A Salon Owner's Honest Review: 6 Months With a WhatsApp AI Receptionist

Six months ago I paired an AI receptionist into the WhatsApp Business number for our salon. I wrote down what happened week by week, partly because I am a measurement nerd and partly because I half expected to write a teardown post if it went sideways. It did not go sideways. It also did not go the smooth way the demo videos suggested.

This is the in chair version. What worked, what did not, the numbers behind it, and the bits I wish I had known on day one. There is no marketing department in the room. If a salon owner you know is thinking about this, send them this post and skip the rest of the internet.

1. Why I Bought It

The honest reason is that I was answering WhatsApp messages on the toilet at half nine on a Sunday and my partner had given up on having a conversation with me at dinner. The salon does about thirty bookings a week, mostly through WhatsApp, and we are not a chain. Every cut, colour, and trim runs through one phone, my phone, and that phone never stops vibrating between five and nine in the evenings.

The trigger to actually pay for something was a Saturday in the autumn when I missed three colour enquiries because I was halfway through a balayage. Two of them booked with the salon down the road by the time I read their messages. Each colour is around a hundred and fifty pounds. That single afternoon paid for the AI receptionist for nine months.

I picked the platform on the basis of two things. One, the bot paired into our existing WhatsApp number and I was not asked to set up a new line. Two, I could try it for free for the first thirty days without entering a card. Both turned out to be load bearing decisions.

2. What It Replaced

Before the bot we had a chaotic mix.

• WhatsApp Business app on my personal phone with a smart greeting message.
• A paper diary at the front desk that one of the senior stylists kept up to date.
• An ancient Acuity calendar that two clients out of fifty actually used.
• Voicemail nobody listened to.
• Me, replying to messages while doing a regrowth.

The replacement was not "fire all of these and let the bot handle everything". It was "let the bot do the first reply, the qualifying, and the standard bookings. Keep me in the loop on anything weird". The Acuity calendar stayed because Google Calendar is what we already use and the bot writes to it cleanly.

3. The First Month: Mostly Awkward

The first month was rougher than I expected. Not because the bot was bad. Because I was bad at configuring it. I learned three things the hard way.

The first thing I did wrong was give it our service menu without prices. The bot would happily quote a colour at "around one hundred pounds" because the platform's default price band was generic. We had to redo the menu with explicit ranges. Cut and finish from forty five. Half head highlights from one ten. Full head balayage from one fifty. Once I gave it the real numbers, the quoting stopped feeling weird.

The second thing was not setting stylist names. We have two Sarahs and a Sophie. A client typing "Sarah for Saturday" needed a follow up question, "which Sarah, the senior or the junior". I had not configured that and the bot just picked one. The first Saturday after launch I had three angry clients in the chair when they realised they were not with the Sarah they had asked for. Configurable. Embarrassing.

The third thing was not configuring an emergency keyword for complaints. A client typing "Hi I want to complain about Saturday" got a polite reply offering booking slots for the following week. I noticed three days later when she escalated on Instagram. After that I added "complaint", "unhappy", "refund", and "wrong colour" to the escalation list and the bot started routing those straight to my phone.

By the end of week four the configuration was tight and the bot was actually replying the way I would have, only at three in the morning to a question I did not need to see.

4. What It Got Right

The biggest win is not the booking automation. It is the speed of the first reply.

Before, a client messaging me at eight in the evening waited until eleven for me to reply. By then she had already messaged two other salons. After, she got an acknowledgement in five seconds, the qualifying questions answered in two minutes, and a slot offered in five. The number of "thanks but I just booked elsewhere" replies dropped to almost zero. Speed of first reply is the real product. Everything else is paperwork.

The second win is repeat client management. The bot remembers the regulars. A client who books every six weeks for the same service gets the same stylist offered automatically. The bot reads the calendar and suggests Saturday at the usual time. I did not write any of that logic. It just worked.

The third win is the Sunday and Monday. The salon is closed. The bot is not. Two days a week we used to lose to "no reply, will try Tuesday". Now those two days quietly add another six bookings to the week.

The fourth win is silence. The single biggest difference six months in is that my phone does not buzz at half nine on a Sunday any more. The bot has already handled it.

5. What It Got Wrong

I want to be specific about what failed because the marketing pages never are.

Two name confusion. The two Sarahs problem above. Resolved by configuration but it took three angry clients to find.

Outdated colour pricing. I raised colour prices in February and forgot to update the bot's config. For two weeks it was quoting clients twelve pounds under the new rate. I noticed when a client showed up expecting one twenty and got told one thirty two. We honoured the bot's quote and I felt sick about it for a day. Now the prices live in a spreadsheet that the bot reads from, and my admin updates the spreadsheet when she updates the door price list.

None of these are AI failures in the dramatic sense. They are configuration failures. The bot does what you tell it. Tell it the wrong thing and it confidently gives the wrong answer to your customer. The quality of the bot is mostly the quality of the configuration. Plan an afternoon for setup. Plan another two hours every quarter to tidy.

6. The Real Numbers After Six Months

Here is what six months actually looked like.

• Total inbound WhatsApp messages: about 3,100. About twenty a day on average, peaking at fifty on Saturdays.
• Bookings closed by the bot end to end: just over 400. Roughly sixty five percent of all enquiries that ended in a booking.
• Bookings that needed me to step in: about 130. Mostly colour consultations where the client wanted to send a photo of the result they wanted, or first time clients who needed a chat first.
• Complaints routed correctly to me after configuration fix: every one. Zero missed.
• Bookings made outside opening hours: 47%. Roughly half of everything we now book happens when the salon is closed. That number alone changed how I think about the bot.
• Clients who explicitly asked if I was a bot: 11. Out of more than a thousand active conversations. Of those eleven, none cared once I confirmed it.

The financial side is harder to measure cleanly because we changed our pricing twice in the same period. My honest estimate is that the bot recovered around fifteen to twenty bookings a month that I would otherwise have lost. At an average ticket of seventy five pounds, that is twelve to fifteen hundred a month. The platform costs less than ten percent of that.

7. Would I Do It Again?

Yes, but I would do four things differently on day one.

1. Configure stylist names properly before launch, not after the first angry client.
2. Put prices in a spreadsheet the bot reads from, so updates happen in one place.
3. Set complaint and refund as escalation keywords from the start.
4. Tell the regulars in person before they get a bot reply, just so they know the salon is the same.

The brand voice is the bit nobody warns you about. I spent an hour rewriting the default greeting messages in our salon's actual voice, friendly with a bit of cheek, not the formal "Thank you for contacting our establishment" thing the bot starts with. That hour was the most valuable hour of the whole project. Every client now reads a reply that sounds like the salon, not like a customer service bot.

If you are a salon owner reading this, I would not get one if you are doing fewer than ten messages a week. See the five honest cases where you should not buy an AI receptionist. If you are doing thirty a week or more, the maths is loud and the only real question is which one. See the wider numbers behind WhatsApp bots and small business results.

The Bottom Line

Six months in, the bot is the best hire I never put on payroll. It does not call in sick, does not need a Saturday off, does not get tired at five. It does, however, do exactly what you configure it to do, including quoting last quarter's prices to a paying client. That is on you, not on the AI.

If you are sitting in a chair right now reading this between clients, you already know the messages are piling up. Try the free starter plan, configure it on a Sunday afternoon, and see what your phone feels like by next weekend.

Frequently Asked Questions

Did the AI receptionist actually book real salon appointments?

Yes. In the six month window the agent took just over four hundred bookings end to end. About sixty five percent of all incoming WhatsApp enquiries were closed without me touching the chat. The rest needed a human handoff, mostly for colour consultations and complaints.

Did regular salon clients notice they were talking to an AI?

A handful did. Most did not. The ones who noticed were not bothered because the bot replied in seconds at ten at night, which is what they actually wanted. The few who explicitly asked for me by name were handed off in under a minute. The brand voice mattered more than the bot label.

What did the bot get wrong over six months?

Three things repeatedly. Booking the wrong stylist when a client just said by first name and we have two of the same name. Quoting an outdated colour price for two weeks because I forgot to update the rate after a price rise. Not catching that a complaint was a complaint until the client typed in capitals. All three were configuration fixes, not AI failures.

Was the cost worth it for a single chair salon?

Yes, but only because I was already losing evening enquiries every week. If I had been quiet, the bot would not have paid for itself. The break even calculation for me was three recovered bookings a month. I was easily clearing that by month two.

Would you recommend a salon owner do this themselves?

Yes if you are comfortable with the WhatsApp Business app and willing to spend an afternoon writing your service list, prices, and stylist names properly. The DIY setup is real and works. If you would rather not, the done for you setup is the right call. The mistake is paying for the cheapest plan and never configuring it.