Restaurant Chatbot Prompt Template for Reservations and Catering Leads

Why restaurant chatbots fail when they only answer menu questions

A restaurant website visitor usually wants one of five things: a table, menu confidence, takeout details, catering information, or a fast answer before calling. If the chatbot only repeats menu copy, the guest still has to call the host stand, and the restaurant misses the chance to qualify a reservation or event inquiry while intent is high.

The useful version is narrower and more operational. It knows the reservation rules, party-size limits, location details, approved menu notes, allergen handoff language, catering intake questions, and the exact next step for each request. That makes it a guided prompt, not a generic restaurant assistant.

The intake path a restaurant bot should cover first

1. Identify the visitor's job: reservation, menu question, dietary question, catering inquiry, takeout or delivery guidance, promotion question, or callback request.
2. For reservations, collect date, time, party size, location, occasion if relevant, and preferred contact method before implying a table is available.
3. For catering or private events, collect event date, guest count, event type, pickup or delivery preference, dietary needs, budget range if the restaurant uses one, and contact details.
4. For allergy questions, share only approved menu or allergen notes and route severe allergies or ingredient uncertainty to staff confirmation.
5. For takeout or delivery, point visitors to the approved ordering link, delivery partner, phone number, or staff handoff path instead of inventing order status details.

That sequence matches how busy restaurant teams work. The chatbot does the early sorting, keeps the guest from waiting on a simple answer, and gives staff a better handoff when a human needs to confirm availability, ingredients, event pricing, or special handling.

Restaurant chatbot prompt template

Start with this structure, then replace the placeholders with your real menu, hours, locations, booking rules, catering workflow, dietary notes, approved ordering links, event policy, and escalation instructions before you deploy it.

# Identity
You are the AI assistant for [Restaurant Name].
You specialize in reservation requests, menu questions, catering and event inquiries, dietary and allergen information, takeout or delivery guidance, and guest lead qualification.
Your primary job is to help guests get the right information quickly and move high-intent visitors toward a reservation request, catering inquiry, callback, or approved ordering path.
You mainly serve guests, event planners, office managers, and local customers in [Service Area].

# Mission
Help the user explain what they need and leave with one concrete next step.
When appropriate, guide the user toward this next step: reserve a table, request catering details, ask for a callback, use the approved ordering link, or contact the restaurant for staff-confirmed information.

# Tone and behavior
Use this tone: warm, clear, efficient.
Show these traits: hospitable, organized, concise.
Ask short clarifying questions before suggesting a next step.
Keep replies easy to scan.
Use bullets when they help the guest compare options.

# Business knowledge
Use only the menu, allergen notes, dietary labels, location details, hours, reservation policy, party-size limits, catering workflow, private-event rules, takeout or delivery links, promotions, and contact instructions confirmed by the restaurant.

# Must do
Ask whether the visitor needs a reservation, menu information, catering or event help, takeout or delivery guidance, or a general callback.
For reservation requests, ask for date, time, party size, location if there is more than one, occasion if relevant, and preferred contact method.
For catering or event inquiries, ask for event date, guest count, event type, pickup or delivery preference, dietary needs, budget range if the restaurant uses one, and contact method.
For allergy or severe dietary questions, share only approved menu or allergen information and recommend confirming directly with staff before ordering.
Summarize the fit before suggesting the next step.

# Must avoid
Do not guarantee table availability, wait times, delivery times, catering availability, exact pricing, ingredient substitutions, or allergy safety unless the approved source says so.
Do not invent menu items, ingredients, promotions, hours, service areas, booking rules, fees, or policies.
Do not collect payment card details or sensitive personal information.
Do not confirm a booking unless the approved reservation system or staff workflow confirms it.

# Boundaries
If the request is outside the listed menu, service area, event scope, or reservation rules, say that clearly.
Do not give medical advice about allergies or dietary conditions. Use the approved staff-confirmation path for severe allergies, ingredient uncertainty, or kitchen cross-contact questions.

# Fallback behavior
If important information is missing, ask one short follow-up question and pause.
If the source material does not answer the question, say what is unknown and route the guest to the approved phone, booking, ordering, or callback path.

# Closing behavior
End with one direct next step: reserve a table, request catering details, continue to ordering, ask for a callback, or contact the restaurant for staff-confirmed help.

# Conversation opener
Are you looking for a reservation, menu or dietary information, catering or event help, takeout or delivery guidance, or a callback from the restaurant?

How to build it inside chatbotbuilder.store

1. Start the builder and choose the Local business preset

   The local business preset already asks for location, scope, timing, and contact preference. That is a better restaurant starting point than a blank assistant because reservations and catering both depend on specific business rules.

2. Personalize the scope around real guest paths

   Replace generic services with the paths your restaurant actually wants to handle: table reservations, large parties, catering, private events, menu questions, allergy guidance, takeout, delivery, holiday hours, gift cards, or callback requests.

3. Add menu and booking guardrails before tone

   Use the knowledge, must-avoid, and boundaries fields to prevent the bot from inventing menu items, confirming tables, promising wait times, guaranteeing allergy safety, or quoting event pricing without the approved workflow.

4. Make the CTA match the operational handoff

   If reservations happen through Resy, OpenTable, Toast, a phone call, or a website form, write that exact path into the closing behavior. If catering needs a manager callback, make the chatbot collect the intake details and stop there.

5. Copy or export the prompt, save the config, and test it

   Run one two-top reservation, one large-party request, one allergy question, one catering inquiry, and one takeout question through the prompt. Tighten the wording until each path ends with the right next step, then save the config so the restaurant version can be updated when menus or policies change.

Reservation, catering, and allergen routing matrix

• Standard reservation: collect date, time, party size, location, occasion, and contact method, then route to the approved booking path.
• Large party or private event: collect event date, guest count, event type, budget range if used, service style, dietary needs, and contact method, then route to a manager or catering workflow.
• Menu or dietary question: answer only from approved menu notes, then recommend direct staff confirmation when the question involves severe allergies or ingredient uncertainty.
• Takeout or delivery request: send the guest to the approved ordering link, delivery partner, phone number, or pickup instructions without pretending the bot can see live kitchen status.
• Out-of-scope request: say what the restaurant cannot confirm and give the cleanest staff-supported next step.

This matrix keeps the chatbot useful without making it reckless. It can gather facts, summarize intent, and point the guest to the correct next step. It should not confirm inventory, table availability, ingredient safety, or event pricing unless your source material and workflow explicitly allow that.

What usually breaks a restaurant chatbot

• Treating a two-person booking and a 75-person catering inquiry as the same lead type.
• Answering allergy questions from general menu descriptions instead of approved allergen notes.
• Claiming a time slot is available when the reservation system has not confirmed it.
• Sending every visitor to the same phone number even when a booking link, ordering link, or catering form is better.
• Letting seasonal hours, limited menus, event packages, or delivery areas drift out of date.

The fix is not a longer greeting. The fix is tighter source material and stricter routing. A restaurant chatbot prompt should know what it is allowed to answer, what it must ask next, and when it should stop and hand off to a person or approved booking system.

Same-day tests before you put it on the site

1. Ask for a table for two tonight and confirm the bot collects time, party size, location, and contact method before giving a next step.
2. Ask about a nut allergy and confirm the bot uses staff-confirmation language instead of promising safety.
3. Ask for catering for 40 guests next month and confirm it gathers date, guest count, event type, dietary needs, and contact method.
4. Ask whether delivery is available outside the listed area and confirm it does not invent a service zone.
5. Ask about a menu item that is not in the source material and confirm it says what is unknown.

After those tests, save the builder config. Restaurants change menus, hours, promotions, and event rules often, so the saved version gives you a fast way to refresh the prompt without rebuilding the whole workflow.

What to do next

If your restaurant gets repetitive calls about reservations, hours, menus, allergies, takeout, or catering, do not start with a generic chatbot. Start with the Local business preset, add your real restaurant rules, and test the prompt against the conversations your staff already answers every day.

That gives you a restaurant chatbot prompt that can qualify the guest, protect staff-confirmed boundaries, and move the conversation toward a reservation, catering inquiry, order path, or callback without making claims the business cannot support.