A guest message is often the start of a job for somebody else. The cot needs carrying up to the second floor. The blocked drain needs a plumber. The 1pm arrival depends on whether the cleaner can finish by half twelve. Answering the guest is the easy part.
That second half, the coordination, is where much of a property manager's day goes. It's also the part that most guest messaging tools leave alone.
The work hiding in the inbox
Look back at a week of guest conversations and mark every one that needed someone to do something besides reply. In most portfolios it's a sizeable share. Extra towels, a missing remote, a request for a high chair, a smoke alarm chirping at 3am, a parcel arriving before the guest does.
Each of those used to run the same loop. Someone on your team reads the message, works out who should deal with it, sends that person the details, waits, checks it's done, and tells the guest. Five small steps, several apps, and a fair chance one of them slips on a busy Saturday.
What an autopilot does with it
An AI autopilot runs that loop for the routine cases. It recognises that the message needs action, works out what kind (housekeeping, maintenance, an owner decision), and creates the task with everything the person needs. The property, the reservation, access details, the guest's own words and any photo they sent.
Then it closes the loop. When the cleaner confirms the cot's in place, the guest hears about it. When the handyman's running late, the guest gets a revised time before they have to ask. All of it is written back to Guesty or Hostaway, so the reservation tells the whole story.
The same goes for questions that depend on operations. An early check-in can only be offered if the turnover allows it, so the autopilot checks the cleaning schedule first, and where it isn't sure, asks. The best systems message the housekeeper on WhatsApp, wait for the reply, confirm to the guest and update the arrival time in the PMS, with no person in the middle. A request to drop bags early gets the same treatment.
The next step is the schedule itself. When a cleaner calls in sick, someone has to rework the day, ask the others, and tell any guests affected. Scheduling tools show the gap. An AI agent can fill it. We cover that in AI staff scheduling for short-term rentals.
What stays with people
Some tasks need a decision before they need doing. Whether to send a plumber out at midnight or first thing. Whether a damaged sofa is wear and tear or a deposit claim. Whether an owner's happy to allow a dog this once. Those go to your team, with the context assembled.
Urgent problems should always reach a person immediately. A leak, a lockout, a guest who feels unsafe. The autopilot's job is to recognise them, alert the right people and reassure the guest that someone's on it.
Making it work with your current set-up
You don't need to change how your cleaners or maintenance team work. A good autopilot fits around them, whether they get jobs through your PMS, a separate operations app, email or a messaging app on their phone.
What helps most is clear property information. Where the spare linen lives, which handyman covers which area, what the cut-off is for same-day cleans. The more of that is written down, the more the autopilot can handle without asking. Many teams find that writing it down is useful in its own right.
Read more on what an AI autopilot is, how agentic AI handles a request from start to finish, or how the same approach applies to upselling.