AI Voice Agent That Answers Missed Booking Calls for a Massage Franchise
Let's Connect
Overview
What we built
A 45-location massage franchise was missing roughly one in three booking calls while its front desks looked after the guests in the room. We built a voice agent that picks up every time.
In plain terms: when a spa's front desk is checking in a guest or taking a payment, the phone loses. At peak hours roughly one in three inbound booking calls went unanswered, audits showed callers giving up after four rings and booking elsewhere, and voicemail converted fewer than 15% of the people who left a message. Franchisees were paying overtime just to keep someone near the phone, and calls still slipped through.
We built an AI voice agent that answers overflow and after-hours calls in a natural conversational tone. It checks live therapist availability, books or reschedules appointments directly in the franchise scheduling system, and answers everyday questions about pricing and membership; anything complex is transferred to staff with a summary. Within 90 days the missed-call rate fell from 32% to under 6%, the agent now handles 71% of after-hours calls end-to-end, and the franchise books an estimated 1,100 additional appointments per month across all 45 locations.
The Problem
One in three calls missed
The franchise's problem was not demand, it was simultaneity. Booking calls peaked at exactly the moments front-desk staff were busiest, checking in arriving guests and taking payments at the end of appointments. A therapist's schedule might have plenty of open slots, but if the phone rang while the desk was occupied, that availability may as well not have existed.
The fallback layers failed one by one. Call audits showed callers abandoning after four rings and simply booking with a competitor. Voicemail caught some of the rest, but fewer than 15% of the people who left messages ever converted into an appointment. Franchisees responded the only way they could, paying front-desk overtime for phone cover, which protected some calls at a real and recurring cost.
Across 45 locations the pattern repeated daily, and nobody could see it clearly. Each missed call was invisible in the moment, each abandoned booking left no record, and franchise leadership only knew the aggregate: roughly one in three booking calls at peak hours was going unanswered, and revenue was walking out through the phone line.
Peak-hour collisions
Calls peaked precisely when front desks were checking in guests or taking payments, so the busiest moments were also the ones when the phone went unanswered.
Four rings, then gone
Audits showed callers abandoning after four rings and booking elsewhere, turning a brief wait at the desk into a permanently lost appointment.
Voicemail black hole
Fewer than 15% of missed callers who reached voicemail ever converted into a booking, making the safety net barely better than no answer at all.
Overtime as a patch
Franchisees paid front-desk overtime purely for phone cover, an ongoing cost that still could not keep pace with peak-hour call volume.
What it was costing them
Every unanswered call was a bookable appointment offered to a competitor. With roughly one in three peak-hour calls missed across 45 locations, voicemail recovering fewer than 15% of them, and overtime being spent just to keep a human near the phone, the franchise was paying twice: once in lost bookings and again in labour spent defending against the losses.
The Solution
Always-on AI booking agent
We built the voice agent to behave like a capable colleague rather than a phone tree. It answers the overflow calls the desk cannot reach and every after-hours call, speaks in a natural conversational tone, and gets to the purpose of the call quickly: a booking, a reschedule, or a question about pricing or membership.
The agent works against live data, not scripts alone. It checks real therapist availability before offering times, books or reschedules the appointment directly in the franchise scheduling system while the caller is still on the line, and confirms the details back. There is no callback queue and no message pad; the caller hangs up with an appointment.
We were equally deliberate about what the agent does not do. Complex or unusual calls transfer to staff with a summary of the conversation so far, so the caller never repeats themselves. And every conversation is logged for franchise review, giving each location a record of what its callers asked for and how the agent responded.
Key decisions
Answer overflow, not replace staff
The agent takes the calls front desks physically cannot, overflow at peak and everything after hours, leaving in-person guests with the staff's full attention.
Live availability, real bookings
The agent reads live therapist availability and writes bookings straight into the scheduling system, so callers leave with confirmed appointments rather than callback promises.
A natural voice, deliberately
The agent speaks in a natural conversational tone, because booking a massage is a personal call and callers who feel processed simply hang up.
Warm transfers with a summary
Complex calls hand over to staff with a summary of the conversation so far, so escalation feels like being helped further rather than starting again.
Every call logged
Full conversation logs give franchisees and the franchise office a reviewable record of caller demand, common questions and agent behaviour at every location.
Measurable Impact
What changed after launch
The missed-call problem effectively disappeared: within 90 days the franchise-wide missed-call rate dropped from 32% to under 6%. After hours, when previously every call went to voicemail, the agent now handles 71% of calls end-to-end without any human transfer, and the bookings followed: an estimated 1,100 additional appointments per month across all 45 locations.
The economics at the front desk changed too. Overtime spend for phone cover fell by 38% in the first quarter, because nobody needs to be paid to sit near a phone the agent already answers. Staff give arriving guests their full attention, transfers arrive with context attached, and franchisees can finally review what their callers actually ask for.
Missed calls
Roughly one in three peak-hour calls unanswered
Missed-call rate under 6% within 90 days
After-hours cover
Voicemail converting fewer than 15% of callers
71% of after-hours calls handled end-to-end
Bookings
Callers abandoning after four rings, booking elsewhere
An estimated 1,100 extra appointments booked monthly
Phone-cover overtime
Franchisees paying overtime just for phone cover
Overtime spend down 38% in the first quarter
Headline results
Missed-call rate across the franchise dropped from 32% to under 6% within 90 days
71% of after-hours calls handled end-to-end by the voice agent without human transfer
An estimated 1,100 additional appointments booked per month across all 45 locations
Front-desk overtime spend for phone cover reduced by 38% in the first quarter
Tech & Tools Used
What powered the build
Every tool below earned its place in this engagement. Here is the part each one played.
Twilio Programmable Voice
Carries every call, routing overflow and after-hours traffic to the agent and handling the warm transfer path back to front-desk staff.
Deepgram Speech-to-Text
Converts caller speech to text in real time, giving the agent an accurate live transcript to reason over even with background noise on the line.
OpenAI GPT-4o
Acts as the conversational brain, interpreting what each caller wants, driving the booking dialogue and deciding when a call should transfer to staff.
ElevenLabs Text-to-Speech
Gives the agent its natural speaking voice, so conversations across the franchise's phone lines sound warm and human rather than robotic.
Node.js
Runs the real-time conversation service that stitches telephony, transcription, the language model and the scheduler together within the tight timing a live call demands.
Redis
Holds live call state, keeping track of where each conversation stands so responses stay instant while many calls run at once.
PostgreSQL
Stores the full conversation logs and booking outcomes that franchisees review, building a durable record of caller demand at each location.
AWS Lambda
Executes the event-driven work around each call, including generating the transfer summaries and writing post-call logs ready for franchise review.
Scheduling REST API Integration
Connects the agent to live therapist availability and writes bookings and reschedules directly into the franchise's scheduling system during the call.
Ready to Build your Spa & Wellness Business with AI Voice Agents
Ask Byte
Ask Byte
Typically replies instantly
just Now
Hi! I'm OrganByte's assistant. How can I help you today?
AI-generated content may be incorrect
