Retellect detects wrong product selections at the lane and speeds up checkout flow — reducing shrinkage, staff interventions and customer friction on the checkouts.
At self-checkout, shoppers pick loose items like fruit and vegetables from a list on the screen. Sometimes the wrong item is chosen — by mistake, like peaches selected as apricots, or on purpose, like expensive alcohol rung up as cheap potatoes. The self-checkout system has no way to tell. That is the loss Retellect is designed to detect.
Rimi
StrongPoint
Overkill
Retellect starts with the core layer: item recognition and picklist loss prevention. From there, we tune the system around your checkout process, product catalogue and store data — then extend into additional scenarios where the signal is clear and the value is measurable.
We identify the checkout flows, product groups and loss patterns where Retellect can create the most value.
We test the recognition layer in your real checkout environment: your products, camera setup, checkout events and store conditions.
We track recognition accuracy, false positives, wrong-pick patterns, staff interventions, checkout flow impact and operational value.
Additional checkout scenarios are co-developed around your highest-value loss or flow problems, where the data supports a measurable outcome.
Retellect's AI is live at self-checkout lanes in three Rimi stores in Lithuania — 98% recognition accuracy on weighed produce, ~10 seconds saved per item, and a foundation for real-time loss prevention.
Read articleRetellect received its first investment from Overkill Ventures Fund following the Accelerator Program — accelerating go-to-market for retail checkout AI.
Read articleA conversation about your stores and where Retellect can help.
