A restaurant AI chatbot that takes every reservation, even at the dinner rush

Service is exactly when the phone rings

A restaurant's busiest hours for enquiries are its busiest hours for everything else. The booking line rings mid-service, a stack of reservation DMs builds up on Instagram while the floor is full, and the same allergen question arrives for the tenth time that day. Every one of them is a table you could have filled, and a missed booking does not come back tomorrow — it goes to whoever answered first.

None of it needs a person. "Do you have a table for four on Friday", "is the pasta gluten-free", "what time do you close on Sunday", "can I move my booking to 8pm" — all answerable from your menu and your booking system, all repetitive, and all arriving when no one has a hand free. An always-on agent takes the booking and answers the menu across every channel a guest actually uses.

What it handles, and what reaches your team

The restaurant pages start from the built-in restaurant template, so menu, allergen, hours and booking questions are covered from the first day, and you tune the wording against real guest questions instead of writing from scratch.

Taking the booking on every channel a guest uses

Guests do not all arrive through your website. Some message on WhatsApp, many slide into Instagram DMs after seeing a photo, and the rest phone. The agent handles the reservation in all of them, against real availability from your booking system, so the DM that used to sit unread until the next morning becomes a confirmed table in the moment. One agent, one set of answers, every inbox.

A restaurant answering service for the dinner rush

Plenty of bookings still come by voice, which is why the widget alone does not cover it, and the phone rings loudest during the dinner rush itself — exactly when nobody on the floor has a hand free to answer it. The AI voice agent answers inbound calls with the same knowledge as your chat agent, in natural speech: it can take the booking, answer the menu and hours questions, and take a message when a person is needed.

It books the table into your live availability while the caller is still on the line, so a party of four asking for eight o'clock gets eight o'clock confirmed on the spot rather than a note for someone to call back once service calms down. Forward the line after hours, or run the agent on it through the whole shift, and the calls that used to go straight to voicemail during the rush get answered the same way a quiet Tuesday afternoon call would be.

Every call lands in the unified inbox as a transcript, so a missed call during service is no longer a lost table — a manager can see exactly what the caller wanted before calling back, rather than starting from a blank voicemail. Phone runs on the same AI answering service as the rest of the platform, so a booking made by voice, on WhatsApp or through the website all lands in the same place. See how the voice channel works.

Where the growth actually comes from

Bookings you were dropping. A table request answered at 22:00 with a confirmed slot is a different outcome from one answered at 10:00 the next day with an apology. The demand was already there; the agent simply stops it walking.

Front-of-house time back on the floor. The same team, freed from the booking line and the DM backlog, spends the shift with the guests in the room. The analytics dashboard reports covers booked, enquiries captured and hours saved, so you can see what the agent absorbed.

Bigger-ticket enquiries caught early. Private dining, set menus and large parties are the enquiries worth a person — the agent captures them with the detail your events lead needs and routes them on, instead of letting them expire in an unread inbox.

The line it does not cross

It does not overpromise the kitchen. It answers from your published menu and allergen information and stops there. Bespoke dietary requests and "can the chef do X" are captured and passed to your team, not invented.

It does not confirm what it cannot see. If the booking system shows no availability, it says so and offers the nearest slots, rather than double-booking a full service.

It does not handle complaints. Anything about a meal already served goes straight to a person with the full transcript, because a bad night needs a human reply, fast.

What to watch in the first month

Watch bookings taken outside service hours, the share of Instagram and WhatsApp enquiries answered without staff, and the questions the agent missed — knowledge gaps point straight at the menu, allergen and opening-hours details your site never made clear.

Taking bookings and orders on WhatsApp

Diners already message you on WhatsApp — for a table on Friday, to change a party size, to ask if the kitchen is still open. The agent answers from your own menu and booking rules, holds the reservation, and passes anything it cannot settle to a person with the whole thread attached. Nothing sits unread until the next quiet moment between services. See how the WhatsApp channel works.

Reservation DMs on Instagram

Your Instagram is where people find the room and then ask "can I book for six?". The same agent replies in DMs with your hours, your set-menu details and a booking link, so a discovery scroll turns into a table instead of an unanswered message. See how the Instagram channel works.