Bahandi: delivery-app orders and automated purchasing on iiko

Challenge

The chain ran on many disconnected systems. Delivery-app orders went through an iiko transit server, and the chain found out about a missing order from the customer.

Every morning, purchase orders for the locations were calculated by hand from iiko data — by the night shift.

What we built

• Order pipelineeach order is first saved on the chain's side, then created directly in iiko; the system checks that it reached the POS and, on failure, recreates it without duplicates.
• iikoFront POS pluginstop lists, order contents and statuses flow from the POS to the delivery apps automatically.
• Auto-orderingnightly exports of stock, sales and write-offs from iiko; by 10:00, tomorrow's order is ready, based on a comparable day, with limits and exceptions.
• Log & monitoringevery order has a step-by-step log; failures are reported to Telegram.
• Back officea dashboard with a map of locations, supplier orders, stock by warehouse, limits and acceptable losses.

Result

Automation fully replaced the manual night work, errors dropped by 70%, and inventory levels fell by 30%. Beyond the original scope, we added stop lists, menu sync to the delivery apps and order-ready statuses for customers.

Technology

• NestJS
• PostgreSQL
• React
• C# / .NET (iikoFront API)
• Telegram Bot API
• Docker