text and text-to-speech with a human-sounding voice and natural filler phrases to eliminate dead silence.

Reasoning Engine

Anthropic Claude Haiku 4.5

Sub-second reasoning latency optimized for voice. Temperature 0.3 for accuracy; max 150 tokens for natural phone-friendly responses.

Backend API

FastAPI on AWS EC2

Exposes 11 tool endpoints — availability checks, bookings, cancellations, reschedules, confirmations, and escalation emails.

Scheduling

Google Calendar API

Live calendar queries for each provider and location, with server-side rule enforcement before any slot is offered to the caller.

Notifications

Microsoft 365 + Twilio

Transactional emails via Microsoft Graph API and SMS confirmations via Twilio — both triggered automatically at call end.

Records

MySQL + Admin Portal

Every call transcript, summary, and booking reference stored and searchable in a secure admin portal for office managers.

Per-Location, Per-Provider Scheduling Rules

Alex enforces the clinic's real scheduling logic — not a generic calendar picker. Every slot offered is one that genuinely exists.

Kissimmee

Mon–Fri · 08:00–16:00
Dr. Valentine, Resource 1, Resource 2

Tampa

Resources: Tue & Fri · 08:00–16:00
Dr. Vernier: Wed only · 08:00–14:30

Lakeland

Mon–Thu · 07:00–16:00 general
06:00–18:00 for EEG appointments


Key Features

What Alex Delivers on Every Call

🕐

Always-On Availability

24/7/365 — no missed calls, no voicemail, no dead zones on evenings or weekends.

🌐

Automatic Bilingual Support

No "press 2 for Spanish." Language detection is instant, and the entire call — including confirmations — adapts accordingly.

📍

Multi-Location Scheduling

Books across all three clinics, routing callers to the right provider based on patient type, insurance, and location.

👨‍👩‍👧

Booking for Others

Parents and caregivers can book for family members. Patient and caller details are collected separately.

💊

Prescription Refill Routing

Refill requests are validated, collected, and emailed directly to the clinical team as structured data — not verbal relay.

🔁

Fallback & Escalation

If any tool fails, Alex collects caller details and emails the team. No caller is ever left without a resolution path.

Read-Back Confirmation

Before every action, Alex reads back all collected details — including email addresses letter by letter — and requires explicit confirmation.

🔒

HIPAA-Aware Data Collection

Collects only the minimum required. Never requests SSNs, insurance IDs, diagnoses, or medical history.


Results & Impact

What Changed After Deployment

Since going live, CMC WeCare has transformed patient intake and scheduling across all three locations.

Operational

  • Zero missed calls — every inbound call answered within seconds
  • Front-desk staff freed from routine scheduling to focus on in-clinic care
  • After-hours bookings and cancellations captured automatically
  • No-show rates reduced as patients self-service changes at any hour
  • Consistent scheduling rules applied across all three locations

Patient Experience

  • Immediate answer on the first ring — zero hold time
  • Bilingual service available to all Spanish-speaking patients
  • Instant email and SMS confirmation after every interaction
  • Family bookings handled naturally — no awkward workarounds

Staff & Management

  • Complete call logs with transcripts and collected data in the admin portal
  • Every unresolved call generates a structured follow-up email
  • Prescription refills arrive as formatted emails, not verbal relay chains
  • Managers can complete any call from the portal without re-contacting the patient

Technology Stack

Built on Industry-Leading Platforms

Component Technology
Voice AI PlatformElevenLabs Conversational AI
Language ModelAnthropic Claude Haiku 4.5
Backend FrameworkPython / FastAPI
Cloud InfrastructureAWS EC2 (Amazon Linux)
Calendar / SchedulingGoogle Calendar API
Transactional EmailMicrosoft 365 / Graph API
SMS NotificationsTwilio
DatabaseMySQL
Web ServerGunicorn + Uvicorn (ASGI)
Admin Portal AuthHMAC-signed stateless cookies
TLS / HTTPSAWS + Nginx

Is This Right for You?

This Solution Is Built for Practices That Experience:

Missed calls or after-hours call volume that results in lost patients
Front-desk staff spending 40–60% of their time on routine scheduling calls
A bilingual patient population underserved by English-only phone staff
Multi-location complexity making it hard to maintain consistent scheduling rules
High no-show rates driven by patients being unable to cancel or reschedule after hours
Manual prescription refill relay processes that slow clinical workflows

The solution adapts to any specialty, any scheduling system, and any set of business rules. Implementation typically takes 4–6 weeks from kickoff to live deployment.

See Alex Answer the Phone Live

No slides. No decks. Just a real phone call showing exactly what your patients would experience.

Schedule a Live Demo

"We don't replace your staff. We eliminate the 80% of calls that don't need them."