Starbucks Ends AI Inventory System After 9 Months Due to Counting Errors

The coffee giant has pulled the plug on an artificial intelligence inventory management system this week, just nine months after rolling it out to all North American locations, according to company documents obtained by Reuters and confirmed by two sources familiar with the matter.

The technology was implemented as part of CEO Brian Niccol’s strategy to address ongoing product shortages that he has identified as a factor damaging the company’s sales performance.

An internal company communication from Monday stated: “Starting today, Automated Counting will be retired. Beverage components and milk will now be counted the same way you count other inventory categories in your coffeehouse.”

The artificial intelligence application, which was created to give Starbucks better insight into store-level shortages, regularly made errors in counting and identifying products, including mixing up different types of milk or completely overlooking items, as Reuters previously documented in February. Company footage demonstrated the system’s failure to detect a peppermint syrup bottle while successfully counting other bottles nearby.

When previously questioned about the technology, Starbucks had maintained that implementing the tool resulted in better product availability at store locations, which represents a key performance indicator in Niccol’s company-wide improvement initiative.

In response to Reuters’ inquiry on Thursday, Starbucks explained that ending the program for milk and beverage products resulted from choosing to “standardize how inventory is counted across coffeehouses as we continue to focus on consistency and execution at scale.” The company added it is pursuing more regular daily deliveries to locations along with ongoing supply chain enhancements.