AI & Automation · 2026

Appointment Businesses · Voice AIStudio concept

A voice assistant that answers a salon's phone around the clock, books into the real calendar, texts a confirmation and passes tricky calls to staff.

Project
Studio concept
Discipline
AI & Automation
Tools
Python · Twilio · OpenAI Realtime API · Next.js

Overview

Salons, clinics and workshops lose bookings every time the phone rings while everyone's hands are busy. Callmate is our concept for an AI receptionist that picks up instead: it speaks naturally, understands what the caller wants, checks the real calendar, books or reschedules, and sends a text confirmation. Anything unusual – a complaint, a complicated colour correction – is transferred to a person with a summary of the call so far.

Challenge

The problem to solve

Callers hang up on robots. The assistant had to sound and behave like a good receptionist – quick, polite, sure of the facts – and know when to stop and hand over.

Deliverables

What was produced

  • 01

    Voice assistant for inbound calls with natural turn-taking and interruptions.

  • 02

    Live calendar integration for booking, rescheduling and cancellations.

  • 03

    SMS confirmations and reminders for every booking.

  • 04

    Owner dashboard with live calls, transcripts, call log and summaries.

  • 05

    Transfer rules that pass calls to staff, with a summary of what was said.

Approach

How it was designed

  1. 01

    The assistant only offers times it has just read from the calendar, so it never promises a slot that is gone.

  2. 02

    Every call becomes a short written summary, so the owner can review a day of calls in a minute.

  3. 03

    Transfer is quick and polite, and the person who picks up already knows why the caller is ringing.

  4. 04

    Bookings made by the assistant are marked in the calendar, so the team can see at a glance what it is doing.

Outcome

Where it landed

Conversation design, voice pipeline architecture and dashboard design, with a prototype handling bookings against a test calendar – scoped for a pilot in one salon.

A self-initiated studio concept, not client work – which is why it has no client name, testimonial or result figures.

Process

How the project ran

  1. 1

    1 week

    Call analysis

    Listened to 150 recorded calls; mapped the requests and where they get stuck.

  2. 2

    4 weeks

    Prototype

    Speech pipeline, calendar tools, SMS confirmations and transfer rules.

  3. 3

    2 weeks

    Dashboard

    Live call view, call log, summaries and calendar badges.

Want results
like these?

Tell us about your goals and we'll reply within one business day with first ideas and a free 30-minute discovery call.