Get started

After-hours bookings: how an AI assistant fills a restaurant’s tables

Most restaurant questions arrive when nobody can pick up the phone. Here is how an AI assistant answers them from your own menu — allergens included — and turns late-night browsing into booked tables.

A cosy restaurant terrace at dusk with a phone showing a chat conversation on the table

It is Friday, 22:47. Someone is planning tomorrow’s dinner with friends. They find your restaurant, like the photos, and have two questions: is there a table for four around seven, and can one of them eat gluten-free? The kitchen has closed, the phone goes to voicemail, and the menu is a PDF that says nothing about gluten. So they open the next restaurant in the search results.

You never find out that this guest existed. That is the real cost of unanswered questions in hospitality: not the complaints you get, but the bookings that silently go elsewhere.

The same twenty questions, every day

Ask anyone who has worked a restaurant phone. The questions barely change:

  • Are you open on Sunday? On the public holiday? In August?

  • Is there a table tonight / on Saturday for six?

  • Do you have vegan, gluten-free or lactose-free options?

  • How much is the lunch menu?

  • Where can I park? Do you deliver?

  • Can we bring the dog? Is there a high chair?

Each one takes thirty seconds to answer. The problem is when they arrive: in the middle of service, or after closing — exactly when nobody has thirty seconds.

An Enéh chat window for Trattoria Bella Vita: a guest asks for a table for four with a gluten-free option late on Friday, and the assistant records the booking request
A late-night question, answered from the restaurant’s own menu — and turned into a booking request.

What changes when the answers are instant

Bookings that don’t wait for the morning

The assistant collects everything a booking needs — date, time, number of guests, name, phone, any special request — and passes it to you by email and on your dashboard. On every plan it works as a booking request you confirm. On the Pro plan it can go further and book straight into your free slots, with Google Calendar sync, email confirmation and a reminder.

Either way, the guest leaves the conversation with a clear next step instead of a voicemail.

Allergens without guesswork

This is where a restaurant assistant has to be strict. It answers allergen and diet questions only from the menu you approved: if your data says any pasta can be made gluten-free for +490 Ft, it says exactly that. If your data says nothing about sesame in the dessert, it says it doesn’t know and offers to put the guest in touch with the team. It never “assumes”.

Tourists in their own language

If your restaurant is anywhere near a tourist route, a good share of your website visitors don’t read Hungarian. The assistant replies in the visitor’s browser language automatically — English, German, Italian and most European languages — from the same approved menu. Prices, opening hours and the booking flow stay identical.

Groups and private events still come to you

A birthday for twenty, a company dinner, a custom wedding menu — these are conversations you want to have yourself. The assistant recognises them, collects the basics (date, headcount, budget, contact) and hands them to your team, so you call back with context instead of starting from zero.

Diagram: the assistant answers opening hours, menu, allergens, parking, booking requests and foreign-language questions; complaints, private events and anything outside approved knowledge go to the team
A clean split: routine questions to the assistant, judgement calls to your team.

Keep the menu honest

An assistant is only as current as its data. Seasonal menus, a new lunch price, the two weeks you close in August — these are exactly the things guests ask about, and exactly the things that go stale.

With Enéh you don’t maintain a knowledge base. We build it from your website and menu, a person reviews it with you before it goes live, and when something changes you tell us (or edit it yourself in a minute). The holiday closure is in the data, so on 15 August the assistant tells the guest you are closed — instead of happily taking a booking.

What to measure in the first month

  1. How many booking requests came through the assistant.

  2. How many of them arrived outside opening hours. These are bookings you were simply losing before.

  3. The ten most frequent questions. If “do you have parking?” is number one, put the answer on your homepage too.

  4. How often it handed a conversation to your team, and why. Every handover points to a gap in your data — or a guest worth calling.

Try it on a real restaurant site

Our demo restaurant, Trattoria Bella Vita, runs the same assistant our customers use. Ask it for a gluten-free option, a table for six on Saturday, or whether it is open on 15 August — then try to catch it out. Open the restaurant demo, or request a demo built on your own website.

Magyarul olvasnád? Foglalás zárás után: így tölti meg az AI asszisztens az étterem asztalait.

See it working

Seven industry demo sites, each with a trained Enéh assistant on it — ask them anything.

Browse the demos