In the business landscape we’re navigating today, few components impact the bottom line quite like supply chain management. Perhaps more importantly, few elements of a business affect the overall experience for human end-users like a supply chain operating at maximum efficiency.
Rapidly advancing technologies like artificial intelligence are revolutionizing how factories strategize, how goods move, and how high the efficiency bar can be raised. But for Emily LeVasseur, the self-described “supply chain geek” and Co-Founder and Managing Director of Waypost Advisors, it’s the synthesis of human expertise and institutional knowledge with cutting-edge technology that will lead us into the future. From the reality of AI-driven decision-making to the dawn of digital twins, LeVasseur details how manufacturers are leveraging data to transform the way companies handle supply chain management. The following is a condensed version of our conversation edited for length and clarity.
Design World:
Hi Emily! Thank you for being here. You previously mentioned the statistic that, according to the Data Manufacturing Institute, 97% of manufacturers are concerned about losing institutional knowledge as their workforce retires. The question is: how are manufacturing companies using artificial intelligence to capture this silent knowledge now?
Emily LeVasseur:
Yeah, great question. I can’t verify the 97% — I’m not aware of that — but what I can tell you is that the companies we work with, which are middle-market manufacturers across many different industries, are very concerned about it. In fact, I was just having a discussion with a data modeler about a project we have for a company whose production scheduling decisions are highly sensitive. They risk a lot of changeover loss and downtime if they don’t optimally schedule their production facility — and all those decisions go through one person who’s been there 25 years.
So this is a big concern. That person is starting to think about retirement, or even just how to delegate and democratize decisions. What we’re starting to see is companies asking: how do we use data, information, and AI to help democratize those decisions while keeping them as high-quality as possible?
Depending on your definition of AI — since we work primarily in supply chain, though that’s inextricable from production and manufacturing — AI means a few things. I think a lot of people think of ChatGPT or large language models, but in supply chain it’s really about data analytics and machine learning.
Large data models can bring in far more data than a human could process in a reasonable time, and they help structure decision-making so that outcomes can be presented as scenarios to a human decision-maker. That’s where AI really shines.
We’re also seeing things like digital twins — which I think I saw an article about in Design World, right?
DW:
Yeah, it’s a big part of what DW covers from the factory automation standpoint. It’s always on everyone’s mind.
EL:
Exactly. The digital twin factory is really helpful because it generates so much data around the supply side. You can also bring in data from the demand side — orders, forecasts, macroeconomic trends — and combine it all to optimize inventory: having enough to meet demand, but not so much that you’re sitting on excess or obsolete stock.
DW:
That’s such a good point. We frequently find ourselves at shows — IMTS comes to mind — and hear people say, ‘We’ve collected mountains of data, but if we can’t analyze it and make it actionable, what’s the point?’ and, ‘Hiring a data analytics team is expensive and time-consuming.’ Especially with digital twins, I have this conversation constantly. It’s interesting to see where AI might bring solutions for people.
EL:
I absolutely think it will. Data is the unsexy foundation of all these powerful tools. I was talking yesterday to a woman who’s a data governance director at a large manufacturer in the Twin Cities, and she said it’s getting huge — because data is the fuel for the AI engine.
DW:
Right.
EL:
And one thing I’ll say — I’m super passionate about the supply chain planning function: demand forecasting, inventory strategy, capacity modeling. It’s the mathy, sciencey side of supply chain. But those roles are probably the first ones that’ll be replaced by AI once companies figure out their data, because so much of it is math.
DW:
Right. People avoid that topic, but it’s true. AI will integrate with many jobs and help people do them better, but we’re lying to ourselves if we think a few roles won’t disappear altogether. Once it’s cost-effective and easy to implement, those data-heavy jobs might go first.
EL:
Yes. In a previous career, we had a team of 20 people doing supply and inventory planning. So much of that was just “plug and chug,” exporting Excel spreadsheets and doing analysis. You won’t need 20 people to do that once the AI systems are set up — you’ll need maybe two.
DW:
Exactly — one or two people who know how to use and maintain the technology.
EL:
Correct. And back to your question about losing institutional knowledge — the big thing is understanding how we make decisions, getting that tribal knowledge out of people’s heads, and democratizing decisions using tools.
DW:
Yes, and that leads into my next question: what ways can AI capture and learn from essential institutional knowledge?
EL:
I really think it’s through data analysis — especially if you can build a machine that continuously processes data and learns from outcomes. It takes time and investment to “teach” the system, but once you do, it can start making higher-quality decisions and reveal insights that humans might not have had the bandwidth to discover.
DW:
Could you give us some real-world examples of manufacturing clients using AI?
EL:
Sure. This wasn’t one of our clients, but a company someone in my network worked with used AI to track and trend absentee rates at their facilities. They correlated it with weather, local events, holidays, and more. Eventually, they could forecast absentee rates to 97% accuracy — which helped them plan temp labor better.
DW:
That’s incredible.
EL:
Yeah — it minimized throughput constraints and improved planning. Another example: we worked with a company that was building a demand-sensing model to improve inventory management. They had a lot of “roving inventory”— stock in trucks or distributed across regions — and they wanted to use external data like weather and economic trends to predict where demand would rise. That helped them reposition inventory efficiently instead of stockpiling everywhere.
DW:
So as a self-proclaimed “supply chain geek” (which we love), what do you think is the next big challenge for supply chains — and can AI help?
EL:
That’s a big question. There are so many challenges. Climate change comes to mind — it’s going to impact infrastructure: ports, rivers, the Panama Canal, transportation routes. Physical movement of goods in a changing climate, plus labor shortage — that’s huge.
Also, demographics — in the U.S. and China, populations are aging, and labor forces are shrinking. Even if automation improves, the redistribution of workers doesn’t necessarily fill demand where it exists.
So, yeah. I think the physical movement of goods and infrastructure challenges will remain major concerns, especially with climate change and global instability.
DW:
You’re absolutely right — climate change is already here for many, and we don’t have a clear path forward yet.
EL:
Until we learn how to teleport food and goods, that’s going to stay a challenge. But I’ll put a positive spin on it: human ingenuity is amazing. We’re adaptable. Whether it’s climate change, labor shortages, or supply chain disruptions, we always find ways to adapt.
Yes, it’s going to cost more, and yes, it’ll cause short-term pain — but it’s also a breeding ground for innovation.
Design World
Filed Under: Commentaries • insights • Technical thinking, NEWS • PROFILES • EDITORIALS, AUTOMATION