AI-Powered Resilience in Global Supply Chains (9/23/25)

Ellie Thornton

Alright, welcome back to Milestones Behind the Freight Curtain! I'm Ellie Thornton, and, as always, I'm here with the one and only Professor Steve DeNunzio. Steve how's your week going?

Steve DeNunzio

Hey Ellie, doing well, thanks. We're finally getting rain in Columbus, and it's been one of those weeks where every supply chain feels like it's also on the brink of... a storm. Lotta moving parts lately—some of them powered by AI actually. Which fits well with what we're diving into today, right?

Ellie Thornton

Yeah, spot on. We're peeling back the curtain on how artificial intelligence, especially at giants like General Motors, is changing the game for supply chain resilience. I mean Steve did you see that piece on how GM dodged a hurricane shutdown? It had me like—woah, okay, this isn't just dashboards and buzzwords anymore.

Steve DeNunzio

Absolutely. So, GM's approach right now? It's very much "proactive, not reactive." They've built a four-pronged AI system that basically lets them look—almost in real-time—across their whole supplier map. It's mapping their tier ones, twos, all the way down the chain, constantly scanning for risks.

Ellie Thornton

Wait, can you break down those four prongs for anyone who hasn't kept up? I read it's all about mapping and risk, but—full confession—sometimes my eyes glaze over with all the acronyms.

Steve DeNunzio

Yeah, definitely. At its core, it's: supplier mapping, risk intel from live data, news scanning with machine learning, and a dashboard pulling it all together. For example, during Hurricane Helene last year, the system flagged a supplier in North Carolina—a factory cutting carpets for GM SUVs—that was in the storm's direct path. Even before the hurricane arrived, GM already knew they'd need backup, and got crews over to help the plant restore water. That meant production didn't hit a snag. Imagine pulling that off without AI—just wouldn't happen at that scale.

Ellie Thornton

It's wild. I remember this one time—totally different scale—but at a smaller UK retailer, we had a factory shutdown because someone missed—oh, I dunno, a weather warning buried in their inbox. No clever system flagged it, so thousands of kits just... sat there, waiting. It really hammered home for me the difference it makes if you've got these eyes everywhere, tracking things humans absolutely can't. I suppose that's... that's the point of all this, right?

Steve DeNunzio

Exactly. And GM's not even saying they'll replace people; it's about letting their data folks focus on real threats, not spending days sorting emails. What I find fascinating is that they're actually scaling up how many suppliers they're watching, like ten times as many since before the pandemic. This isn't just pie-in-the-sky stuff, it's—

Ellie Thornton

—oh and during those chip shortages, right? I mean, GM got whacked back then, had to stop production in, what, eight plants? That's kind of the motivation for them to go so hard on data mapping since then. It's a bit of a pivot—data management and analytics as the real secret sauce for supply chain, not just having contracts.

Steve DeNunzio

Yeah, and it's also about helping their suppliers see risks sooner than they could on their own. It’s a win-win. GM keeps its lines running; their suppliers don't get blindsided. I just wish more mid-market companies could get that level of real-time insight.

Ellie Thornton

Totally, and maybe we'll see a trickle-down as these tools mature. Anyway, let's keep this going—because AI isn’t only about hurricanes and chips, right? There’s this whole new frontier with weather data and planning that’s getting smarter by the day...

Ellie Thornton

So Steve we’ve talked about AI spotting supply disruptions before they hit—but what about unpredictable weather? I mean, my train gets delayed because of, like, “leaves on the line,” and meanwhile, there are entire ships rerouting around storms with AI-driven forecasts now.

Steve DeNunzio

Yeah, that's the big leap. Traditional weather forecasts are kind of blunt instruments; now, platforms like GraphCast or Joro—yeah, Joro's that NASA-backed one for insurers—they use AI to process dozens of climate models and real-time data sources. So instead of a single prediction, companies can see probabilistic risk across a whole region, or even for specific assets. It’s, uh, a real shift in how decision makers build resilience into their plans.

Ellie Thornton

Exactly! These models give you, like, what are the odds of, say, a port getting lashed by a cyclone in the next six weeks—or how a heatwave might spike electricity demand in California. Instead of just “here’s the forecast,” you get an entire risk profile. For supply chains, that's absolutely golden, right?

Steve DeNunzio

It really is. And in sectors like shipping or insurance, it’s not just about moving goods or insuring cargo—it’s about pre-emptively changing routes, prepping alternative transport, or adjusting insurance premiums when the probability of a big event rises. Joro, for example, can give a full temperature probability curve for any global location. Suddenly, you aren’t just hoping the weather holds—you’re actually modeling risk in ways we couldn’t before.

Ellie Thornton

And what’s interesting is, these AI models aren’t replacing traditional forecasting—they’re running side by side, right? I read this bit from the Met Office where they said, “We’ll always have classic models, but now it’s layered with AI for that hyper-local, rapid update.” It’s the blend of old and new that’s so powerful.

Steve DeNunzio

Yeah, it’s definitely an “and” not an “or.” Companies get a lot more nuance. And to your point Ellie, it takes solid data integration to make all this work—if you’re missing pieces, even the smartest AI falls over. The idea is to combine reliability from traditional models with the speed and agility of AI so that when climate volatility creeps up, leaders can actually plan, not just react. Let me throw a question back at you: if you’re a supply chain leader, how would you mix the old-school planning routines with these flashy new AI models? What’s the best way forward?

Ellie Thornton

Ooh, that’s a good one. I reckon it comes down to, like, trust and training, right? You keep what’s proven—maybe run both forecasts in parallel for a bit—and get teams confident interpreting AI-generated risks. I think it’s about getting everyone, not just the analysts, comfortable with data-driven contingency plans. Also, don’t ignore local smarts! Sometimes the person driving the lorry knows a shortcut the model hasn’t seen—so, best blend is tech plus human instinct.

Steve DeNunzio

Perfectly put. And as we’ve seen with everything from droughts to wildfires to random “once-in-a-century” storms happening every few months, the only real way forward is agility. You layer up your insights—from old and new systems—and move from what just happened to what could happen and what are we going to do about it.

Ellie Thornton

I love that framing. Alright, so, weather and data are huge, but there’s another, even trickier area making headlines lately—cyber disruptions. And not just for banks or shops, but end-to-end logistics networks...

Ellie Thornton

Okay, so Steve—this past month, aviation just got clobbered by a ransomware hit, right? It wasn’t even a direct airline attack—just a vendor went down, and suddenly airports from Heathrow to Berlin were at a standstill. Makes you realise supply chain resilience is about way more than what you can see moving on a ship or truck.

Steve DeNunzio

Absolutely. It’s, uh, almost a textbook example—one exploited vendor, and the whole continent hit pause. The scary bit is, even airports with all their classic backup plans weren’t ready for one software provider taking a hit. As we just discussed, you can’t silo physical and digital supply chains anymore—they’re intertwined. If your vendor’s IT isn’t rock solid, you could have all the weather models in the world and still be offline.

Ellie Thornton

I know, right? And it’s not just airports; fashion brands are feeling it too. They're using AI now for “what-if” planning—like, what if tariffs spike, or what if factory B is compromised? But all those clever tools still need, like, really robust, connected data. Otherwise, as we saw this summer, companies just scramble with knee-jerk reactions, which sometimes does more harm than the threat itself.

Steve DeNunzio

And we’ve seen this back in the WannaCry crisis. Vendor ecosystems at logistics providers can be a maze. Every missed software patch was a potential domino. When the ransomware hit, no one really knew which systems were linked to which partners. That confusion caused way more downtime than the virus itself. It’s—uh—a wake-up call. You have to map your digital connections as thoroughly as your physical shipping lanes.

Ellie Thornton

Exactly. If you don't have collaborative risk management, if you're not stress-testing your plans for things like data breaches or ransomware attacks, you're basically hoping for the best. Regulators are stepping up, but still, it takes a village—a whole network—making sure every vendor can prove, “Yeah, we've got security sorted.”

Steve DeNunzio

What it comes down to is visibility—real, validated, connected data across the board. If you only track what’s in your own warehouse, you’re missing blind spots. If everyone in the network is stress-testing, sharing threat updates, and treating digital resilience like the real-world supply chain, you’re miles ahead. It’s not about compliance for the sake of it, it’s about building trust and, honestly, making sure that when the next crisis rolls around, you’re not the weak link.

Ellie Thornton

Hear, hear. Alright Steve, shall we call it there for today? I feel like we’ve zipped from AI hurricanes to cyber meltdowns and barely scratched the surface! Plenty more to cover next time—but for now, cheers for joining me and sharing your stories.

Steve DeNunzio

Always a pleasure Ellie. Thanks to everyone listening, and if you’ve got a hot take or disaster story from the trenches, send it our way. We’ll be back next week—until then, stay resilient!

Ellie Thornton

And remember, you can always email us at milestones at Steve DeNunzio dot com. See you next time! Bye Steve. Bye everyone!