Sign Up For OPN Connect
When Stefan Kalb was running Molly’s Salads, a grab-and-go company he’d founded in 2009, he noticed that a lot of product was going to waste. While he needed a good amount of inventory to ensure shelves were stocked, Kalb thought there had to be a better way to forecast it. So he reached out to his friend Bede Jordan, then a software engineer at Microsoft, for help creating a solution.
Bede Jordan, Co-Founder and CTO, and Stefan Kalb, Co-Founder and CEO, Shelf Engine
The program Kalb and Jordan developed ended up cutting Molly’s food waste in half—to 13 percent. The pair knew they’d created a valuable tool that could be used more widely, and by 2016, they had turned it into a company called Shelf Engine.
“Our mission is to reduce food waste through automation,” says Amanda Sonenberg, Shelf Engine’s director of partnerships for vendors & associations. “Today, Shelf Engine uses its proprietary artificial intelligence and machine learning technology to forecast and order highly perishable foods for the nation’s top grocers.”
Amanda Sonenberg, Director of Partnerships for Vendors & Associations, Shelf Engine
For decades, retailers have ordered perishable items manually or via computer-assisted ordering software designed for non-perishable inventory. “Both processes are laborious and error-prone, especially when applied across thousands of items,” Sonenberg says.
“Today, Shelf Engine uses its proprietary artificial intelligence and machine learning technology to forecast and order highly perishable foods for the nation’s top grocers.” - Amanda Sonenberg
By contrast, Shelf Engine offers a highly accurate, automated solution that also reduces waste. “Our technology forecasts the precise order volume that optimizes each item’s sales potential,” Sonenberg says. “In doing so, we’re able to drastically reduce the food waste that results from over-ordering. Shelf Engine’s platform makes over one million predictions each week to anticipate consumer demand under a variety of conditions. We then marry these predictions with a store’s historical and daily sales data to generate probabilistic models for each unique SKU we manage in every store, every day.”