Starbucks to Discontinue AI Inventory System
Starbucks has decided to phase out its AI-driven inventory management tool less than a year after its launch, responding to numerous complaints from employees regarding its accuracy and frequent errors.
The coffee company previously introduced automatic counting software, developed in collaboration with technology firm NomadGo, to all North American locations in September 2025. The goal was to make inventory tracking easier by allowing staff to use mobile devices to scan shelves and identify items like different types of milk and syrups.
Initially, the system aimed to revolutionize inventory management, cutting down on the time employees spent on manual counts and improving accuracy. In a now-deleted blog post, Deb Hall Lefebvre, Starbucks’ chief technology officer, excitedly endorsed the new technology, emphasizing how it would allow partners to quickly check their supplies, ensuring that popular items were always in stock.
However, the reality fell short of these ambitious expectations. Many employees reported significant issues, such as the system incorrectly labeling products and miscounting stock. It often confused similar milk types or failed to recognize items entirely. The situation deteriorated to the extent that Starbucks opted to end the program after just nine months.
Interestingly, a promotional video shared in a September blog post had hinted at potential pitfalls, showcasing the system misplacing a bottle of peppermint syrup during an inventory check, which foreshadowed the struggles that would follow.
With the decision to scrap the automated system, Starbucks staff will revert to traditional manual inventory counting methods. An internal newsletter indicated that beverage ingredients and milk would be itemized just like other inventory categories.
This experience underscores the complexities businesses encounter when integrating AI solutions in real-world settings. Although AI technology is evolving and finding success across various fields, this instance illustrates that not all implementations yield the expected outcomes. The difference between what AI can theoretically achieve and what it accomplishes in practice can be quite significant, especially in the intricate world of retail.
Starbucks has been exploring AI and automation to boost its sales, especially in light of recorded declines. The company has also begun deploying new technologies across U.S. stores, including AI robots taking orders at some drive-throughs, and in-store Assistants that assist baristas with recipes and scheduling. Additionally, the old manual inventory counting practices were replaced with scanning tools to address ongoing stock issues.
