On the first evening of each stay, guests get a WhatsApp message asking how the room is. An AI agent sorts every reply into a maintenance task, a housekeeping request or a duty manager alert, and each case is closed and logged before checkout.
Typical results
≈45%
Guests who reply to the first-evening message
≈2 hrs
Typical time from guest report to fixed
100%
Cases logged against the booking before checkout
Figures reflect typical results for this kind of automation.
The challenge
Most guests at a thirty-room hotel don't mention a noisy air conditioner or a missing pillow during their stay. They mention it in the review after checkout, when nothing can be done. The front desk is too busy to call every room, and requests that do come in by phone are passed on verbally and sometimes forgotten.
Before and after
Reviews mentioning a room problem per month
−25%Hours from guest report to fixed
−89%The automation
The property management system triggers a WhatsApp message on each guest's first evening, asking how the room is and whether anything is needed. An AI agent reads each reply and sorts it. Faults such as a noisy air conditioner become maintenance tasks, housekeeping requests go to the floor supervisor, and complaints alert the duty manager with the room number. Closed tasks trigger a follow-up to the guest, and every case is logged against the booking before checkout.
How it works
- 01
First-evening message
The PMS triggers a WhatsApp message on the first evening asking how the room is and whether anything is needed.
- 02
AI reads the reply
Each reply is classified as a fault, a housekeeping request, a complaint, a question or a positive comment.
- 03
Routed to the right team
Faults become maintenance tasks, requests go to the floor supervisor and complaints alert the duty manager.
- 04
Guest follow-up
When the task is closed, the guest gets a short WhatsApp message confirming it has been handled.
- 05
Logged on the booking
Every case is recorded against the booking, so the front desk sees it before checkout.
Starts from
- PMS first-night trigger
- Guest WhatsApp replies
Delivers
- Maintenance task
- Housekeeping request
- Duty manager alert
The impact
Problems are found while the guest is still in the room and are typically fixed within about 2 hours, instead of surfacing at checkout. Complaints mentioned in online reviews fall by about 25%, and the duty manager sees every open case in one place before guests leave.
Questions about this automation
What does the guest receive and when?
Around 7 PM on the first evening, the guest gets a short WhatsApp message in their booking language asking how the room is and whether anything is needed. Guests who arrive late get it the next morning instead. Happy replies get a thank-you, and no further messages are sent unless the guest writes again.
How does the AI decide where a reply goes?
It classifies each reply as a fault, a housekeeping request, a complaint, a question or a positive comment, and takes the room number from the booking. Faults go to maintenance, requests to the floor supervisor and complaints to the duty manager. Anything unclear is passed to the front desk to read and answer.
Which hotel systems does it connect to?
It reads arrivals from the property management system, sends and receives messages through the WhatsApp Business API, and creates tasks in the hotel's maintenance or housekeeping app, or on a shared Google Sheets board. Make runs the steps with OpenAI for the triage. This is a multi-step build, usually rolled out in stages.