Beyond Analytics

A shift to more widespread use of artificial intelligence and machine learning promises to streamline supply chain operations.

By Nick Fortuna

November 11, 2017, is a date that sticks in the mind of Jim Tompkins, chief executive of MHI member Tompkins International, and he’s working hard to make the events of that day commonplace in the supply chain industry.

Two days earlier, Michael Zakkour, a vice president with Tompkins International, was in China to speak as one of the “global influencers” identified by the Chinese retail giant Alibaba as leading experts in global e-commerce. Zakkour was talking with Joseph C. Tsai, Alibaba’s executive vice chairman and co-founder, about the November 11 Singles’ Day Shopping Festival, when Alibaba offers deep discounts to produce the single-biggest day of retail sales in the world.

Tsai gave Zakkour an estimate of Alibaba’s sales revenue for Singles’ Day, and two days later, when the record figure of $25.4 billion was in the books, Tsai looked like a genius. His estimate had been off by only several hundred million dollars, a feat Tompkins compared with predicting the score of a basketball game within a point or two. Even more impressive, Tsai had accurately predicted the sales figures for each region that Alibaba serves allowing the company to optimize its inventory allocation among its distribution centers.

Needless to say, Tsai wasn’t using psychic powers to predict sales figures. Alibaba has created an artificial intelligence (AI) program that uses real-time sales data to increase Alibaba’s margins and eliminate lost sales—those times when a customer wants to buy a sold-out item or when that item is warehoused so far from the shipping destination that the margin is eroded and the company might as well have lost the sale.

“Alibaba did a phenomenal job of predicting sales, and I was really impressed with how far they had come with it,” Tompkins said. “Artificial intelligence represents a huge opportunity for us to make supply chains work better, so I’m excited about it.”

In early January, Tompkins was focused on developing his own demand-shaping AI program in time for use during the 2018 holiday shopping season at five of Tompkins International’s MonarchFx distribution centers. Using data such as past sales figures and weather forecasts, Tompkins was planning to develop several different sales forecasts and then evaluate which one proved most accurate. That set of computer algorithms then would be used to optimize MonarchFx’s inventory allocation during the holiday season.

Read More…