The Same Arrival Question, Forty Times a Week
Most after-hours guest messages are the same handful of questions, and the answers already exist in your booking record. Here is why they still reach you at eleven at night.
Count the guest messages you answered last week and sort them by content rather than by guest. The list gets short very quickly.
Where do I park. What time can we get in. Is the code the same one from the email. Has the house been cleaned yet. Can we drop bags early. Which door does the key box open.
That is most of it. And every one of those answers already exists somewhere in your systems.
Why an answer that exists still costs you an evening
The information is not missing. It is scattered, and it is scattered in a way that makes it useless at the moment it is needed.
The arrival time is in the booking. The code is in the lock app. The cleaning status is in the turnover tool. Whether bags can be dropped early depends on all three at once, plus whether the previous guest has actually left.
No single system can answer "can we drop bags at one" because no single system knows. So the question routes to the only component that can see all of them, which is you, and it does that at whatever hour the guest happens to ask.
The volume is not the problem
Forty messages a week is not much typing. If it were only typing, nobody would mind.
What it costs is different:
Interruption. A message at ten at night is not fifteen seconds of work. It is the end of being off duty. The cost is the evening, not the reply.
Latency. Messages that arrive after the desk closes wait until morning. For a guest comparing three properties, that gap is where the booking goes elsewhere. The enquiries you never answered are invisible on every report you have.
Inconsistency. Answered by four different people across a season, the same question gets four slightly different answers. One of them will be wrong about the cleaning window, and that becomes a review.
The test that tells you whether automation will help
Take your last fifty guest messages and mark each one:
- Answerable from the listing. Wifi password, house rules, nearest supermarket.
- Answerable from the booking record. Arrival time, code, what was paid for.
- Answerable only from live operational state. Has the cleaner finished, is the previous guest out, can we move the arrival.
- Needs a person. Complaints, damage, anything with an upset tone.
Category one is what the AI messaging bundled into your property management system already handles. If most of your fifty land there, switch it on and stop reading.
The interesting number is categories two and three. On most portfolios they are the majority, and they are precisely the ones a standalone messaging tool escalates to you, because it cannot see the booking record or the cleaning app while it is composing a reply.
What connected actually means
A reply layer that can read the operation answers the same question differently:
"You are in Casa Verdera from four. The cleaner finishes at two, so bags from two thirty is fine. Your code is 4417 and it activates at three."
That is not better language. It is the same sentence with four facts in it that the messaging tool did not have. Getting those four facts into the reply is an integration problem, which is why buying a more articulate bot does not fix it.
Where to start
Pick the single question you answer most often after eight in the evening. Just one.
Work out which systems hold the pieces of its answer. If the pieces all sit in your property management system, your existing tools can probably handle it and you should make them. If the answer needs the lock app and the cleaning app to agree, you have found a real integration, and it is worth building before next season rather than during it.
