Starbucks retires AI inventory tool after workers report errors

Starbucks has shut down its worker-facing AI inventory system just nine months after rolling it out across North American stores, a sharp reset for a tool that was supposed to cut manual work and help prevent stockouts. Instead, employees described a system that miscounted product, mislabeled items and added friction to one of the most time-sensitive tasks on the floor: knowing what is actually in the cooler, the syrup rack and the back room before the rush hits.

An internal company newsletter said, “Starting today, Automated Counting will be retired,” and instructed stores to count beverage components and milk the same way as every other inventory category. Starbucks said the move was meant to “standardize how inventory is counted across coffeehouses” and support consistency and execution at scale. The company also said it remains committed to investing in technology, even as it refines which tools hold up in real stores.

The problems with the system went straight to the daily realities baristas and shift supervisors deal with. Reuters reported that the software frequently confused similar milk types or missed them altogether, which matters when a store is trying to place accurate orders and avoid running out of the one carton or base needed to keep drinks moving. A Starbucks video also showed the tool failing to recognize a peppermint syrup bottle while it counted neighboring bottles, a small but telling example of how an automation promise can turn into extra correction work for partners already balancing drive-thru times, mobile orders and lobby traffic.

The rollback also fits Brian Niccol’s larger push to fix the chain’s product availability problems. Niccol has said Starbucks wants daily replenishment by the end of calendar year 2026, tying that goal to shortages that have hurt sales. At Starbucks’ March 25 annual meeting, the company said it had invested more than $500 million in hours and tools for Green Apron Service, its store-operating push focused on throughput, accuracy and connection. Starbucks also said hourly turnover was less than half the industry average and that it receives more than a million applications for barista jobs every year, underscoring how much the turnaround depends on store labor.

For workers, the significance is practical rather than flashy. Inventory systems affect ordering, waste, shrink and the amount of time a shift supervisor spends reconciling what the screen says with what is actually on hand. By retiring the AI count tool and moving back to a single process, Starbucks is signaling that store-floor reliability now matters more than novelty. The company may still be betting on technology, but this time the bar is simple: if it cannot count a bottle of syrup or the right kind of milk, it does not save labor, it creates it.