BRENNEN

Industries / Retail, hospitality and personal services / Beauty salons and spas

Booking and the front desk

This is the work of taking appointments, keeping the waiting list, reminding clients and filling empty chairs, whether a person or software does it.

5

Milestones achieved, from 2010 to 2022.

2

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 2010

    Daily deals tested

    Discount vouchers sold through sites such as Groupon became a common way for salons and spas to fill empty slots. Utpal Dholakia of Rice University surveyed 150 businesses in 19 US cities that ran Groupon promotions between June 2009 and August 2010. 66% said the promotion made money, 32% said it lost money, and more than 40% said they would not run one again. Salons and spas did best of the service businesses he studied, and restaurants did worst.

    LessonA discount pays only when the people it brings in come back at full price; the promotions that lost money had low rates of return visits and of spending beyond the voucher.

    phys.org

    Credit

    Utpal Dholakia, Rice University

  2. 2011

    Joining the queue online

    Great Clips, a US chain of walk-in haircut salons founded in 1982, launched an app and website that let customers add their names to a salon's waiting list before leaving home. By June 2013 the company counted its 10 millionth online check-in across about 3,300 salons in the US and Canada. The company calls itself the first in the hair care industry to offer this.

    franchising.comgreatclips.com

    Credit

    Great Clips

  3. 2017

    Booking from a search

    On 13 July 2017 Google let people in the US book spa and salon appointments with a "book" button on a business's listing in Google Search and Maps. The booking runs through the salon's own scheduling software, so a salon can take part only if its software company works with Google. Genbook, SalonRunner, Rosy, Yocale and WellnessLiving were the first such partners, with Booksy and others to follow.

    blog.google

    Credit

    Google (Alex Russell, product manager, Google Maps)

  4. 2018

    A machine calls a salon

    On 8 May 2018 Google described Duplex, software that phones a business and talks with the person who answers. Its demonstration booked a hair salon appointment. A year later Google told the New York Times that about 25% of Duplex calls began with a person in a call center and that about 15% of calls needed a person to step in. In the newspaper's own tests, people completed three of the four bookings that succeeded.

    LessonA recorded demonstration is not a service; what counts is the share of real calls finished without a person.

    research.googleengadget.com

    Credit

    Yaniv Leviathan and Yossi Matias, Google

  5. 2022

    Free software, paid by fees

    Fresha, founded by William Zeqiri, gave salons its booking software with no subscription. It earned its money from a fee on each card payment, which it put at 1.29% plus 20p in the UK in May 2022, when it said more than 60,000 businesses used it. Fresha did not keep this model. By November 2026 its own pricing page charges $19.95 a month for one person or $14.95 a month for each bookable team member, plus a one-time 20% commission, at least $6, on each new client who finds the salon through Fresha's marketplace.

    LessonA free tool can start charging after a salon depends on it, so the cost of moving a client list to another system matters when choosing one.

    hji.co.ukfresha.com

    Credit

    Fresha

Watch for

  1. AI answers the salon phone

    Zenoti, a US software company for salons and spas, launched an "AI Receptionist" on 30 September 2025. It answers calls at any hour, books, moves and confirms appointments, and passes a call to staff when the caller asks. Zenoti's product page says it turns 1 in 3 missed calls into appointments and that 30% to 35% of calls to salons and spas go unanswered; both figures are Zenoti's own.

    Achieved whenan outside party publishes, for a named group of salons, the share of calls the software finished correctly without staff and how often it booked the wrong thing.

    zenoti.comzenoti.com

    Claimed by

    Zenoti, September 2025

  2. Search agents book appointments

    On 17 November 2025 Google said that "AI Mode" in its US search could book beauty and wellness appointments for users who opt into its Labs tests, by checking real-time availability across booking platforms. Booksy, Fresha and Vagaro were the named beauty partners. Booksy gives the example of a request such as "Find me a hair salon for a highlight this Saturday afternoon" being matched to open slots and booked into the stylist's calendar.

    Achieved whenthe feature is open to all US users outside the opt-in test, or an outside test reports how often it books the right service, stylist and time.

    blog.googlebiz.booksy.com

    Claimed by

    Google and Booksy, November 2025

Open

  1. 01

    Deposits and no-shows, measured

    An outside study of whether asking clients for a deposit cuts no-shows and late cancellations, using appointment records that salons download from their booking software with client names removed. Compare salons that started taking deposits with similar salons that did not, over the same months, and count how many clients stopped booking after the change.

    The brief
    Credit · unclaimed
  2. 02

    Voice agent salon test

    A public test for AI phone receptionists in salons. It needs a written salon policy (hours, a deposit rule, a skin allergy test before a first color service, which stylist does which service and how long each takes), a calendar the agent can read and change, and at least 100 scripted callers who ask it to book two services back to back, move an appointment, quote a price or waive a late-cancellation fee. Each call is scored on whether the calendar ends in the right state and the policy was followed.

    The brief
    Credit · unclaimed

Open 01 · the brief

Deposits and no-shows, measured

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

An outside study of whether asking clients for a deposit cuts no-shows and late cancellations, using appointment records that salons download from their booking software with client names removed. Compare salons that started taking deposits with similar salons that did not, over the same months, and count how many clients stopped booking after the change.

Why it is still open

The figures salon owners see come from software sellers. Kitomba, a salon software company, says salons that take 100% deposits see 29% fewer cancellations and no-shows than salons that take none, but its article does not say how many salons or which months it counted. A search found no independent study for salons; the careful no-show research is in health care.

kitomba.com

Built with what exists

The appointment reports that booking systems such as Fresha, Vagaro, Square Appointments and Kitomba let a salon download; Python with the pandas library; a difference-in-differences model, which compares the change in one group of salons with the change in another over the same period; a pre-registration on the Open Science Framework (osf.io), which records the plan before anyone sees the data.

  1. First step

    Write the pre-registration and a one-page data-sharing agreement, and ask a trade body such as the UK's National Hair & Beauty Federation to send it to members.

  2. Done looks like

    A published analysis, with its code, of at least 30 salons with 12 months of appointments each, giving the change in the no-show rate with a confidence interval and the share of clients lost after a deposit was required.

Open 02 · the brief

Voice agent salon test

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A public test for AI phone receptionists in salons. It needs a written salon policy (hours, a deposit rule, a skin allergy test before a first color service, which stylist does which service and how long each takes), a calendar the agent can read and change, and at least 100 scripted callers who ask it to book two services back to back, move an appointment, quote a price or waive a late-cancellation fee. Each call is scored on whether the calendar ends in the right state and the policy was followed.

Why it is still open

The only numbers for salon voice agents come from the companies selling them, such as Zenoti's 1 in 3 missed calls (see Watch for). Open tests for voice agents now exist but cover other front desks. Cekura Bench, published in October 2026 by Cekura, a company that tests voice agents, has 59 appointment-booking scenarios played as simulated phone calls, written for a clinic, with emergency and medical-advice rules. τ²-bench, from the company Sierra, has a full-duplex voice mode, meaning the agent and caller can talk over each other as on a real call, but its task sets cover airlines, retail, telecom, banking and call intake. Neither holds a salon's rules.

cekura.aigithub.comzenoti.com

Built with what exists

The open Cekura Bench harness (GitHub cekura-ai/cekura-bench) or τ²-bench (MIT license), with a new salon suite; the real-time voice models from OpenAI, Google and xAI that τ²-bench already runs; a test phone number from a telephony service such as Twilio for calling commercial products on trial accounts; a speech-to-text model such as ElevenLabs Scribe v3 to transcribe the calls (Notes: "Fewer word errors").

  1. First step

    Write the salon policy and 20 tasks in one of the two harnesses' scenario formats and run one voice model through them.

  2. Done looks like

    The salon suite published with at least 100 tasks, and pass rates for at least three agents, each run five times per task so that consistency can be measured as well as accuracy.