🟢 The AI first-mover disadvantage

In chaotic markets dealing with unproven technology, companies currently spending the most on AI process automation are doing their competitors' R&D for them.

🟢 The AI first-mover disadvantage
Photo by Elisa Kennemer / Unsplash

Starbucks spent nine months deploying an AI inventory management tool across its North American locations, only to quietly scrap the whole thing just a few weeks ago. The app, which tracked beverage components like milk and syrups, hallucinated inventories by failing to identify bottles on shelves, and required stores to rearrange back-of-house storage (a time-intensive process that made employees' workflow more challenging than necessary). To quote one shift supervisor with nine years at the company:

It started off not particularly accurate and got less accurate over time.

In the same spirit, McDonald's ran voice-ordering AI across more than 100 drive-thrus for three years, generating an endless flow of viral TikTok footage of customers receiving nine sweet teas and bacon on their ice cream. Why? Mostly, the system pulled in audio from adjacent lanes, then sometimes misread regional accents, and (wait for it...) couldn't parse families ordering simultaneously. The plug was pulled in July 2024.