Call · 15 min
Sector · Restaurants & Bars

The phone rings always during service.

Restaurants, trattorias, bars. Whoever calls to book does so exactly when nobody on the floor can answer; booked tables fall through without notice; reviews go unanswered because at midnight there is nobody left with the clarity to write a reply.

·A typical day

A restaurant's margin is decided outside the kitchen.

The quality of the dish is already solved: it is the reason the place still exists. What remains unsolved lies all around the service.

12:30

The phone, at the peak

It rings three times during service. Once someone answers in a rush, twice nobody does. Whoever got no answer has already called the next place.

Unhandled calls and messages are collected and receive a first reply with availability and confirmation. Inbox AI
15:00

Tonight's tables

Nobody has confirmed the evening's bookings. Confirming them by hand means twelve phone calls in an hour that was meant for preparing service.

The reminder goes out on its own; whoever doesn't confirm is contacted again before the table stays empty. Booking recovery
19:45

The table for four that doesn't arrive

Booked, never cancelled, never showed up. There are people waiting in the dining room, but the table stayed blocked for forty minutes on a booking that no longer existed.

The freed slot is detected in time and can be reassigned instead of staying empty. Booking recovery
23:40

The review that just came in

Three stars, a remark about the waiting time. It would deserve a calm reply, and at this hour nobody has a calm reply to give.

The draft reply arrives ready the next morning, to approve or correct before publication. Review follow-up
Monday

The supplier's invoice, and ours

The supplier has already sent two reminders. Nobody has chased the invoice the restaurant has to collect from a catering job three weeks ago.

A sequence of increasing courtesy follows up overdue invoices, without depending on the morning's mood. Reminders

A lost cover doesn't come back the next day: it is margin that simply disappears. The difference between a full evening and a half-empty one rarely lies in the menu.

·The sector's numbers

Not our opinion. Public research, with the source alongside.

12,8%

is the average no-show rate measured across more than 212,000 bookings in 300 Italian restaurants. With a deposit or pre-authorisation it drops to 5.4%.

Osservatorio Incrementoo · Ristorazione 2025
−32%

fewer no-shows with automatic reminders by SMS, WhatsApp or email, measured on the same sample. It is a reduction that adds to the one obtained with a deposit.

Osservatorio Incrementoo · Ristorazione 2025
97%

of consumers read reviews of local businesses before choosing, and 80% prefer those that reply to all of them. BrightLocal 2026 survey of 1,002 US consumers.

BrightLocal
100×

more likely to make contact by replying within five minutes instead of thirty: the window in which a booking is still yours is very short.

Harvard Business Review · InsideSales
·What we build here

The pieces that in this sector pay for themselves first.

They aren't all built at once. You start from one, the one that costs most today, and add the next only if the first has paid off.

Booking Recovery

Reminders and reassigned tables

A reminder goes out before the booked time; whoever doesn't confirm is contacted again. The table that would otherwise go empty can be reassigned in time.

We measure: no-shows avoided.
Why hereIt is the piece with the most direct return and the easiest to verify: covers recovered against covers lost, month on month. Full page for the system
Inbox AI

The calls and messages lost during service

Requests from phone, WhatsApp, email and website collected in a single point and classified straight away. Recurring questions, such as opening hours, availability, allergens and parking, have an answer approved by the restaurant.

We measure: average time to first response.
Why hereThe two hours when you can't answer are exactly the two hours when people book for the evening. Full page for the system
Reviews

Replies ready the next morning

The system collects the reviews and prepares a draft reply, ready for the owner's approval before publication.

We measure: average time to reply to a review.
Why hereIn local restaurants the review profile is the shop window: it is what a new customer looks at instead of the sign. Full page for the system
Reminders

The receivables nobody chases

A three-level sequence of increasing courtesy follows up overdue invoices for catering, events and supplies, instead of depending on whoever remembers to do it.

We measure: average days to collect.
Why hereIn places that also do events and catering it is often the most neglected cash item of the year. Full page for the system
·Real applications

Not a concept: a prototype already built and put under stress.

Built with a cash problem in mind, not a courtesy one: the table that frees up too late to be resold. Anonymised, as it is, with synthetic data and put under stress before the first conversation.

Prototype · Restaurant or bar with booked service

The booked table that falls through isn't a question of manners: it is takings that vanish in silence.

It doesn't matter whether the customer apologises or not. What matters is the seat left empty for an hour while people were waiting in the dining room. The system doesn't chase courtesy: it detects who is at risk of not showing up, sends a warning in time for a reply, and if the table frees up anyway it puts it back on sale before it is too late.

On one evening's flow (real test)Today, by handWith the system
Bookings assessed for riskby instinct, if someone notices24 out of 24, automatic
Reminders to at-risk bookingsrarely, depends on who is on the floor16 out of 24 sent
Table freed in time and resoldalmost never: it stays blocked until the booked time8 out of 12 no-shows reassigned
Euros tracked per eveningnobody adds them uprecovered vs lost, per booking
Detects

The Risk

Every booking, as soon as it comes in, receives a score: first visit, no confirmation, large group, peak time.

Warns

The Reminder

It goes out only to those who need it, in time for a reply: not to everyone, not at random.

Recovers

The Waiting List

The table freed in time is offered again straight away, instead of staying empty until the end of the evening.

24
bookings in a test
0
errors
0,3s
per reply
€1,056recovered in this test, out of €1,568 at risk in a single evening
12 no-shows detected out of 24 67% of freed tables recovered from the waiting list €512 still lost, in the same test
What the test says, and what it doesn't
Bookings assessed in the test24
At risk, because not confirmed in time12
Offered again to the waiting list and recovered8
Still lost, same test4

The bookings were generated for the test, they weren't your customers: the test says the mechanism works, not how much it is worth at your place. You know what a table is worth, and we do the sums together.

The first step, here

The covers lost to a missed confirmation. Thirty days to really count them.

In restaurants the piece that pays for itself first is almost always booking recovery: it is the quickest to get running and the only one whose effect shows in the till within the month, not after a quarter.

1Process tackled
30Days to a running system
1Indicator agreed beforehand
0Commitments beyond 30 days

We choose the indicator for this sector together before starting: no-shows avoided in the month, or average time to first reply to a booking request. It is the yardstick you will judge us by.

If the piece has returned more than it cost, the next one is assessed. If it hasn't, we stop there: no automatic renewal, no penalty, and the system stays yours.

Work out which piece to start from
·This sector's questions

The objections we hear every time.

The automatic message to customers feels cold. It isn't like us.

A badly written reminder is worse than no reminder, and that is why you write the texts, or we write them together in your words, and you approve them before they go out. There are no anonymous system messages.

The sector's experience is that a well-made reminder is read as attentiveness, not automation: it is the same reason your dentist sends you one.

We are a small place. This is chain stuff.

It's the opposite: a chain has an office that handles it, a family-run place doesn't. The system's value is proportional to how much a dedicated person is missing, not to the size of the turnover.

That said, the signal for which we would tell you to forget it does exist: if you take ten bookings a week and handle them all without strain, it isn't worth the expense, and we tell you so before sending a quote.

We don't have time to learn another program.

There isn't one to learn. The design criterion is that things arrive where the place already looks: a message, the email inbox, the spreadsheet you use. If a piece requires someone to learn a new tool to keep service running, that piece is badly designed.

The real effort is in the analysis phase, at the start, and it is a few hours in total: we need to see how you really work, not a questionnaire filled in in a hurry.

Other questions, such as price, data and time required, are answered on the Method page.

How many covers fell through last month without anyone writing it down anywhere?

With Itria we start from the outside to build custom digital systems for those who work with service open. For you that translates into more bookings collected, fewer tables lost and less work after closing. Write us a line about what weighs on you. We take the first step: what a customer sees when they look you up, and what we found there. Even if we don't end up working together.

Write us a line

Or measure in five minutes (in Italian)