How Mars, Church & Dwight, PDC Brands, Lowe's Build Intelligent Supply Chains

When it comes to supply chains, volatility comes with the territory. From unpredictable consumer demands to global disruptions, a linear supply chain is no longer a realistic or efficient model. 

AI can help retail and consumer goods companies make decisions in real time, pinpoint growth opportunities and pain points, and speed up systems performance, but to what extent?

At the recent Analytics Unite event, Janice Burk, VP of technology supply chain for Lowe's, Alexander Cunningham, director of advanced analytics for Church and Dwight Co., Kristen Daihes, global VP of supply chain for Mars, and Sulabh Jain, head of supply chain for PDC Brands, spoke about how they implemented AI within their own supply chains, and what they learned along the way.

Endless Opportunities, With Guardrails in Place

According to Cunningham, supply chains have been thinking for themselves for quite some time.

“Operations research won us World War II, but that research was theoretical and we didn’t have the compute abilities to execute those theories at scale. Now we have the solutions that unlock these problems,” he said.

He believes an agentic layer can access output and take action. “But how do we take those insights and put them into play so they don’t just become interesting trivia, but drive the organization forward?” he posed. 

Also: Church & Dwight looks to AI to capture consumer attention in the vast online landscape

Burk agreed the space is exciting, “but we’re not there yet,” she said. In terms of progression, “many of us are at that point where we've started to build out the data platforms that are going to enable a lot of the autonomous and agentic work to be performed, but on top of that is the consideration of how agents will interact with that platform.” 

Use cases for agents can include demand forecasting and taking on repetitive tasks that can be set within guardrails and adjusted as needed. 

“There’s a lot of activity we’re doing from the standpoint of ensuring that we're working through processes, first with a human in the loop and then humans managing by exception," said Burk. 

Daihes also believes in creating the right foundation, but it must drive automation, and Jain said it’s less about the buzzword but more about closing the loop. 

“It’s about something we see, notify and recommend. If an agent can see delays in production planning or shipments, it can notify the right person as a phase-one step,” he said.

Jain sees agents eventually being able to help drive the top line, manage investments, improve working capital and create “a much more collaborative supply chain.” 

And according to Burk, the best way to balance AI-human collaboration is to “take your teams along with you,” especially if they are legacy teams, for which “some level of explainability is required.”

Adding AI to supply chains can also help teams respond to disruption. Cunningham credited the technology for helping them work through pricing hurdles and workflow, and Daihes said it was a heavy lift to figure out, but it allowed more end-to-end visibility. 

“Once we've got that connection, there's a much easier way for us to start connecting the dots between the data to derive important insights,” she said.