Booking recovery. The place nobody cancelled stays blocked until the last moment.
A booking that doesn't show up doesn't leave a hole in the register. It leaves a place assigned to someone who won't arrive, and keeps it occupied until it is too late to give it to someone else.
The damage is how late the absence is discovered. A table cancelled at 18:00 can be resold. The same table, found empty at 20:30, is lost twice: the cover that doesn't arrive and the waiting couple who leave.
The reminder goes out on its own; the cancellation doesn't. The system asks for confirmation, reads the reply on whatever channel it arrives and records the cancellation. Cancelling a booking the customer hasn't cancelled, and offering the place to someone else, remain a person's decisions.
There is one indicator and it is agreed beforehand: the percentage of no-shows out of total bookings. You measure how things stand today over the previous thirty days, set the threshold the work must pass, and compare afterwards.
This is the page for a single piece of the system. The other pieces, and the criterion for choosing which to start from, are on the services page.
The place becomes sellable again while there is still time to sell it
The gain from this piece is notice: knowing two hours in advance that a place has come free, instead of discovering it once service has started. A booking that falls through without warning takes up two things. It takes up the place, which stays unavailable to whoever would have taken it, and it takes up the attention of the staff, who keep holding it for someone who isn't coming.
Asking for confirmation by hand works, and indeed almost nobody does it every day. Twelve phone calls in the hour that was meant for preparing the service are real work, and the first busy day is the day it gets skipped.
The empty place costs as much as the lost cover. The empty place discovered late also costs the customer who could have taken it.
How much is recovered, and on what sample it is measured
The most solid measure available on this mechanism comes from healthcare, where absences have been counted for decades. A systematic review published in BMJ Open in 2016, gathering 21 controlled studies and 16,076 patients, finds 15% absences among those who receive a notification against 21% among those who don't: a quarter fewer.
The same authors add a practical detail: several notifications work better than one. In their words, «sending multiple notifications could improve attendance further». That sample, though, is clinical appointments, not tables and not rooms, and the patient who skips a visit has no alternative venue to go to.
On the restaurant side, the number going around in Italy comes from an analysis of 212,000 bookings in 300 restaurants between January and August 2025: no-shows at 12.8% on average, down to 5.4% where there is a deposit, and a further reduction of 32% with automatic reminders.
That analysis is signed by a company that sells booking technology, and it reached the trade press as a press release, with no study available alongside it. The number remains the broadest there is on the Italian market, and these two lines are why we don't treat it as a promise.
What the system does, step by step
Every booking enters the same route, whatever channel it arrived through, and every step leaves a log entry saying what happened and when.
| Step | What happens | What you get |
|---|---|---|
| Reminderbefore the appointment | A message goes out on the channel the customer booked through, with date, time, number of people and a single question: do you confirm. |
A reply instead of silence, because the question is closed and can be answered in one word. |
| Reading the replyon any channel | Confirmations, cancellations and changes of time are recognised even when written in the customer's own way, and they update the booking. |
The diary updates anyway, even when the customer replies in their own way instead of pressing the expected button. |
| List of unconfirmed bookingswho hasn't replied | Whoever doesn't reply stays booked and goes onto a separate list, ordered by time, which a person looks at before service. |
Before service you know exactly who hasn't confirmed, without depending on anyone's memory. |
| Place back on offeras soon as the cancellation arrives | The cancellation frees the slot in the diary and flags it immediately, with the time by which to reassign it for it still to make sense. |
The place comes back to you while it is still worth something, with a note of the time by which it is worth reassigning. |
| Reassignmentdecided by a person | The system suggests whom to contact among the requests turned down and the waiting list for that day; a person sends the offer. |
A ready list of whom to call first, and the last word on the offer stays yours. |
No customer cancelled by mistake, and the rule that guarantees it
A customer cancelled by mistake doesn't come back, and this system is built so that it can't happen. A system that reads human replies makes mistakes, and its two mistakes don't cost the same. Reading a cancellation as a confirmation leaves a place blocked, which is the original problem. Reading a silence or an ambiguous sentence as a cancellation deletes a customer who would have come.
That is why the system is deliberately lopsided: only an explicit cancellation cancels anything. Silence leaves the booking valid. A reply the model can't classify is shown to a person instead of being interpreted, and stays marked as unread until someone looks at it.
Each row records whether it was handled by the model or passed to a person, so the share of ambiguous cases gets counted instead of disappearing. If it grows, the problem shows before it becomes an angry customer on the phone.
The engine is built and tested on its test branches, with no real customers involved. It is a functional check, and we state it for what it is.
What goes out on its own, and what waits for a person
The reminder, reading the reply, recording the cancellation and flagging the place that has come free all go out on their own. They are all operations on an appointment the customer has already made: they repeat an agreed fact, they don't add a new one.
Everything that commits the business waits for a person: cancelling a booking nobody has cancelled, offering the freed-up place to another customer, applying a financial condition, asking for or keeping a deposit. The reasoning behind this line is on the page about the principles we build with.
The owner switches the automatic part on and off channel by channel, with a switch, without touching the system again. Some businesses want the reminder always on, others keep it on only on busy days.
Reminding someone of an appointment is repeating something already decided. Cancelling it is deciding something new.
The message states that it is a system
From 2 August 2026 Article 50 of the European regulation on artificial intelligence applies, and for whoever uses a system like this it translates into a duty to inform: the person must know they are talking to a system, and must know it before talking to it, not after.
A reminder invites a reply, so it opens a conversation. It says so in one line, within the text itself, and the statement holds for the whole exchange. What really changes for a small business we have written in the AI Act and Italian SMEs, and the systems we use are listed on the AI transparency page.
The statement weighs little on the result, because whoever booked expects to be contacted again. It removes the only serious risk of the function, which is a customer convinced they wrote to the owner.
The same thing changes name with the trade
The mechanism is identical; the place that gets lost isn't. It is worth looking at your own case, because that is where you see how much it costs today.
| Sector | The place that gets lost | Where you see it |
|---|---|---|
| Hospitalityaccommodation and events | Two bookings for the day after tomorrow that have never confirmed: if they don't arrive, the rooms stay empty without anyone having been able to reassign them. |
|
| Restaurantsand bars | The table for four, booked, never cancelled and never shown up, held free for forty minutes while people wait in the dining room. |
The service of a dining room where the phone rings unanswered |
| Appointment diariespractices, workshops, personal services | The hour of work reserved for someone who doesn't show up, which could have gone to one of the customers pushed back to the following week. |
What this piece doesn't do
It doesn't replace the booking software, the channel manager or the diary. It works on what happens outside those systems today, where confirmation is asked for by voice and the cancellation arrives on a channel the diary doesn't read.
It doesn't bring in new bookings. It works on the existing ones, and that is why the return is calculated on the current volume, without depending on a campaign that still has to work. The requests that today don't even become bookings are a different problem, and that is the work of Inbox AI, the first reply to every request.
It doesn't introduce the deposit on your behalf. It is the lever with the biggest effect among those measured, and it remains a commercial decision of the owner, with consequences for the relationship with the customer that no tool can assess.
Questions and answers
Is a reminder enough to bring no-shows down?
It moves them, and by how much depends on the venue. The most solid measure available comes from healthcare: a systematic review published in BMJ Open in 2016, 21 controlled studies and 16,076 patients, finds 15% absences among those who receive a notification against 21% among those who don't, that is a quarter fewer.
In Italian restaurants an analysis of 212,000 bookings attributes a 32% reduction to automatic reminders. These are samples that differ from a restaurant or an accommodation business, so the direction holds and the size is measured on site, over the previous thirty days.
What happens if the customer doesn't reply to the reminder?
Nothing automatic, and that is a design choice. Silence is never read as a cancellation: the booking stays valid and goes onto a list of unconfirmed bookings that a person looks at before service.
The same goes for a reply the model can't classify, which is shown instead of being interpreted. Cancelling a booked table because of a reading error costs more than the empty place you wanted to avoid.
Is a deposit needed? Is it something the software does?
No, the deposit is a commercial decision, and it is worth saying so because in the same analysis of 212,000 bookings it is the lever with the biggest effect: no-shows go from 12.8% to 5.4% where there is a deposit or pre-authorisation, with an average amount of 20 euros per person.
The system can ask for it, record it and remind people of it. The decision to introduce it stays with the owner, and it has consequences for the relationship with the customer that no tool can assess on their behalf.
Do we have to change our booking software?
No. The booking software, the channel manager and the diary stay where they are. The work lies in what happens outside those systems today: asking for confirmation, reading the reply on whatever channel it arrives, recording the cancellation and flagging the place that has come free while there is still time to reassign it.
We propose replacing a tool only if it is the bottleneck measured during analysis, never out of preference.
How do you measure whether it's working?
With the percentage of no-shows out of total bookings, agreed as the indicator before starting. The current value over the previous thirty days is measured, the threshold the work must pass not to be considered a failure is set, and the two are compared afterwards.
The second number is how many freed-up places were actually reassigned, which is the most honest count because it counts heads present, not messages sent.
Notes on sources
- The figure of a quarter fewer absences comes from Robotham et al., Using digital notifications to improve attendance in clinic, BMJ Open, 2016: a systematic review and meta-analysis of 21 controlled studies, 8,345 patients with notification and 7,731 without. It measures healthcare appointments, not tables and not rooms. We state this because the sample isn't the reader's.
- The figures on no-shows at 12.8%, deposits at 5.4% and reminders at 32% come from an analysis of 212,000 bookings in 300 Italian restaurants, January to August 2025, reported by the trade press. It is signed by a company that sells booking technology and there is no study available alongside the press release: we report it because it is the broadest Italian sample available, and we flag the limitation.
- This page doesn't report results obtained for a client, because this piece hasn't yet been delivered to a client. The tests cited are functional checks carried out in testing, and that is why you won't find a recovered figure here.
- No share of freed-up places that can be reassigned is published: it depends on how early the cancellation arrives, on the day of the week and on the venue's demand. The number is calculated on your previous months, before any quote.
The other pieces in this group
Missed bookings and customers who don't come backFifteen minutes, with your case in front of us.
How many bookings, last month, didn't show up? And how many of the cancellations were reassigned? If the answer is that nobody knows, that is already the most useful information, and it can be counted in thirty days. In fifteen minutes on the phone we look at it together and tell you where it makes sense to start, even if we don't end up working together.
You get Mattia Esposito, who then builds the system: there's no salesperson in between. If you'd rather measure on your own before talking, the Diagnostico (in Italian) is twenty questions and five minutes.