Starbucks is pulling the plug on its AI-powered inventory management system after just nine months of deployment. The coffee giant rolled out the Automated Counting software across North American stores in September 2025, partnering with NomadGo to automate tedious inventory tracking through mobile device scanning.

The tool promised to streamline supply chain operations by instantly identifying stock levels of popular items like oat milk and caramel drizzle. However, employees reported frequent mislabeling and miscounting issues, with the system mixing up similar milk types or skipping items entirely. A promotional video inadvertently revealed the technology failing to detect a peppermint syrup bottle during scanning.

Workers welcomed the return to manual counting methods, expressing relief that the problematic system was discontinued. One employee noted the concept was sound but execution proved challenging. This failure highlights the gap between AI hype and real-world retail implementation, showing that automation doesn't always deliver promised efficiency gains in complex inventory environments.

For Starbucks, the decision represents a costly lesson in deploying unproven technology at scale. The company will revert to traditional inventory methods while competitors continue exploring AI solutions.