Shiny Fridges, Misread Bottles Doom New Starbucks AI-Powered Inventory System
Last September, Starbucks rolled out a new automated AI-assisted inventory counting system, which they believed would cut down on counting product considerably. Now that it's been junked, the fallout continues.
Seattle Area Start Up Sent Reeling by Starbucks Dumping the Program
NomadGo is a Seattle company that developed the program which cuts down the time it take for product inventory. Founded in 2017, NomadGo's system uses AI-assisted tech to rapidly scan items, according to Geekwire:
"The company’s tech runs on-device and counts each item within rows. It displays results on the device and syncs data with existing inventory management platforms."
The program, called Green Dot Assist, was put into place in 11,300 Starbucks' locations, to allow workers to scan and log product on shelves. However, 9 months later, it's gone.
What Happened in the Stores?
Geekwire says the idea was to turn an hour or two long process into one that only took 10-15 minutes. But despite a 99 percent success rate in strenuous testing, real-world conditions affected the program. According to Geekwire:
"Baristas reported camera errors — such as shiny refrigerator reflections doubling milk counts or the app misidentifying syrups and trash cans — while stores with spotty Wi-Fi frequently had their counting progress wiped out entirely mid-scan."
Was Coffee Giant's Back-End System Incompatible with New Tech?
Officials also discovered Starbuck's backend network utilized systems dating back to the 1990's making it hard to process modern AI-backed tech. Starbucks abruptly pulled the plug in May, and now NomadGo has been left reeling.
They've had to lay off much of their 30-person workforce, and not long ago Starbucks informed workers to rip off all the barcode tags with the system. They're now back to the 'old' method of inventory. Starbucks does continue to use Green Dot Assist for other duties, just not inventory.
KEEP READING: Scroll to see what the big headlines were the year you were born
Gallery Credit: Andrew Lisa