Resources / Chat and FAQs

Introduce a restaurant chatbot’s actual scope

A chat greeting sets the visitor’s expectations before the first question.

01

The decision this workflow supports

A chat greeting sets the visitor’s expectations before the first question. If it sounds like a human host with access to live tables, later explanations may come too late. State the assistant’s role and the work it can genuinely do.

02

Put the review process in place

Have the owner approve a short opening that identifies the AI assistant, approved information and request capture. Keep the limitation beside the invitation to ask a question. Avoid promising immediate staff response or platform connections that have not been tested.

03

Worked example for your team

A fictional visitor opens chat and immediately asks for a table tonight. The greeting and first response should establish that Ava can collect a request but cannot confirm availability. Ask the tester to explain what they believe the assistant can do before continuing. If they assume it has access to the reservation platform, revise the introduction. Then test an ordinary hours question, which should receive a concise approved answer without a long warning. The goal is accurate expectations that fit the task, not a repetitive disclaimer before every sentence.

04

Sample wording to approve

“Hi, I am Ava, the restaurant’s AI assistant. I can share approved information and help record a request for staff. I cannot confirm reservations, accept food orders or take payments here.” This wording is a draft for the restaurant owner to approve before use.

05

Fields for the staff handoff

Staff need approved greeting version, supported question categories, staff request route, owner, and review date. Mark uncertain details explicitly and keep the requested action separate from anything restaurant staff have already confirmed.

06

Acceptance tests before use

The greeting identifies the assistant. Booking language remains conditional. The visitor can find the staff-request route without first exhausting the chatbot. Record the observed result, assign any mismatch, and repeat the affected scenario after correction.

Further reading

Catering inquiry value calculator ↗

Try Ava with your business’s public information ↗

Related guides

Your workflow, first

Let’s map your first week.

Bring the calls you want covered. We’ll agree the script, handoff and limits.

Start a pilot