Silpo App
Mobile App · Retail · July 2026
Silpo. Receive goods with one hand.
Silpo App
Retail
Mobile App
Goods receiving moved from three legacy systems and paper into one mobile app. The user path shrank from 21 clicks on desktop to 7 on mobile, and the clerk now only confirms what the system has already recognised — so receiving with one hand became possible.
21 → 7
Fewer clicks: measured on a fixed scenario — receiving without an order, one item.
+
Store visits for research.
4:53
Average interaction time per session.
96.4%
Average scroll depth.
01
Take a photo of the delivery note from the supplier.
02
The system identifies the supplier and the items to be received.
03
The user confirms and accepts the item.
Orders matched to arrivals
The system automatically matches orders to the truck’s arrival time.
Short-delivery report
A report on missing items is created automatically, with no manual retyping.
Sound feedback
Successful and failed actions get their own sounds, so the clerk can keep their eyes on the goods.
Scope of work
Silpo is one of the largest grocery chains in Ukraine. For decades, store operations — receiving, stock transfers between the stockroom and the sales floor, write-offs — ran on desktop systems built for desk work, while store staff work in the stockroom, the cold room, or during stock recounts.
The Silpo operations office asked for goods receiving to move from three legacy systems and paper into one mobile app.
The situation
To understand how receiving really works, I visited stores and watched it on the spot. Invoices can stay open for long periods when the clerk is pulled to other tasks. Each clerk adapts the process to suit themselves. Data goes through an Excel spreadsheet that is later imported into the old system, long quantity lists are sometimes retyped by hand, temperatures are logged on paper, and drivers may wait while documents print.
First hypotheses
Two candidate flows were considered: one linear, one with maximum user freedom. Low-fidelity prototypes were tested with real store staff, and neither option fit well.
The first flow forced people to go through the full process, from creating an invoice to filling it in, every time. The second gave too much freedom — while in 87% of invoices the quantity is already correct.
The idea
What if the system does the routine and the clerk only confirms?
The approach relies on the smartphone camera recognising the delivery note: the clerk photographs it, the system finds the supplier and the items, and the user confirms and accepts the goods.
The outcome
The user path was reduced from 21 clicks on desktop to 7 on mobile. Cognitive load — measured in moderated sessions as how often a person has to stop and decide — dropped too, because much of the work went into removing decisions the system can make itself.
