The AI receptionist for salons and gyms that actually books the visit
DMs, WeCom and website enquiries land while your front desk is with a customer. TOPPP turns the playbook of your best closer into a team that answers around the clock.
They came. They cooled off in these five steps
A local sale doesn’t happen on one page. It happens across five steps, and each one leaks in the same place.
01Step 01 / First enquiry
Nobody replies for minutes
DMs, WeCom and website enquiries arrive while the front desk is with a customer. No reply, and the next shop gets the question.
02Step 02 / Booking the visit
The price question ends it
You quote, they say they’ll think about it. The line that turns that into a booked visit is one only a few people know.
03Step 03 / No-show follow-up
Nobody nudges them
The slot passes, nobody walks in. Whether anyone nudges them that day depends on who is on shift.
04Step 04 / Closing in person
This step belongs to people
The real close happens in the room, with your manager and senior staff — when they have time. The first three steps ate it.
05Step 05 / Renewals
You notice when it expires
Packages run out, memberships lapse, regulars drift away. Nobody watches those dates, so a loyal customer becomes a paid lead.
Four stages, on a shop floor
The engine is shared, the business logic is yours: your playbook, your roles, your standards, your limits — one system.
Codify the playbook
The opening, the way you quote, the objection answers, the line that books the visit — captured as your own SOP.
Scale the execution
Booking, re-engagement and Q&A each get a role, a lead agent routes the work, and it runs around the clock.
Compound the learning
The review workbench checks quality before go-live; after it, problem conversations feed a loop. Fix once, the team updates.
Humans where they matter
You draw the line on discounts, big packages and refunds. Cross it and the chat goes back to a person, unnoticed.
Custom does not mean built from scratch. Your shop, your roles, your standards, your limits — configured on one system.
It runs your playbook because the rules make it
“Follows the SOP” isn’t a description. It’s five hard rules in the product, and you can see each one in the configuration.
Stages have entry conditions
From first enquiry to booked visit, you set what each step requires. Unmet, and it doesn’t move on — no skipping ahead.
No criteria, no pass
Every case is scored before go-live. A case with no criteria to judge it won’t pass, and a failing version doesn’t ship.
Changes ship as versions
Nothing is edited in place. A version is saved, reviewed, then applied — and rollback is one step.
One switch per customer
Not a global toggle. Crossing a line hands back that one customer; every other chat runs on, backstage.
Every turn is on record
Each turn’s reasoning and tool calls are kept and replayable turn by turn — if it gets it wrong, you find the step.
Some things don’t go to AI
Limits live in the configuration, not in the moment. You decide where each one sits.
Discounts and big packages
- Discounts and big packages：Pricing authority stays with you; past the amount or discount you set, the chat goes to a person on the spot.
- Health and safety calls：Injuries, allergies, recovery — those calls belong to trainers and specialists. AI only gets people in front of them.
Refund disputes
- Refund disputes：Anything about a contract or money goes to a person; AI takes the full account and hands it to the manager.
- Outcome promises：It won’t promise outcomes; banned phrases and sensitive-topic limits sit in the configuration, checked before go-live.
A limit isn’t a gap in capability. It’s the judgement you owe customers — written down, not left to whoever is on shift.
Three things shop owners always ask
Are DMs, WeCom and website chat three systems?
Three doors for the customer, one team for you. Your site, WeCom and social contacts all run on the same configuration, so a conversation picks up where it started and nothing is copied across by hand. It can also connect with a CRM. During the demo, we map what can be connected against your current channels.
We’re a chain and our locations don’t sell the same way.
They shouldn’t. In TOPPP, configuration is scoped to a project — AI workers, knowledge bases and review criteria all sit inside one, so locations or regions can each run their own: share what is shared, split what isn’t. You can also start with the location that closes best, prove it there, then copy it out.
What if the AI says the wrong thing and they walk in tomorrow?
Three lines of defence. Before launch, the review workbench runs quality checks — no pass, no deployment. After launch, problem conversations feed an improvement loop, so the same mistake isn’t repeated. A human can take over at any moment, invisible to the customer. On a shop floor a wrong answer is paid for in person, so red lines are hard rules, not reminders.
Bring your shop’s playbook — watch it become a team
A demo takes about 30 minutes and covers three things:
- 01Map the playbook your shop runs on today and find what AI should take
- 02Watch it live: you play a customer while agent routing runs on the right
- 03Run both ledgers on your numbers: reply speed and booking rate against headcount
Or email us directly business@toppp.ai