Imran Munawar
AI & AutomationWorking demo2026

AI Ordering Agent

WhatsApp ordering for restaurant chains where the backend owns the money

My role: Designed and built end to end: order engine, agent tools, branch routing, live order board.

In-house product demonstrated with fictional restaurant brands.

Overview

Customers order in plain language, including messy corrections like "two meals, one Coke one Sprite, make the Sprite large, no mayo". The model talks; a deterministic order engine owns menu, pricing, promotions, branch routing, delivery quotes and status.

The problem

Phone and DM ordering is slow and error-prone for multi-branch food businesses. Letting a language model compute prices or apply promotions is a liability.

The solution

Twenty-one typed tools connect the agent to an order engine that resolves menu options, computes totals and promo eligibility, quotes delivery and routes to the nearest open branch. The model never decides a price.

Principal features

  • Understands real orders

    Per-item drinks and sizes, multiple corrections in one sentence, Roman Urdu. Options resolve on the server.

  • Backend owns the money

    Cart, tax, delivery fee and promotions are computed in code; the model relays the numbers.

  • Branch routing

    A typed address or a shared location pin goes to the nearest open branch with the right fee and ETA.

  • Rule-based upsell

    One relevant add-on per conversation, with acceptance tracked.

  • Live order board

    Staff move orders from Preparing to Delivered; each step lands as a WhatsApp message.

  • Reorder and takeover

    "Same as last time" rebuilds a past order. Staff can take over any chat and hand it back.

Technical decisions

  • An event bus decouples the order engine from POS and kitchen-display integrations.
  • Four demo brands run on one engine to prove the menu and promo model is data, not code.

More AI & Automation