Read more at:
Amazon warehouses are constantly buzzing with activity as employees stock, pick, and restock merchandise to be shipped to customers. The tech giant has added systems over the years to make the processes even more efficient, including specialized robots to rearrange shelves. The company hasn’t had as much success introducing new technology to manage employee schedules, though. A report from Business Insider discussed how managers pushed back on a new AI-powered labor management system, repeatedly overriding the computer’s recommendations when it clashed with their own judgment.
The main pitfall? Human workers simply didn’t trust the algorithm recommendations, particularly when it appeared to possibly threaten their area’s productivity with too few employees. Many managers, instead, thought the warehouse was too dynamic a work environment for the algorithm to make effective recommendations, and so frequently found what Business Insider sources called “loopholes” to add staff hours. Amazon appeared to struggle with making the system better because constantly overriding the algorithm’s recommendations seemed to keep it from working properly in the first place.
There are examples where technology can work better than a human. For instance, automotive company Stellantis has used AI-powered robots to do a human’s job in a fraction of the time. For every success story like Stellantis, there’s an equally newsworthy failure. In one case, Ford replaced engineers with AI but had to quietly hire them back after the computers failed to replicate the experience of human workers.
Amazon is no stranger to failed AI experiments. It also quietly abandoned a warehouse robot innovation project after just six months due to problematic implementation and high costs.
Will Amazon’s managers prevail over the machine?
Humans and AI butting heads in the workplace is fairly common. Amazon’s experience simply adds another data point to a larger statistic. Executives are pushing AI hard, but some workers are rejecting it with full force. According to a 2026 survey by AI consulting firm Section, 40% of employees in the “trenches” report no time savings from the technology.
What may have hurt the implementation in Amazon’s case is the lack of “guardrails” ensuring the system’s recommendations are actually followed. The company concluded that the recommendations the software was churning out could not lead to the expected benefits if they weren’t followed. Instead, managers actively ignored software recommendations and even went as far as to ask higher-ups to shut off the staffing controls.
Managers reportedly pushed back so much because they believed the system generally struggles with context. For instance, managers Business Insider spoke with said the system may not understand urgency, so it would pull workers away from areas that required extra hands. In another example, a manager complained about the software not accounting for specific strengths or weaknesses between different employees.
The retail giant’s spokespeople defended the system, telling reporters that the transcripts are a small part of the picture, as it reflects a fairly limited set of experiences by a few managers. Business Insider reports that Amazon plans to implement stricter controls that will effectively reduce manual overrides by managers. Whether it will work better once human decision-makers are out of the loop remains to be seen.
