Human-centered design systems and product experiences
Core services
01UI/UX Design02Web Development03Software Development042D/3D Experiences05App Development06AI Agents & Automation07Website DesignTechnical services
Frontend Engineering
React DevelopmentNext.js DevelopmentWeb ApplicationsEnterprise DashboardsSaaS PlatformsBackend & Infrastructure
API DevelopmentCloud InfrastructureDevOps & CI/CDBackend SystemsCustom SoftwareIndustries we serve
Technology & Software
AI Restaurant Orders and Reservations Agent Development
Orders prepared, tables requested
An agent for a restaurant's messages and calls.
An agent for a restaurant's messages and calls. It answers questions about the menu from your own menu data, takes a reservation against real table availability, prepares an order for the guest to confirm, and passes allergy questions to your staff.
During service nobody can answer the phone. A guest wants a table for six on Friday, another asks whether a dish contains nuts, a third wants to order for collection. The calls go unanswered, the messages are read after closing, and the booking goes to the restaurant that picked up. Delivery platforms solve the ordering and take a large share of every sale.
This page describes a workflow we propose to build as a custom project. It does not exist as software yet. It would reuse the order draft and the catalogue lookup that run in our WhatsApp demonstration on test data, where an order is prepared and held for confirmation, and the booking logic from our appointment booking demonstration. Menu data, table availability and connection to a point-of-sale system are not built.
From the First Signal to the Record
An example of one case, in 6 steps. The same steps play in the picture at the top of this page.
Guest writes
A guest messages asking for a table for four on Saturday at eight.
Availability checked
Eight is full. Seven thirty and nine are offered from your reservation system.
Booked
The guest takes seven thirty. The booking is confirmed with the address.
Question asked
They ask whether the lamb dish contains nuts. The published allergen sheet is shown, and the question is passed to staff.
Order prepared
Another guest orders for collection. The items are read back and confirmed.
Reminder sent
On Saturday afternoon the party of four is reminded and confirms.
Six jobs it carries through
Menu answers
Answers from your menu data.
- Dishes, prices and descriptions
- Opening hours and location
- Today's availability, where you update it
- Says when it does not know
Allergen handling
Treats allergies as a matter for staff.
- Shows the allergen information you publish
- Never improvises about ingredients
- Passes specific allergy questions to staff
- Notes the allergy on the order or booking
Reservations
Takes a booking against real availability.
- Party size, date and time
- Table availability from your system
- Special requests noted
- Confirmation and reminder
Order preparation
Builds an order for the guest to confirm.
- Items and modifications read back
- Sold-out items caught
- Collection time from your kitchen's rule
- Held until the guest confirms
Large and special bookings
Routes what needs a person.
- Groups above your size limit
- Private events
- Deposits and set menus
- Passed to the manager with details
Changes and no-shows
Keeps the book accurate.
- Easy changes and cancellations
- A reminder before the booking
- Late cancellations per your policy
- Released tables offered again
What We Need to Scope This
- The process as it runs today, in a few lines.
- The systems it touches, for example reservations, point of sale, menu, guest channels.
- Rough monthly volume, and who approves what.
Six Things That Slow This Down
Calls during service
The phone rings while every hand is busy.
Tables booked on paper
Two parties are given the same table.
Allergy questions answered from memory
A guess about ingredients by whoever picks up.
Orders taken wrong
A modification is misheard over a noisy line.
Sold-out dishes
The order is accepted and the kitchen has none left.
No-shows
A table for eight is held and nobody arrives.
The Systems Around It
The systems named here are the kind this agent works with. Which ones it connects to in your case is settled in the scope.
Reservations
- OpenTable
- SevenRooms
- ResDiary
- or your booking system
Point of sale
- Your till and kitchen order system
Menu
- Dishes, prices, availability and published allergens
Guest channels
- phone
- your website
Payments
- Your own payment provider, for deposits or orders
Staff
- The manager on duty
It Answers From the Menu and Leaves Allergies to Your Staff
Food allergy is a safety matter. The agent shows the allergen information you publish and passes any specific question to a member of staff. It does not reason about ingredients, substitutions or cross-contamination.
A reservation or an order is confirmed only when your own system accepts it, and the guest confirms an order before it reaches the kitchen.
- Answers only from your current menu
- Published allergen information only
- Allergy questions passed to staff
- Bookings confirmed by your system
- Orders confirmed by the guest
- Your cancellation policy as written
Straight Answers
It shows the allergen information your restaurant publishes and records the allergy on the booking or order. Specific questions go to your staff, and the guest is told that. It never says a dish is safe.
A restaurant's work is tables and food orders, with allergens as the main risk. A hotel's is rooms, stays and requests during the stay. A hotel restaurant can use this agent inside the other.
It can take orders through your own channels, such as WhatsApp or your website, for collection or your own delivery. Whether that suits you is a commercial decision. It does not connect to a delivery marketplace unless you ask for that.
Not as software yet. This page describes a proposed workflow. It would reuse the order draft and the catalogue lookup that run in our WhatsApp demonstration on test data, where an order is prepared and held for confirmation, and the booking logic from our appointment booking demonstration. Menu data, table availability and connection to a point-of-sale system are not built. A project starts by mapping your process, then proving the first steps on test data before anything touches a live system.
Count Friday Night's Missed Calls
And what they wanted. Most restaurants find that reservations and opening-hours questions make up nearly all of them.