How Supply Chain Leadership Is Evolving in the Age of AI and Decision Intelligence

As AI reshapes supply chains, leadership is evolving just as quickly — driven by the rise of decision intelligence. In this episode, Florian Selch, CEO and Co-Founder of Masters of Supply Chain, shares perspectives from the global network of leaders he has cultivated across industries. Today’s organizations face a new mandate: turning insights into action by evaluating technologies, identifying high-impact use cases, and building trust in AI-driven decisions. As companies scale AI, the challenge is not just automation, but rethinking how decisions are made, executed, and improved over time to drive meaningful results at scale.

www.aeratechnology.com

Download the Gartner Magic Quadrant for Decision Intelligence Platforms 

Listen to the Masters of Supply Chain Podcast here

Transcript

Fred Laluyaux 1:30
Hello, and welcome to the Decision Intelligence Podcast, where we have real, candid conversations with thought leaders, practitioners, and tech experts about AI, decision intelligence, and the future of work. I'm your host, Fred Laluyaux, co-founder and CEO of Aera Technology. Today, we are talking to Florian Selch. Did I pronounce it properly? I'm not sure.

Florian Selch 2:08
I love the French pronunciation, it's perfect!

Fred Laluyaux 2:11
Yeah, exactly. CEO and co-founder of Masters of Supply Chain, where he's built a powerful global community of leaders shaping how supply chains are evolving today. We'll focus on the impact of AI and decision intelligence on supply chains, which promises to be a very rich conversation. So, Florian, welcome, and thank you for joining the podcast.

Florian Selch 2:36
Thank you for having me. Good to see you, Fred.

Fred Laluyaux 2:38
Good to see you too. Actually, I was talking to you not that long ago. Before we start, tell us a little bit about yourself, your career. You've had a very interesting journey. You work for NASA, you did a bunch of stuff. Can you walk us through what led you to where you are today? And then we'll deep dive into the Masters of Supply Chain.

Florian Selch 2:59
Absolutely, no, thank you. And I feel like my journey or background is one of those that only makes sense after the fact, in hindsight. It doesn't make sense ahead of time when you go through it. As you just said, I previously spent six years in the science and technology world at NASA, specifically NASA Ames Research Center, which is in the middle of Silicon Valley. But one of the things, aside from being fascinated by NASA and technology and all that cool stuff, I also was always curious about the disconnect between the more technology-minded people and the more business or leadership-minded people, and you see some of that within NASA, but you kind of also see that within Silicon Valley, right, you have the nerdy, geeky computer programmers that come up with genius ideas, but then they need to figure out how to work with business leaders and how to work with investors. And so you see that, and then you see it in supply chain now as well. In supply chain we have so much new technology, and it's the technology folks that have to connect, like literally speaking and understanding each other with the business leaders, with the function leaders of supply chains, and again, that disconnect I find personally really exciting and interesting, and try to be a mediator in all of this, and so that's definitely part of my personal passion in starting this group,

Fred Laluyaux 4:26
That's awesome. So, let's talk about the group, the Masters of Supply Chain. It's a really amazing organization and community that you've created. So, I think you touched a bit already on it, but what was your vision for creating a community as an interesting evolution in your journey? So, talk about your vision, how it's evolved. I'm sure it's still evolving, as technology is changing very fast. And then maybe touch on the value that it creates for the members of that community.

Florian Selch 4:54
Absolutely, and like you said, it keeps evolving, right? There's something that's been true since we started the group five years ago, which is the challenge that a lot of supply chain leaders are facing this rapid onset of new technologies. And it really started with the onset of generative AI, which drove a lot of this new buzz. I mean, we were just at Manifest (Las Vegas), and there were 1,500 technology companies there. I'm not saying that everyone is great – in fact, many of them probably won't be there next year –but the point is, if I'm a busy practitioner, I don't have time to sit through 1,500 (companies) and these are pre-Gartner conference companies, there's so much out there. So back in the days, I jokingly say you would call up Accenture or some other consultant, and they would install SAP and Oracle for you, and you were good to go. In today's world, you have to have the supply chain leaders on the front line pick their own systems, pick their own tools that they want, and that makes it quite challenging for them to figure out how to navigate the technology environment. So we said, why don't we create a neutral, trusted, non-salesy environment for practitioners to come together and listen directly from their peers, right? Not from an analyst, not from surveys, but directly from others at other companies. What are they doing? What is working? What technologies have they tried? And then also hear from technologists and tech experts, AI experts, what do we need to pay attention to? Like, what is happening? Where's the field moving to? And again, you notice better than anybody. Six months ago, I think there probably weren't any MCP servers. Now everybody has MCP connectors, right? Most people don't even know what those are, and that's kind of where I think a lot of the education needs to come from to hear from each other.

Fred Laluyaux 6:57
So, you're curating that group, it's peer to peer education, peer to peer exchange, the true value of a community, and I think in this day and age it's absolutely critical, because things are moving so fast, people need to talk about their real experiences with the technology, with the change, and having that safe place where people can talk is really critical to their own success. So I think it's a remarkable initiative, and now you said you've been five years in the journey. So you're exposed, you're on the front line of seeing the evolution of supply chains, and especially as you talked about in the era of AI and decision intelligence and agentics. So you touched already on the fact that now those supply chain leaders, they have to almost select their solution, working with their IT, but talk a bit about the evolution of their role. What do you see? What do you hear that's different today in 2026 from when you started – when it comes to supply chain leaders, when it comes to technology?

Florian Selch 8:05
Yeah, that transition has certainly happened over the last five years, and even then it has changed, like five years ago the buzzwords and marketing language was always ‘visibility,’ right? I get end-to-end visibility for everything, and everybody is like, what does that mean, and so I think the bigger picture challenges are let's pick somebody in logistics, right? Like I'm a senior manager, I'm a director level leader in a logistics company, I'm an expert on FTL, LTL, I'm an expert on routing, I'm an expert on all these TMS, even, right? But I'm not necessarily an expert in understanding what AI can do for my logistics operation. Right, I'm not an expert in even managing my data and my systems. I probably have somebody who is doing that for me, or it used to be the IT group that used to do that, but now, first of all, I get bombarded with sales emails and I have to sift through those. I do find interesting opportunities. Maybe I'm curious about that, and so the onus now falls much more on these business leaders to make these decisions to try to figure out what could be valuable. I would argue there's probably a bit of a skill set that we need to add to to our quiver here that can not only understand what this technology can do for me, and in terms of use cases, in terms of my problems, what can it solve, but also the questions of how do I evaluate a company like that, right? Some of these companies may only be two or three years old, some of them may be 10 years old. Do I evaluate the team? Like, how can I learn from a venture capitalist? How do they evaluate companies, right? This whole thing of evaluating a technology in a company, again, it's a new skill set that was not really required when I'm managing FTL and LTL loads, you know?

Fred Laluyaux 10:03
Yeah, and we're moving from an age, as I like to say, from fear of missing out to certainty of missing out. So, the pressure is increasing dramatically, and people have to make decisions, and making those decisions and selecting the right partner in your journey is getting very, very tricky and very complex. So, that's where you provide a lot of help, but it's not just about automating processes, but it's about redefining them. It's about learning how to work with that AI capability as a partner in the way you run your supply chain, so when the folks in your community. The people that you talk to think about that shift from automation to just redefinition, co-working with that technology. Where do people struggle the most? You talked about selecting the right vendor, but beyond that, where are the big areas of struggle?

Florian Selch 11:01
Well, first of all, I think many of us are still very much at the beginning of that journey. I think individuals using ChatGPT for our own purposes are more advanced than our organizations using some of these AI tools, right? Like, most of them are really still struggling with fundamental questions like, what kind of processes can I address? And, my data is not ready to support this tool, right? We're still very much at the beginning of a lot of these challenges, so I don't think anybody actually at this point is really thinking about a new process yet, right? I think everybody is kind of stuck in that mindset of here's my problem, here's how I've done it so far - I can automate that? But I do always like to invite folks to broaden their horizon a little bit, and I think the opportunity of these tools to connect our systems together, even outside of supply chain, is hugely exciting, because so far, again, I'm a logistics leader within my logistics function. I need to understand how my decisions impact maybe planning, or procurement, or even management, right? But if I look at the bigger picture, my logistics decisions may impact product development. They may have an impact on sales or marketing or other enterprise business, right? Or cross enterprise, and I think all of that was hard to grasp in the past, but I think if I have the ability to connect my systems together, I am able to really see how my decisions affect various others, either within my company or outside, and so I think that will require the rethinking of my processes, maybe even my org structure, my teams, cross-functional teams, and so forth.

Fred Laluyaux 12:59
I think the point you made is incredibly valid, but the unlock that we're seeing right now with our clients is a little different. The shift from data driven, which means that people are making decisions with tools and collaboration platforms, to decision centric, is really happening. I think Gartner predicted that shift would be well on the way by 2028. Which means that I start with the decision, and then I look at how the decision is made – do I want to automate it? Do I want to augment it? Human in the loop, on the loop, out of the loop? What data do I need? – So you start from the decision, not from the data, which has been the problem that you've talked about, and it's been the problem in enterprise for forever, as far as I can remember. But that shift forces you to think about a decision in its context, and the context is the value chain, exactly what you said, the impact of logistics to and from. Procurement and planning is real, and we're starting to see now the fact that once decisions are getting digitized, you can now manipulate. You can now orchestrate them in a much more nimble way, because you're letting the system decide, guided by people, and that kind of forces a rethinking of how you orchestrate your supply chain within the traditional walls of supply chain, but expanding, if I'm predicting that I'm going to have some excess inventory, I need to generate dynamically a promotion or a change in pricing, and vice versa. If a campaign is working very well on social media, maybe I need more inventory to be shipped to that specific zone where I can monitor and measure where the campaign is performing, so it's quite interesting to see, and I agree with you that the market is getting there. We still have conversations about, “oh, I don't have enough data”, but we're starting to have really good conversations, and seeing the early signs of how digitizing your decisions allows you to orchestrate them in a very different manner. So, any reactions to that?

Florian Selch 15:09
No, I think what you're describing is that there are companies like Aera that do a bit better of a job of hand holding some companies, some practitioners, and their organizations showing them what is already possible. This analogy I've been thinking about lately is there's always this thing of like don't chase the shiny AI object for the sake of AI, right? Chase a business problem that you try to solve,

Fred Laluyaux 15:42
Totally!

Florian Selch 15:43
But on the flip side, there's an argument I also like to throw in, which is, if I don't understand what AI is already capable of doing, I can't think of the right use case. My metaphor for this is, imagine a logistics leader lives in a time period where there weren't airplanes yet. Their options for shipping were rail, ocean, boats, and cars – they would not think about airplanes because they didn't exist. I think that's the kind of situation we're in, where if we can get a practitioner to be open-minded enough to say, “Yes, I'm curious, I would like to learn what is possible,” to then say. “You know, let me show you, let me give you an example,” and then that sparks ideas of, “Oh, actually, I have a business problem that is very similar to this, can you address that as well?” And I think those kinds of conversations are super helpful, because then it opens up again inspiration and new ideas for new projects.

Fred Laluyaux 16:42
Yeah, and when you project a little bit and you think about the performance gap that's going to come up between those who are thinking the way you are, to scale AI as part of their operation and those who don't, what is your view on that gap? Has it started already? Do you see big differences between the haves and the haves-not when it comes to leveraging AI and automation supply chain? Or, do you think the gap will build up over time? Where do you think we are on the curve here?

Florian Selch 17:22
I feel like I'm sometimes surprised that we're still as early as we are, because I feel like the technology is not what holds us back, and the productivity gains from technology, or just from AI, is just massive. I mean, we all know that, even as an individual, right now with various AI bots here on my computer, I can get literally five times as much done as I used to before. It's just massive, right? And we do hear it, I mean, even in the last roundtable that we had on planning, right? Like there were conversations, like this one company mentioned, “Yeah, we used to have 300 planners, and now the job is done by 13 people,” right? Like, gains are real. The productivity gains are real. I would say the adoption is just slowed down by red tape and bureaucracy and mindset, maybe? Because the technology is certainly ready, it's there for the taking, and if you're not jumping on board yet, you're already falling behind.

Fred Laluyaux 18:25
Yeah, and I think to me, I observed that quite a bit. This gap has not yet been publicized so much, like the companies who are doing amazing stuff don't talk about it so much because they don't want to be caught up. I think the gap has started, and when I talk about gap, I talk about financial performance of companies, I'm talking about EBITDA, I'm talking about results. I was watching an interview with one of our customers, Estée Lauder, and their CEO has five pillars on their transformation, right, their evolution, beauty reimagined, I think that's how it's labeled, and the fifth pillar is basically around we're going to digitize decisions, so it's a strategic pillar, and they're expecting, you know, real results behind it. I believe that in the next, I don't know, six months, 12 months, 18 months, you're going to see that gap increasing. We had noticed that at the end of last month, Hershey had their Investor Day, and when it comes to supply chain, they were presenting their results and their plan, and one of the slides was around the digitization of their supply chain, and they were saying we expect, I believe it was, $50 million in savings and $100 million in reduction of inventory within the next two years. So now you're starting to see decision intelligence, AI, the digitization of the supply chain, being put forward by CEOs, either in their plan or in their analyst reports, and I think that's going to create that, you know, fear for those who are just not yet started on a journey. Any reactions to that? Have you started seeing that, is that part of the conversations with your community members?

Florian Selch 20:15
I think what is interesting is even in our group, right, like the network is around 700 people at this point, and even in our round tables, you've been joining us regularly, and it's sometimes up to 100 people joining, and even with these executives that we bring into the conversation, you do notice a difference, right? Like there are some companies that are well on their journey, and there are definitely some that are very proactive, they're implementing, they're experimenting – I think that's another thing – they allow themselves to fail, they have an ability to set up pilots, and they're okay to fail and learn from that and iterate again and try something else. Versus there are still companies out there that I've heard from that simply say no, we're not using AI at all. That it’s a policy, right? And so for me, it's interesting to observe that as a neutral party that brings everybody together. I would say we try to make it easier to take the first steps to experiment to de-risk maybe some of these opportunities that people can try. It's a learning process, right? But I personally don't have a doubt that you need to be on that journey, right? Yeah, and if somebody tells me no, there's no way we'll do any AI for whatever reason. Then I was like, well, then somebody else will take over and eat your lunch, because...

Fred Laluyaux 21:50
It reminds me of a couple of meetings I've had in the last few years with European companies that are producing very high-end luxury goods, and one of them was like, "Wow, this is phenomenal, what you're talking about, but our strategy is to go slow; our strategy is to be manual!" And they're doing incredibly well, so that's good for them, right? Another one was a Swiss watchmaker, and I was talking about stock outs, and they said, "Can you help us engineer stock outs? Because that's how we build the scarcity around our brand, and this is part of our strategy." And it was like, "No, I can't help you with that!" But, with the exception of, you know, those very prestigious brands that have a very deep understanding of how they do things, I think whenever you talk to an executive in supply chain in a boardroom, they want to have more accuracy, faster decisions. Those were the exceptions, but I want to double click on something you've said earlier, Florian, and you recently had a session on data and decision readiness in an AI agentic world. The reason I bring this up is I feel like a lot of companies want to go, but they're still stuck in this paradigm, or in this belief that their data is not good enough. You touched on that a little bit, but I'm curious if you can share a little bit of the outcome of that conversation. So, the data and decision readiness in an agentic AI world, what came out from that panel, from that conversation?

Florian Selch 23:28
I think what's really interesting for that one, it's such a foundational topic. We have a round table tomorrow on procurement, and again, guess what the number one topic is that brought everybody back from procurement automation?

Fred Laluyaux 23:33
Data!

Florian Selch 23:34
Exactly right. We set up the session specifically to try to tackle some of these myths and stories and just put them to rest and move forward. Apparently we didn't succeed fully because it's still there. What's so interesting is on the flip side, I have conversations with technology experts like yourself on a daily basis, and every one of them tells me the data issue is not an issue, like we can fix that for you, right? Like, we can come into your company, let's talk about the problem we want to solve, right? There's probably subsets of data, we will just use. We don't need to do everything end to end, and that subset of data – yes, it may be in disparate systems and tools, and it may not be clean, but we can clean that up for you. We can do that for you. And I don't know what the hold up is – is it the resistance? Is it a mindset thing? Do we not believe that they can do it right? Like, do we need to let them in and actually show us that they can do it, because that gap is still very strong, very much there. now. If we double click a little bit into it, I would say there's the data within my four walls, and then there's the outside of the four walls. I would argue within the four walls, I would like to put that discussion topic to rest, because I really think we've been there. I understand that externally it may be a little trickier with supplier partners or logistics partners or customers to get some alignment there, but still, it is fascinating to me how much it is still a blocker for folks to move forward.

Fred Laluyaux 25:32
Yeah, there's also a very pragmatic point, which is, well, I've invested x million, sometimes you know, two digit millions, in solving that problem with a data lake and a data warehouse and a data everything, a data swamp, and yet the problem is still not resolved, but I'm not really ready to say that there is a new approach. And, the reality about data –I've done this for close to 30 years – is that the technology that we have available for us today allows us to solve the problems that you're talking about. But there's a bit of a well, I'm already on the path to deploy technology XYZ, and what you're basically doing is you're moving data from point A to point B, but you're not making it necessarily ready for decisions to be made digitally. There's a layer of intelligence that needs to happen. So, the consensus from the conversation around data and decision readiness was really around data, internal, external. I think you also talked about the context layer. Do you have any thoughts on that, or any feedback on that?

Florian Selch 26:49
I mean, context was an important part of the conversation, and context is especially in the sense of AI, right? Like in some of the automation tools, super important that these AI tools have to understand the supply chain context behind some of this information,

Fred Laluyaux 27:56
The memory!...

Florian Selch 27:56
The memory, right? Like, and so that, that is certainly something that at least people start to understand and pay attention to, and I think it's an important point that can generate wrong outcomes, wrong outputs, if the context is not there. So the ability to connect systems together, I think it opens up a whole lot of new opportunities, but I think it creates or adds more of that context challenge into it. So again, it needs partners like yourself to come in and do a bit more of the hand holding there to make sure that that doesn't become an issue, that is something that's baked in from the beginning.

Fred Laluyaux 28:46
Yeah, so I'll move us forward. I've got one more topic, You talked about data, the next one is planning. When you talk to supply chain leaders, and I think the title of your round table was ‘supply chain planning is shifting from forecasting to decision orchestration’. This is right in our wheelhouse, but I love to see the round table on that topic. The reason I bring up data and planning is they’re two of the anchors in how we think of supply chain, but now you're talking about shifting to decision orchestration. So, what came out of that roundtable? Can you share some insights, and what did you think the maturity is from the leaders that you're talking with.

Florian Selch 30:28
I think what was interesting is first of all the starting point, planning forecasting is kind of at the very beginning of a supply chain, obviously. It's also an interesting starting point for using some of the automation tools, because we're starting with a scenario where most people say “I don't trust my forecast anyways.” The starting point is, whatever forecast you give me is probably wrong, and I'll probably just look at it based on my gut feeling. I changed this, and that's the plan, right? And so I think it's almost like a low risk kind of opportunity to get started and having the possibility to connect again systems, data together and then run thousands of scenarios and and get these analyzed and now get many more suggestions that may be much more accurate and better, there's still a human in the loop that can say, here's the top three options, which one do you like best, right? But that top option that now comes out of this may actually be something where the human says, “Oh yeah, that's actually really good, like that's what I would have done with my gut feeling.” I think if we reach that point of trusting these outputs, and it just comes over time, right? Like, the more and more often I get these outputs that I trust, then I can make that jump a little easier to the truly decision automation, right? Like, then I can say, okay, like, let me get that right over to my manufacturing systems and plans. That's kind of what was an interesting outcome of that conversation.

Fred Laluyaux 32:14
There was another quote from Gartner in the same report they released on decision intelligence, talking about how by 2030, decisions that have been touched by automation will be five times faster, obviously, but also 80% more trusted. And one of the trends that I'm seeing right now with our clients is – and we've seen this, we've been doing this for nine years – the shift from, “hey, okay, for augmentation, deliver recommendations, but I need my human operators to actually validate everything,” to “The goal from the get go is automation for specific processes, not for everything,” and why? Because we have enough data now to demonstrate. For example, in forecasting, every time humans change the forecast, they degrade the quality over time. And the only reason we can say that is because we create that memory, our customers create that memory. That's the point you were making earlier, about it takes time to learn, but you learn from the memory, otherwise you restart the same process every time, you don't correct your biases, every time someone else comes into the job, new biases come in. Whereas here with this approach, you memorize the decisions that you've made, and then you learn from them over time, and the system can come back and say, well, Florian, I think maybe you shouldn't do that, last time you did that, this is what happened. And those are complex, multi-dimensional problems that a human brain struggles to, you know, comprehend fully, especially as the planning cycle is shortening, right, we're getting too close to real-time adjustment sometimes, so it's becoming something quite complex.

Florian Selch 32:59
There's something that I remember when you and I talked last. Like you were saying, this idea of tribal knowledge being so ingrained in organizations, and I feel it's specifically in supply chain, right? Like, for some, you got people working in that role for 20, 30, 40 years, like we've always done it this way, like I remember this guy is the expert, he knows how to do it right. And so being able to take that tribal knowledge and be able to understand and learn from it, and let others now tap into it as well. I know that's where you guys shine, of course, but it is again, it's something so important in understanding how my decision making process actually works in my supply chain.

Fred Laluyaux 34:20
Yes, and I think we can continue this conversation for another two hours, but it's probably time to wrap up, but I take from the conversation points that don't surprise me."First of all, the power of the community, and the haves and the haves not - the gap is building. We talked about the data problem, it’s coming from a real problem to maybe a reason not to move, and you mentioned your kind of frustrated with that. And we talked about planning as well. What is the power? What is the power of the community effect? Meaning, are people joining the community to get reassured in their initial point and expectation?" Or, do you see people as a result of this great content that you put together, moving in their perspective, taking actions and establishing a direction for their company that they would have not otherwise set up?

Florian Selch 34:22
First of all, hopefully both!

Fred Laluyaux 34:22
Yeah!

Florian Selch 34:22
But the initial point, actually, entry point that I've heard several times from folks is that we sometimes provide the equivalent of a psychotherapy group session. Yeah, because, especially when I get up there into the higher level leadership position, it gets kind of lonely up there, and the fact that I can share with peers what I'm struggling with, and I hear from other peers what they're struggling with, is sort of reassuring and makes me feel a little better that I'm not the only one struggling with this. So that's one piece. Now, of course, we don't want to get stuck in the struggle, we want to move forward and help everybody towards solutions, and so there is another aspect to this community where I say we live in an age of information overload. We also live in an age where AI itself makes it easier to start an AI company. That's why we had 1,500 of them at Manifest, right?

Fred Laluyaux 35:21
Crazy!

Florian Selch 35:22
And, so it becomes more important for the group themselves to come together and say out of those 1,500, just give me some suggestions, give me ideas. Who has worked with who? Can a fellow human, a peer, giving me a recommendation, it's not my ChatGPT bot. So, it is interesting that in this age of AI, coming back to that human connection is something that is very beneficial, that people trust, and at the end of the day, I also want to say that, like we said earlier, the adoption of these technologies is not a matter of the technology not being ready, it really is a matter of the people and the processes being able to catch up, and so using that community to get everybody ready to change people's mindsets, to open them up to new opportunities and be more curious. That is really, I think, what should be the unlock going forward.

Fred Laluyaux 36:19
Now, that's awesome. Look, I'm a big fan of what you've done, and. I've seen the same way you look at plenty of software, and there are some good ones and some bad ones. There are plenty of sources for information. Yours is neutral and intelligent, and I mean that there is quality content that's being shared by quality people. It's curated, so I'm a big fan. So I invite you to check out the Masters of Supply Chain. We'll put the link at the end of the podcast, but Florian, thank you so much for sharing your insights. I know you are used to being on the other side of the dialog here, so I appreciate you playing the game with me, and good luck with everything, and good continuation for the Masters of Supply Chain.

Florian Selch 37:07
Thank you so much, Fred. Great conversation!

Fred Laluyaux 37:09
Thank you. So, this concludes our Decision Intelligence podcast: How Supply Chain Leadership Is Evolving in the Age of AI and Decision Intelligence. You can view today's podcast at our website, www.Aeratechnology.com Thank you very much. See you next time!