AI receptionist · Restaurant
LiveAI receptionist for restaurants
Table, group, banquet and catering requests answered on your web chat and email during service and after close. Menu and prices from your card; the team confirms every booking.
Runs on the Restaurants and catering job description of the AI inbound sales manager. Web chat and email today; phone is planned.
Unistaff · Restaurant
Host stand
In one paragraph
What does an AI receptionist do at a restaurant?
An AI receptionist for a restaurant answers table, group, banquet, catering and takeout questions on your web chat and email. It says it is the restaurant's AI assistant, sorts the request type in its first reply, collects date, time, guests and allergies as a note for the kitchen, quotes events only as “starting at” from your card and leaves every confirmation to the team. The pilot is free for 30 days.
How much does it cost?
The first 30 days are a free pilot: one AI inbound sales manager for one company on web chat and email, no payment and no card, stop at any time. After it, one flat monthly fee per AI employee is agreed on the pilot review. Messages and minutes are unmetered, and no list price is published.
The path of one enquiry
Five steps from a message on your site to a confirmed booking. The AI works the first four; a person owns the last one and anything handed over on the way.
- 01
Message arrives
Web chat on your site or the email address set up for the pilot
- 02
Request type
Table, group over eight, banquet or catering, takeout, a change, a complaint
- 03
The details
Date, time, guests, occasion, contact; allergies as a note for the kitchen
- 04
Request recorded
“Received, the team will confirm”; events get a starting price from your card
- 05
Team confirms
Availability, the table, the event terms and the final quote
Never by itself
Saying a table is booked before the team confirms, promising an allergen-free kitchen, taking card details in chat, signing or changing an event order, quoting a number outside your card.
What arrives at a restaurant desk
The mix below is an illustration of this kind of desk. The timing column shows why replies slip: a lot of it lands while the room is busy or after everyone has gone home.
A table for four tonight
During service
A group of twelve for a birthday
Evenings
A banquet or a rehearsal dinner
Weeks ahead
Office catering, off-site
Weekday mornings
Takeout and pre-order questions
Lunch rush
Gluten, nuts, vegan options
All day
Hours, parking, patio, capacity
Weekends
A complaint about last night
Next morning
What it takes on, what stays with a person
Read the person column as the job that remains, and staff for that job rather than for the whole desk. Every row follows the restaurants and catering job description as it runs today.
| Task | AI receptionist | Person |
|---|---|---|
| The first reply | AI receptionistSays it is the restaurant's AI assistant and sorts the request: a table, a group over eight, a banquet or catering, takeout, a change, a complaint or a question | PersonReads the request on the lead |
| A table request | AI receptionistCollects date and time, number of guests, name and phone, occasion and special requests; tells the guest the request is received and the team will confirm | PersonChecks availability and confirms |
| A group over eight | AI receptionistCollects date, time, number, occasion, room or set menu and a contact | PersonSets the terms |
| A banquet or catering | AI receptionistAsks about the event, then the seven fields: date and time, guests, place, service style, diet, contact and deadline; names only a “starting at” price from your card | PersonWrites the proposal and the final quote |
| Allergies | AI receptionistRecords them as a note for the kitchen, checks the priority allergens for events and promises no allergen-free kitchen; symptoms get 911 and a person | PersonThe kitchen decides what it can serve |
| Menu, specials, hours, parking, capacity | AI receptionistAnswers only from your knowledge folder; a missing fact gets “I will check with the team” | PersonKeeps the card current |
| Complaints, “the manager please”, delivery status | AI receptionistHands over at once with a summary; it lands in the cabinet with an email copy | PersonThe manager answers |
Where it hands over
Complaints, a request for the manager, delivery status, card details, signing or changing an event order, and every table until the team confirms it.
What the owner keeps
The book, the event terms and every complaint.
The seven fields before an event price
A banquet or catering enquiry gets a “starting at” figure only after these seven are on the lead. They are asked one per message, and nothing the guest already wrote is asked twice.
Date and time, and how flexible
Decides whether the room or the kitchen is free at all.
Number of guests
Sets the room, the staffing and the per-guest part of the quote.
At the restaurant or off-site, with the address
Off-site adds travel and equipment to the quote.
Service style
Plated, buffet, family style or drop-off catering price differently.
Dietary needs and allergies
Recorded for the kitchen, with no allergen-free promise.
A contact and who decides
The proposal goes to the person who can say yes.
The deadline for a decision
Tells the team how fast the proposal has to go out.
The kitchen reads the same lead, so the allergy note and the guest count arrive together.
Text-first: what about phone calls?
The AI receptionist plans compared on the pricing page (Dialzara, Rosie, Goodcall, Smith.ai) answer the phone first. This one is text-first: it works on the web chat on your site and on email to a dedicated address, both set up with you for the pilot. Phone and SMS are planned as a separate desk.
The practical read: if a good share of your enquiries already arrive in writing, a text-first desk covers them today. If the desk lives on the phone, keep calls with a person or an answering service, and point the repeat questions to the chat on your site.
Web chat on your site
LiveThe widget goes on your site with you or your webmaster during pilot setup. Replies come in the customer's language, and the chat keeps an AI badge.
Email to a dedicated address
LiveSet up for your pilot. Answered from the same knowledge folder, with the same price check and handover rules as the chat.
Handover alerts
LiveEach handover lands in the cabinet with a summary, and an email copy goes to the handover contact agreed for the pilot.
Leads as CSV
LiveLeads carry the qualification answers and what is still missing. If you stop, the leads and the knowledge folder stay with you.
Phone answering and SMS
PlannedPlanned as a separate desk. Until then, phone enquiries stay with a person or your answering service.
Calendar booking
PlannedUntil a calendar is connected, a booking is a request with the details written down, and a person confirms the time.
Instagram, WhatsApp, Telegram
PlannedOn the roadmap. A link from your profile to the page that carries the web chat is the bridge until then.
Every connection with its status: integrations
What the written enquiries cost you now
The table is the site’s cost model run on this desk. Twelve hours a week is the assumption to change first: it stands for written table and event enquiries, separate from the phone at the host stand.
Read it as hours. The share left over carries judgement — the exception, the upset customer, the decision — and it still needs a person, who now spends the shift on those hours instead of retyping the same answers.
Model: hours and share are assumptions, $35/h sits near the BLS median total compensation (March 2026)
C$3,870
In-house receptionist in Ottawa, per month, before benefits
Job Bank Canada, NOC 14101: median $20/h plus CPP, EI and vacation pay
$29–800
AI receptionist plans for small business, per month, most with minute or call caps
Dialzara, Rosie, Goodcall and Smith.ai price pages, USD, checked 2026-10-31
$0
The Unistaff pilot, 30 days
One AI inbound sales manager, web chat and email, no card
What you write before it starts
The Unistaff team reads your website into a five-section knowledge folder draft. You add the lines below, approve it with your price list, hours, contacts and “never” rules, and the AI reads that folder on every reply.
The menu with prices, specials and the hours for each service.
Group and private-room rules: size limits, set menus, deposits, minimum spend.
The banquet and catering card with “starting at” prices.
The allergen answer your kitchen already gives, written down once.
The “never” rules and the handover contact for complaints and confirmations.
Before release it sits an exam of twelve cases from the trade, each run three times, then works 30 days of probation as a Trainee whose replies you approve. Every sitting of the demo employees is in the exam log.
AI receptionist for a restaurant: questions
It takes the reservation request: date and time, guests, name and phone, occasion and allergies as a note. The guest is told the request is received and the team will confirm. Booking-system and calendar connections are planned, so no table is confirmed by the AI alone.
It is an AI employee with a job description, a knowledge folder, an exam and a weekly review. On your site it looks like a chat, says it is the restaurant's AI assistant and hands over to a person when the rules say so.
Phone and SMS are planned as a separate desk. Today it works on the web chat on your site and on email, so calls stay with the host stand.
It records them as a note for the kitchen and never promises an allergen-free kitchen. For events it asks about the priority allergens. A guest with symptoms is told to call 911 and the conversation goes to a person.
The first 30 days are a free pilot: one AI inbound sales manager for one company on web chat and email, no payment and no card, stop at any time. After it, one flat monthly fee per AI employee is agreed on the pilot review. Messages and minutes are unmetered, and no list price is published.
Event enquiries answered in time, enquiries that became booked events and handovers that reached a person. The product's targets for this job description: a first reply within 15 minutes, a quarter of requests booked, three in ten event quotes closed.
It sits an exam of twelve restaurant cases, each run three times. Pysanka, the group's own restaurant, went from 5 of 12 on October 9 to 11 of 12 on October 11 and graduated. Then it works 30 days of probation as a Trainee.
Try it on your restaurant enquiries for 30 days
The pilot is free, with no card. The Unistaff team sets up the chat on your site and the email address with you, and replies within one business day.
The same desk, another room
01
LiveAI receptionist for coffee shops
The phone is on the bar and the inbox has no owner. Catering and hire enquiries arrive during the morning rush and get answered after close.
02
LiveAI receptionist for salons and barbershops
Bookings and colour questions arrive while every chair is full. The reply that wins the booking is the one that goes out first.
03
On requestAI receptionist for dental practices
The desk is with a patient, the phone is ringing and the inbox holds the new-patient enquiries that fill next quarter's schedule.
04
On requestAI receptionist for auto repair shops
The advisor is under a hood, the phone is on the counter, and every quote request needs seven facts before anyone can price it.
05
On requestMedical clinic
One person covers phones, check-in and scheduling. Requests that arrive during a consultation wait for a gap that never comes.
06
LiveAI receptionist for HVAC and plumbing companies
The no-heat message lands at 9 p.m. while the techs are on calls. The company that answers first books the diagnostic visit.
07
LiveAI receptionist and intake for law firms
New enquiries arrive while the lawyers are in meetings or in court. Each one needs a few answers before a lawyer can tell whether the matter fits.
Every venue on one page: AI front desk for small business
Continue reading
Hub
AI front desk for small business: every venue, one page
Venue
AI receptionist for coffee shops: catering and room hire
Live role
AI inbound sales manager: the Live role behind this page
Evidence
Exam log: how Pysanka went from 5 to 11 of 12
Guide
AI employee vs chatbot: what the difference is in practice
Pricing