Starbucks Ends AI Inventory Counting Program Across North America

Starbucks has retired an automated inventory-counting program used for milk and other beverage components across North America, Reuters reports. An internal newsletter reviewed by Reuters stated, "Starting today, Automated Counting will be retired," and said beverage components and milk will be counted the same way as other inventory categories. Reuters also quoted a company statement that the decision came from a desire to "standardize how inventory is counted across coffeehouses as we continue to focus on consistency and execution at scale." Reuters additionally reported the deployment began nine months earlier and that the tool was part of CEO Brian Niccol's efforts to address product shortages.

Nation's Restaurant News and Restaurant Business describe the retired system as combining computer vision, 3D spatial intelligence and augmented reality, with baristas using a handheld tablet to scan shelves. Nation's Restaurant News reports the company had previously said shops using the technology counted inventory eight times more frequently than those that did not. Both outlets document social-media posts and staff comments that the app frequently miscounted or mislabeled similar milk types and sometimes missed items entirely.

Editorial analysis: Operational AI pilots in retail and food service commonly face two friction points: accuracy at SKU-level granularity and workflow integration with frontline staff. Public reporting on this case highlights both issues-reported miscounts and negative staff reactions-which have been recurring challenges in comparable deployments across grocery and quick-service restaurants.

Industry-pattern observations: Teams building shelf- or backroom-counting systems should expect real-world failure modes where visually similar SKUs, packaging changes, and occlusions reduce recall and precision. Reported outcomes here-retiring an automated counting app after widespread miscounts-underscore the importance of robust edge testing, clear fallback procedures for staff, and rapid feedback loops from store teams into model and UI iteration. Reuters notes Starbucks also emphasized supply-chain and replenishment changes alongside the counting decision, quoting the company: "Our goal is simple - if it's on the menu, customers should be able to order it."

For practitioners: Monitor whether Starbucks or vendors publish post-mortem data on false-positive/false-negative rates, SKU-level confusion matrices, or human-in-the-loop remediation workflows. Also watch deployments that combine frequent replenishment with lighter-weight sensing; public reporting suggests companies often pair process changes with automation rollbacks rather than continuing parallel systems indefinitely.