AI Coffee: QR ordering and AI pastry pairings for specialty coffee shops

01 / Overview

About AI Coffee

02 / The challenge

What AI Coffee needed to solve

Coffee shop owners want to sell more drink and pastry pairs, but they need tools that are simple enough to run without technical help.

1. 01

   A good barista knows which pastry suits which coffee, but that knowledge sits in a few heads.

2. 02

   Guests had to be able to order from a QR code on the table without installing an app.

3. 03

   Owners wanted a short menu, orders that reach the bar and a few numbers on what people buy, without needing any technical skill.

4. 04

   One operator had to look after many shops, each with its own menu, QR code and pairing rules.

03 / What we built

How we built AI Coffee

We built a React web app and an Express and MongoDB API. Each shop gets a public menu at its own address, the owner runs orders from a dashboard, and a pairing lab asks a language model for scored suggestions.

• 01

  A menu page for every shop

  Each shop has a public menu at its own address, with coffees and pastries, tasting notes and prices. The page carries the shop's QR code and a share link, so guests can open it from the table or send it to friends.

• 02

  Tray and checkout

  Guests fill a tray, adjust quantities and send an order for a table or for pickup, with a notes field for allergies or brew preference. The server checks every price before an order is saved. The prototype takes no payment and simulates the guest journey.

• 03

  Live orders for the bar

  The owner dashboard lists live orders with the customer, table, items, total and last update. Staff move each order from pending to preparing, ready and completed.

• 04

  AI pairing lab

  Staff pick a coffee focus, a pastry focus and a season, and a language model returns three pairings. Each one comes with a score, the margin, a reason and ready-made marketing lines, and the results export as JSON, CSV or HTML.

• 05

  Owner overview and QR settings

  The overview shows total products, orders received, gross sales and the most popular items. Owners download the table QR code as a PNG and edit the shop description and location that guests see.

• 06

  Admin console

  An operator sees every shop with its address, owner and status, and can mark a shop inactive or delete it. Platform analytics show shops, orders, revenue and products for the last 30 days.

• 07

  Pairing rules

  Admins write pairing rules for each shop and season, such as cinnamon and cardamom pastries with darker roasts in autumn. Each rule can be enabled, edited or deleted.

• 08

  Event tracking

  Nine events are tracked, from QR scans and page views to pairings viewed, items added to the cart, checkouts started and orders placed. The admin console shows the count for each one.

04 / The results

What changed for AI Coffee

The prototype joins ordering, a table QR code and AI-written pairing advice in one product.

• An owner can set up a menu without technical help.
• Guests order for a table or for pickup without installing anything.
• Staff get pairing ideas with a reason and a sales line they can put to use straight away.
• Orders reach the bar and move from pending to ready on one screen.
• One admin console covers every shop, product and pairing rule, with tracked events from scan to order.