Bringing AI to inventory optimization

I don’t often interview solution providers for this column, but last week I had a call with Sivakumar Lakshmanan – Siva to the industry. Lakshmanan is the CEO of, a cloud-based software company, now part of Zebra Technologies, which provides what we used to call multi-echelon inventory optimization right down to the shelf level for retailers. describes what it provides as AI and machine learning-enabled inventory management and allocation solutions specific to forecasting and merchandising for the retail and CPG industries.

I was prompted to have the conversation by two factors. The first is an interview I did in September 2021 with Tom Bianculli, Zebra Technologies’ chief technology officer. Zebra had recently announced the acquisition of, which at the time was one of four acquisitions, including Fetch Robotics and Reflexis Systems, that were taking a company rooted in data collection and mobility into new digital realms. I wanted to understand the strategy.

The second was a brief conversation I had at the booth at this year’s Gartner conference in Orlando. It was the first Gartner event since 2019, and walking the exhibition hall, it was clear to me that three years in technology is a lifetime. What I realized talking to exhibitors is that there are a rash of new companies, new offerings and new approaches to supply chain planning and execution software since those last events, and I wasn’t sure I understood them all or where they fit. In particular, I wanted to learn more about what AI and machine learning are bringing to the table.

The first thing I learned is that has been around a lot longer than I realized. AI was not a thing when it was founded almost 10 years ago in 2013, but the idea was to create “a strong analytics firm that helps retailers,” Lakshmanan said. With a $56 million investment from Goldman Sachs, acquired 5 or 6 other companies and by 2018 had gone to market as a Software-as-a-Service software (SaaS) firm with several offerings to address the inventory-related management and planning challenges faced by retailers and CPG companies. Three years later, was acquired by Zebra.

It comes as no surprise that the company is very data driven – Zebra’s heritage. “Our position has always been that with cloud technology and computing, more data will be available to enable better decision making,” Lakshmanan said. “And, over the last ten years, and especially the last 2 years, that trend has evolved across the value chain. Data that wasn’t even captured before is now captured.”

Examples might include weather data that is used by wine companies to plan their shipments to avoid excessive heat, or pollen data that is being utilized by the likes of Johnson & Johnson to position ZYRTEC®.  The availability of more consumer data allows companies to make better buying decisions when they are buying 6 to 12 months out; and the availability of better short horizon data allows those same companies to be more effective as they allocate inventory across DC’s and stores, right down to the shelf level. * Editor’s note: J&J is not an customer.

Lakshmanan says this has become more critical just in the last two years. “Cyclical patterns in the consumer product market are not new,” he said. “We have always had ups and downs, but they used to take place over 3- to 5-year horizons. Business had time to respond. That started shrinking before the pandemic with omni-channel retailing with more complexity and more consumer choices.”

The pandemic, he added, was like the three weather systems that came together in The Perfect Storm to create one giant calamitous event. “We were already in a transition period and then suddenly the pattern shifted from too much inventory to no inventory and then from no demand to too much demand,” he said. “And with the volatility still in the economy, the data is still not clear. We hear from retailers that say even wealthier customers aren’t spending and other retailers that say they have strong demand.” By the time a retailer develops a strategy for the current environment, it’s shifted again. 
Chasing today’s consumer patterns is a losing game, he believes. “It’s important to take a long-term view so that the next time the pattern shifts, you’ll be ready,” he said.

The solution works by combining the historical data that supply chains have always used as well as new data becoming available, doing it at a scale perhaps not previously used, and then utilizing emerging technologies like AI and machine learning to process that data, make decisions and then learn from the execution of those decisions.

“If I’m a retailer buying from CPG companies to service hundreds of stores, I have to make inventory decisions such as what port to land, what distribution centers to send it to, how to allocate it to the stores down to the shelf level and at what price to sell it,” Lakshmanan explained. “Part of my data equation is knowing what has historically sold, at what price, what promotions I ran, how much inventory did I have and whether there were any external factors, like was it raining. Now, if I know it’s going to rain next week, I have backward and forward-looking data that I can put through an algorithm to determine things like what is the likely demand at a store in Plano, Texas.”

If a manager doesn’t like that answer, she can rerun it with a different promotion to see if the system predicts a different outcome. And, with the improvements in computing technology in the cloud, it can run those scenarios very quickly.

As with my interview with Tom Bianculli last year, I asked Lakshmanan to explain the synergies in the Zebra acquisitions. “ and Reflexis are primarily in planning, addressing inventory and labor,” he said. “Zebra has always been an execution company. We believe you have to be much more responsive today from planning through execution and now you can put them together. Insights from or Reflexis can go into a handheld device; Zebra can capture data and then send it back to our engines. And, if I can go to a CPG company and say with these solutions, we can address labor and inventory costs, you have the attention of the CEO.”

About the Author

Bob Trebilcock Bob Trebilcock, editorial director, has covered materials handling, technology, logistics and supply chain topics for nearly 30 years. In addition to Supply Chain Management Review, he is also Executive Editor of Modern Materials Handling. A graduate of Bowling Green State University, Trebilcock lives in Keene, NH. He can be reached at 603-357-0484.