
Amazon presented the results of testing new robots for the warehouse — Stow and Pick, which are supposed to automate the packaging and selection of goods. Despite promising indicators, robots are still inferior to people in accuracy and speed of work.
Amazon studies indicate that the Stow robot, responsible for placing goods, achieved an accuracy of 85%, while workers — 243 units per hour against 224 in Stow. The Pick robot, which is engaged in the selection of items from shelves, had a success rate of 91% out of 12,000 attempts, but in 19% of cases refused to work — due to unrecognizable objects or the risk of damage.
The systems have machine vision, predictive algorithms, as well as manipulators for careful grasping of goods. They already demonstrate performance close to human, but still need correction, especially in conditions of unpredictability or deformation of packages.
Amazon has long invested in logistics automation. For many years, the company has used robots to transport goods, but now it is trying to completely replace manual labor even in the most delicate tasks – sorting and packaging. In 2024-2025, a test operation of a new generation of robots was launched in real warehouse conditions.
Although Amazon is actively promoting the robotization of warehouses, humans remain more effective in a number of critical operations. The success of robots is growing, but mass implementation is still ahead. People remain indispensable in complex situations that require flexibility, recognition and decision-making.