AI will make supply chains both lean and resilient

Subscribe

Recent seismic world events provide a stark reminder of the fragility of global supply chains. The COVID-19 pandemic resulted in major disruptions to manufacturing, processing, transport, and logistics. Extreme weather events, exacerbated by human-caused climate change, are interfering with transport infrastructure. Global conflict is disrupting food supply and oil prices around the world.

And this uncertainty is stretching traditional supply chain models to their limits.

A lean approach, where companies design their supply chains with maximum efficiency, can leave them vulnerable to sudden disruptions. So this may no longer be a tenable position.

“The world is shifting towards a place where we need to be resilient,” said Mehmet Gumus, Professor of Operations Management at McGill University, on the Delve podcast.

The traditional path to resiliency is through redundancies. But this is not only costly; it also requires a degree of reliable forecasting – a task that was difficult to do in the best of times.

“The idea of forecasting is becoming useless,” said Chris Tang in a recent Delve podcast interview, referring to the increasing unpredictability of large disruptions.

Tang is a professor emeritus in supply chain management from UCLA. He’s taught at several revered academic institutions and advised large companies like Amazon, IBM, and HP.

“If you can’t forecast, what do you do?” he said.

Traditional lean and redundant supply chain models are no longer sufficient, he explained. But another disruption might just change the game: AI.

Redundant vs. lean

To hedge against uncertainties, firms often will introduce redundancies to their supply networks. A grocery chain might source its eggs from multiple farms, in case one of them succumbs to bird flu. A laptop manufacturer might buy more memory chips than they need, in case of a shortage. A shipping company might plan multiple routes to their destination, in case one of them becomes inaccessible. These pre-emptive measures can insulate a firm from major disruptions, but they’re expensive.

That’s why many firms opt for a lean approach. Instead of keeping multiple suppliers on contract, they might do business with only one. This creates savings by 1) avoiding multiple supplier contracts and 2) giving the firm leverage to negotiate a better price. But while this saves money, it does make firms vulnerable. A notebook company can’t manufacture its product if its sole paper supplier shuts down.

Given the current global context, redundancies seem like the best way to go. But even then, decision-makers must create the right redundancies. They’ll need to forecast where disruptions are likely to occur and plan accordingly — which is becoming harder to do.

When a firm can’t lean on a single supplier and can’t hedge against forecasted disruption, what’s left?

“Resilience lean,” said Tang. “You can be lean and resilient, such that we can actually handle the uncertainty and unpredictability.”

And AI could get us there, said Tang.

Resilience lean, with AI

For a supply chain to be both resilient and lean, decision-makers will have to process lots of information quickly to make adjustments on the fly. At the moment, that’s difficult to do. But Tang believes AI models can help.

At one level, some models can handle day-to-day operational tasks like scheduling flights and writing vendor contracts. Then, at a higher level, another model can monitor these operations – along with a slew of other metrics – to quickly detect potential disruptions.

“They can integrate everything into one as a control tower,” said Tang.

If the AI detects a potential risk, it can notify supply chain managers and even make recommendations, giving them time to react accordingly.

For Tang, this doesn’t make humans irrelevant. Rather, it changes their necessary skill set.

“The human becomes an orchestrator, the conductor, to orchestrate all the AI systems,” he said.

For current students entering the field of supply chain management, Tang is designing a course to develop this new skill. He goes into detail on the McGill Delve podcast. Search “McGill Delve” on Apple Podcasts, YouTube, or in your favourite podcast player.

This article was written by Eric Dicaire.