Skip to content
AI setup solutions · Hotel & wellness spa

Don’t lose the booking messages that arrive after hours

Guests message on WhatsApp and Instagram at night and in different languages. The assistant replies with approved answers, pre-qualifies booking requests and hands them to your team — and never answers health questions itself.

Sound familiar?

This setup is for these problems

After-hours messages wait until morning
Replying to international guests in their language is hard
Price, package and opening-hour questions repeat all day
Google and TripAdvisor reviews go unanswered

What we set up

  • WhatsApp or Instagram message assistant (multilingual, with an approved-answer library)
  • Booking-request pre-qualification and hand-over to your team
  • Review reply drafts (published once you approve)
  • Appointment reminders and post-visit review requests (opted-in guests only)

What the assistant won’t do

  • Health or treatment claims — questions like “will it help my back pain?” go to your team
  • Confirming bookings on its own
  • Sharing room, guest or anyone else’s details

Rules we follow in this sector

  • Promotional messages only to opted-in contacts (İYS)
  • Medical aesthetic centres are out of scope for now: health-promotion rules need a separate build
  • Labels on AI-generated promotional visuals

Regulatory notes are research-based, not legal advice. Last checked: 30 September 2026.

Measure first: AI X-ray (Turkish)About 20 questions on which tasks AI could take over first. Your answers are not sent anywhere.
Security & compliance

What every setup includes

Human approval first

The assistant never confirms bookings, quotes or refunds on its own; it takes the request and hands it to your team.

Approved answers only

Prices, offers, hours and policies come only from a library you approve; the assistant does not make them up.

Personal data is masked

Names, phone numbers and IDs are stripped before a message reaches the AI service (Turkish data-transfer rules).

An off switch

Switch the assistant off in one step and route every message straight to your team.

Abuse tests

Before going live we run 12 malicious-message scenarios; if a single one gets through, we do not launch.

Monthly value report

Incoming messages, first-response time, handovers and captured requests — reported together with how each was measured.

How we work

  1. 1

    Conversation

    Where do your messages come from, which ones get missed, what eats the most time — we listen first.

  2. 2

    Inventory

    We write down your services, price source and what you do not offer; that is where the assistant’s limits are drawn.

  3. 3

    Pilot

    A small setup on one channel with written limits. Tests, team training and a hand-over file included.

  4. 4

    Maintenance

    We fix broken flows, update the library and review what works with a monthly report.

What we do not promise

  • We do not promise a specific sales or booking uplift; we measure and report.
  • We do not give legal advice; we build to current rules and leave legal decisions to you and your lawyer.
  • We never present the assistant as a human; it talks to your customers as an AI assistant.

Let’s talk about your business

Pilot scope and price are set after a conversation, based on your needs. It starts with a free intro call.

Request a call

Prefer to try it yourself first?

Our AI guide is free (in Turkish): what each tool does, its free alternatives, and which jobs you can do yourself.