How GM is boosting resiliency through predictive AI tools

When discussing the roadmap for connected and predictive supply chains, GM’s Leich says that the OEM is using a variety of technologies, from supply chain mapping to real time event monitoring, AI, and something GM has been working on for several years which it calls the ‘Supplier Home Dashboard’.

“We use this machine learning tool to give us the parameters of what are the highest risk suppliers that we should be concerned about in a proactive way,” Leich says. So if suppliers that are trending negatively based on a whole bunch of interesting analytics that we use with machine learning, we can see real time risk ratings for every supplier. And that gives our team the top suppliers that we should be engaging with proactively to make sure that they can put actions in place to mitigate issues before something becomes a bigger issue.”

GM has 82 manufacturing sites around the world, excluding joint venture sites in China, with around 27,000 suppliers and over 100,000 part numbers, with a supply chain spanning across 124 countries.

“All of that equals complexity,” he says. “The name of the game is, how do you illuminate risks as early as possible, as proactively as possible, so that you have time to mitigate those.”

Leich says that these digital tools need to be used in conjunction with some clever reverse engineering of crises to help plan for future ones.

“When we would have supply chain disruptions, one thing we’d always ask ourselves is: ‘When did we know? How soon could we have known?’ so we looked at that and reverse engineered it,” he says. “We built this machine learning tool to tell us these leading indicators that are going to point to the next supply chain risk.”

From this, GM has taken the data and analytics and developed standardised workflows.

The tool recently saved GM from a potential disruption when a supplier shut down. The machine learning tool can monitor and search through billions of words and data to constantly be on the look out for potential disruptions, and Leich says it found intelligence that a GM supplier was about to close its doors abruptly. The tool, which is connected to a mapping tool, quickly identified that the supplier was within the GM network and alerted the team, enabling them to quickly look at how many tier 1s the supplier served, so the team could then contact them and communicate with them.

“Several of them did not even know about it,” he says. “So simultaneously, as we’re working with the tier 1s, we’re looking at prioritizations to mitigate risk.”

For more of the latest insights from ALSC Global 2024, follow our live blog

Now, GM wants to roll out this approach end-to-end so that it can be a more predictive supply chain using innovative tools, technology and leading indicators to help the OEM achieve that. The ‘why’ for this is partly due to the rise of EVs causing more complexity in the supply chain, he says. “We know that this is the biggest industry transformation in the past hundred years. We know it’s going to be bumpy, and that drives uncertainty from a demand standpoint,” he says. “So being able to have these connected systems allows us to really be agile. Our goal is to be the most agile supply chain because over the next five to ten years, the most agile supply chains will really be the supply chains that win.”