Automation Case Study

Medical Spa AI Voice Agent System

Inbound receptionist and outbound calling agents with scheduling tools, conversation controls and Twilio SIP telephony for medical spa operations.

Retell AI Twilio SIP Trunking AI Voice Agent Appointment Booking Outbound Calling

Medical Spa · Client identifiers removed

Case Study At A Glance

System overview

Use Case
Medical Spa Receptionist
Voice Layer
Retell AI
Telephony
Twilio (SIP Trunking)
Functionality
Inbound + Outbound
Key Features
Booking + Availability
Verified Evidence
Live Agent Configuration

Business Objective

AI receptionist with smart booking logic

The medical spa needed an AI voice system that could answer service questions, check availability, book appointments, and handle outbound follow-up calls — all while maintaining a warm, professional receptionist experience.

Separate inbound and outbound agents were configured to handle different conversation scenarios with consistent scheduling logic.

Implementation Scope

What Talha Rehamni implemented

The implementation combines Retell AI inbound and outbound agents, Twilio SIP telephony, appointment booking functions, conversation tuning, and service-aware scheduling logic.

Retell AI inbound agent configuration
Retell AI outbound agent configuration
Twilio Elastic SIP trunk setup
check_cal_availability function
book_appointment function
Dynamic variable handling
Conversation tuning (response, interruption, denoising)
US phone number assignment in Twilio

Call Flow

Inbound booking and outbound follow-up

The system uses separate Retell AI agents for inbound and outbound calls, connected through Twilio SIP telephony.

01

Inbound Call

Customer calls the medical spa phone number. The call is routed through Twilio SIP trunk to Retell AI.

Trigger
02

AI Receptionist Answers

Retell AI inbound agent answers with a warm, professional receptionist tone. Answers brief service questions and guides toward booking.

Conversation
03

Check Availability

The agent uses the check_cal_availability function to find open slots. Requires user-confirmed date and time before proceeding.

Function Call
04

Book Appointment

After user confirmation, book_appointment is triggered with RFC3339 datetime. Confirmation is sent to the customer.

Action
05

Outbound Follow-up

Separate outbound agent handles confirmation calls, reminders, and callback handling using dynamic caller fields.

Outbound

Agent Comparison

Inbound vs Outbound agent design

Inbound Agent

  • Answers service questions with warm receptionist tone
  • Uses check_cal_availability function
  • Uses book_appointment function
  • Requires user-confirmed date/time before booking
  • Interprets time in US Eastern Time
  • Call-center background sound enabled

Outbound Agent

  • Uses dynamic caller fields (name, email, phone)
  • Checks if it's a good time to talk
  • Offers callback path when timing is not suitable
  • Internal timezone rules for consistency
  • Connected through Twilio Elastic SIP Trunk

Speech & Call Tuning

Beyond prompt configuration

Response timing, interruption behavior, denoising, transcription mode and call controls are explicitly tuned inside Retell AI.

Response Behavior

Responsiveness and interruption sensitivity are set explicitly rather than left as generic defaults.

Audio Environment

The inbound agent is configured with a call-center background sound setting for optimal clarity.

Transcription

Noise removal is enabled and the transcription mode is set to optimize for speed.

Keypad Handling

User keypad input detection is enabled with a configured timeout for seamless DTMF interaction.

Technology Responsibilities

The role of each platform

Voice Agent

Retell AI

Inbound receptionist and outbound calling agents with booking functions, variable extraction, and conversation tuning controls.

Telephony

Twilio

Elastic SIP trunk for PSTN routing, origination and termination connectivity, and US phone number assignment.

Verified Outcome

A complete AI voice receptionist system

The implementation demonstrates Retell AI inbound and outbound agent configuration, prompt engineering, dynamic variables, service-aware booking logic, and Twilio SIP telephony.

Retell AI inbound agent with booking logic
Retell AI outbound agent with identity confirmation
check_cal_availability and book_appointment functions
Twilio Elastic SIP trunk with US phone number
Response timing, interruption, denoising and transcription tuning
Dynamic variable handling for current time and caller fields
Evidence and performance note

Call volume, booking conversion, revenue and other performance metrics were not supplied. This case study focuses on the implementation evidence and verified agent configuration.

Technical FAQ

About this AI voice receptionist

How does the AI receptionist book appointments?

The inbound agent uses two callable functions: check_cal_availability to verify open slots and book_appointment to schedule. Both functions require a user-confirmed date and time before triggering, preventing premature bookings.

What's the difference between inbound and outbound agents?

The inbound agent handles incoming calls for service questions and bookings. The outbound agent handles follow-up calls with dynamic caller fields for identity confirmation, timing checks, and callback paths when the timing isn't suitable.

How is Twilio configured for this system?

Twilio is configured with an Elastic SIP trunk named for Retell integration. It handles origination (inbound SIP traffic toward Retell) and termination (outbound PSTN routing) with a US phone number assigned under the trunk.

What conversation tuning was applied?

Response behavior, interruption sensitivity, call-center background sound, noise removal, transcription mode (optimized for speed), keypad input detection with timeout, silence handling, and maximum call duration controls were all explicitly configured.

How are dates and times handled consistently?

Current-time context is passed through a dynamic variable and interpreted in US Eastern Time. Internal date/time rules specify an absolute RFC3339 datetime with offset for function calls, ensuring consistent interpretation.

Related Automation Services

Need a similar AI voice receptionist system?

Talha Rehamni builds AI voice agents, Twilio telephony integrations, and conversational booking systems for medical spas, clinics, and service businesses.