Starbucks has withdrawn the artificial intelligence inventory tool it rolled out in its North American stores last September, according to an internal memo reviewed by Reuters and confirmed by the company.
See also: Starbucks: Data breach affects employees
“Starting today, Automated Counting will be retired,” the memo said. “Beverage ingredients and milk will now be counted the same way other inventory categories are counted in your coffee shop.” In other words, by hand.
The tool, developed by Seattle-based NomadGo , used tablet-mounted cameras and LiDAR technology to scan shelves of syrups, milks and other beverage ingredients and produce automatic counts, replacing manual inventories for select categories. It had been in development for several years and was expanded nationally after Brian Niccol took over as CEO in September 2024 as part of his “ Back to Starbucks ” strategy .
The problem, according to a Reuters report in February and internal company documents, was that the tool struggled with the everyday task of distinguishing one white liquid from another. The app often miscounted or mislabeled products, especially similar-looking products like oat milk and cow’s milk. A promotional video released by Starbucks at launch showed the system failing to register a bottle of mint syrup sitting on the shelf as it counted the bottles next to it.
In a statement to Reuters, Starbucks described the move as a standardization exercise, not a retreat. The decision stemmed from “a decision to standardize how we count inventory across all our stores as we continue to focus on consistency and execution at scale,” the company said, adding that it was moving toward more frequent daily replenishments and ongoing improvements to its supply chain.
See also: Starbucks: Suffered ransomware attack via third-party software provider
An internal memo shared by the company included an employee thanking the team for scrapping the program: "The thinking behind it was great, but the execution proved difficult."
The decision matters because inventory was supposed to be the easy part. Four Starbucks CEOs in five years have blamed lost sales on the company’s failure to keep stores adequately stocked. By early 2024, according to the company itself, less than a third of deliveries to Starbucks distribution centers were arriving on time and complete.
Automated Counting was meant to give the chain the immediate store-level visibility it had been lacking, and was one of Niccol’s key operational solutions. It also comes at a time when the broader story of business AI is starting to look less than stellar. NANDA found last year that 95 percent of enterprise AI pilots failed to deliver a measurable impact on bottom lines, despite spending an estimated $30 billion to $40 billion, with only 5 percent reaching production.
The economic backdrop is mixed enough that the decision could be read both ways. Starbucks reported its strongest quarterly sales growth in two and a half years last month, and the stock is up 24% so far in 2026, but operating margins in its core North American market have fallen to 9.9%, from 18% two years ago.
See also: Mercedes-Benz: EV fast charging stations are coming to Starbucks
Niccol continues to invest in other technology options, including AI.
