Starbucks Scraps Automated Inventory Management AI

Starbucks CEO Brian Niccol has officially scrapped the 'Automated Counting' AI program that had been introduced to track inventory at North American stores.

The program was designed to quickly scan ingredients and determine quantities using tablet cameras and Lidar sensors, but it failed due to technical limitations in real store environments. Key failure cases include:

• Failing to distinguish between almond milk and oat milk
• Failing to recognize and missing syrup bottles

Starbucks had claimed that the technology improved product availability, but recently removed the related content from its website. Meanwhile, Starbucks' regional operating margin has plummeted from 18% to 9.9% over the past two years, and the company recently laid off 300 corporate employees.

This case shows that fully automating physical inventory management using computer vision remains a very high technical hurdle. Starbucks plans to return to its existing manual inventory management method for the time being, while addressing supply shortages through a new daily replenishment model.