Starbucks Discontinues AI Inventory Tool and Reintroduces Manual Inventory After Continual Issues Plagued its Performance – CoffeeTalk

Starbucks has decided to discontinue its AI-powered inventory tool named Automated Counting, reintroducing manual inventory counts across its North American stores. This decision marks the end of a technology initiative spearheaded by CEO Brian Niccol, highlighting the challenges faced in integrating AI into retail operations. The Automated Counting tool, developed by NomadGo, utilized tablet-mounted cameras and LiDAR technology to perform automatic counts of beverage components and milk. Despite ambitions to improve efficiency, the system struggled with accurately distinguishing similar products, leading to frequent miscounts and mislabelling issues.

An internal memo declared the retirement of the tool, stating that beverage components would now be counted manually, akin to other inventory categories. Starbucks aimed to standardize inventory counting across locations to enhance consistency and operation effectiveness in its stores. According to company reports, the inaccuracies of the AI tool were deemed a hindrance, and feedback from employees indicated a preference for manual counting despite the innovative intent behind the technology.

Starbucks has faced ongoing challenges with inventory management, being unable to keep its stores consistently stocked, as highlighted by the company’s own delivery performance metrics. This situation has been compounded by broader difficulties in the enterprise AI landscape, where many initiatives fail to yield significant financial returns. Although the company’s recent quarterly sales growth shows promise, the decline in operating margins further complicates its operational narrative.

Moving forward, Niccol continues to invest in technological improvements, including AI tools aimed at optimizing order sequencing and assisting baristas during busy periods. The focus now rests on whether reimplementing manual counts and daily replenishments can resolve the inventory challenges that the AI tool failed to address effectively.

Read More @ TNW