Use Case
AI Restaurant Phone Order Taker
A busy neighbourhood restaurant loses orders every Friday night for one unremarkable reason: nobody can reach the phone. An AI persona built as an order taker answers every time, works through the menu with the guest, and hands a clean, checked ticket to the kitchen.
By EchoMP Editorial · Published 22 August 2026 · Updated 22 August 2026

The Scenario
The restaurant places the persona on its website ordering page and shares the same link through its social profiles and printed materials. A guest opens it at seven on a Friday evening and starts a live face-to-face interaction rather than waiting on hold.
The persona greets them, confirms whether the order is for pickup, works through the items one at a time, checks the modifiers, asks about allergies, reads the whole order back, quotes the total and a ready time, and sends the ticket through.
How the Persona Is Set Up
Essentials
The role is a single-restaurant order taker: warm, efficient, and never chatty during a rush. The system prompt makes three rules non-negotiable — never invent a menu item, never guess a price, and always read the full order back before confirming.
Knowledge
- The full menu with exact item names, descriptions, and prices
- Available sizes, modifiers, sides, and substitutions
- Allergen and dietary information, written per item
- Opening hours, kitchen closing time, and typical preparation times
- Pickup instructions, parking notes, and the address
- Rules for large orders, catering enquiries, and busy-period wait times
Menu knowledge is where accuracy is decided. Written as clean question-and-answer entries with item names spelled exactly as they appear on the menu, the persona quotes rather than paraphrases. See how knowledge and Skills work.
Guided Sessions
A "Place a pickup order" session enforces the sequence: items, modifiers, allergy check, name and phone number, read-back, confirmation, ready time. A "Menu question" session answers ingredient and dietary queries without starting an order. A "Large order enquiry" session collects the details and routes them to a manager.
Skills
- Order submission to the kitchen or point-of-sale system
- Menu availability check for sold-out items
- Wait-time lookup during peak service
- Customer record creation for repeat guests
- Payment handling where the restaurant has enabled it
- Handoff to a staff member for anything unusual
The Read-Back Rule
The single most valuable behaviour is the read-back. Before anything is submitted, the persona repeats the complete order, item by item, with modifiers and the total, and asks the guest to confirm.
It also states prices and quantities in spoken form so that nothing is compressed or misheard, and confirms the pickup name and phone number by repeating them back.
Handling the Awkward Moments
- An item is sold out — the persona says so and offers the closest alternative on the menu
- A guest asks for something that is not on the menu — it declines rather than inventing it
- A serious allergy is mentioned — it records it on the ticket and offers a staff member
- The kitchen is about to close — it states the cut-off before taking the order
- The request becomes a complaint or a refund — it hands off to a person
What the Restaurant Gets
- Every call answered during the rush, with nothing lost to a busy line
- Legible, structured tickets instead of hurried handwriting
- Fewer wrong orders, because the read-back catches them before the kitchen does
- Staff kept on the floor rather than on the phone
- A record of what guests ask for, which quietly informs the next menu
The same persona can be embedded on the restaurant's site or shared as a link — see share, embed, meetings, and earnings.
Related Use Cases
Put an order taker on the front of your phone line.
Load the menu, set the rules, and let every guest reach someone who answers on the first ring.
