00:00:00 What makes someone good at supply chain
00:00:42 Career Companion and gaps in soft skills
00:02:18 Theory vs. practice in supply chain education
00:06:29 Communication mistakes and reframing messages
00:09:55 Why formal education underdelivers in supply chain
00:15:48 LLMS, writing, and thinking in modern education
00:20:30 LLMs as research tools vs. shallow prompts
00:24:58 Teaching leadership through frustrating exercises
00:31:40 The Excel debate and importance of sense-checking
00:36:50 Must-have vs. nice-to-have supply-chain tools
00:42:40 Agentic AI and future digital fluency
00:45:55 AI-driven job disruption in analytical roles
00:48:40 From mechanization to collaboration with partners
00:52:00 Final thoughts: high-level thinking vs. automation
00:53:05 Lifelong curiosity and value of soft skills
Summary
In a dialogue hosted by Conor Doherty, Philip Auinger and Joannes Vermorel explore what makes a great supply chain practitioner. Philip, leveraging his supply chain experience, highlights the importance of bridging analytical prowess with interpersonal skills, a gap often seen in the industry. Founded in 2019, Philip’s company, Career Companion, addresses this shortfall by offering interactive workshops to foster practical application of theories for corporate and academic audiences. Joannes critiques outdated academic models, underscoring the necessity of effective communication. They discuss AI’s impact on education, asserting that core critical thinking remains vital amidst technological advancements. Both emphasize embracing interpersonal interactions in future supply chain roles.
Extended Summary
In the interview hosted by Conor Doherty of LokadTV, a noteworthy exchange unfolds among Philip Auinger, a champion of personal development within supply chain management, and Joannes Vermorel, CEO of Lokad, around the theme “What Really Makes a Great Supply Chain Practitioner.” The discussion delves into the integration of analytical prowess and interpersonal efficacy, with Philip emphasizing a gap often observed in the industry where quantitative skills overshadow people skills, precipitating crucial relational challenges in supply chain roles.
Philip Auinger retraces his career trajectory—counting screws as an intern to leading demand planning teams—highlighting his journey from grappling with supply chain complexities to founding Career Companion in 2019. His narrative casts an illuminating light on the inefficacies of traditional education systems, which, according to Conor, offer inadequate preparation for young professionals. Philip advocates for bridging theoretical learning with practical application, a tenet reflected in the workshops he conducts, tailored for both corporate entities and academic institutions aiming to enhance soft skills among their ranks.
Joannes Vermorel weighs in on the discourse surrounding hard skills, reiterating his skepticism towards outdated academic theories. He posits that effective articulation through writing is central to managing trade-offs, presenting a critical perspective on the pedagogical focus of universities. Philip underscores this through a personal anecdote, illustrating the missteps in communication within demand planning scenarios. These exchanges emphasize the indispensable role of precise communication in aligning stakeholders and advancing supply chain strategies.
The dialogue takes a reflective turn as both Philip and Joannes explore the limitations of formal training, contrasting supply chain education with more rigorously defined fields like engineering and surgery. Philip expands on insights from LinkedIn research—highlighting diverse educational backgrounds among successful practitioners—and critiques certifications for their disconnected relevance and cost. Joannes parallels Philip’s critique, focusing on structural constraints of academic grading that fail to account for essential skills like nuanced problem-solving and complex trade-off presentations.
Amid the conversation, the emergence of AI tools like ChatGPT surfaces, sparking considerations on their implications for education. Philip and Joannes each offer perspectives on mechanizing aspects of writing through AI, while maintaining that core critical thinking and experiential learning remain vital. Philip’s approach juxtaposes AI use with active student engagement during group discussions, fostering genuine assimilation of skills.
Conor guides the conversation towards an exploration of practical competencies, where Philip and Joannes dissect the relevance of traditional tools like Excel and programming languages such as Python. Philip envisions Excel transitioning from a ‘must-have’ to a ’nice-to-have’ as technological advancements make interfaces more user-friendly. Joannes, however, emphasizes the importance of mastering any programming language for its conceptual mindset.
The impact of AI on future supply chain roles forms another crucial segment. Both Philip and Joannes forecast major automation within analytical roles, underscoring the need for adaptability and the fostering of interpersonal skills to thrive in increasingly automated environments. Strategies to future-proof supply chain roles include embracing front-office interactions, thereby mechanizing internal processes, as proposed by Joannes, while Philip underscores the relevance of empathetic and adaptive professional mindsets.
As the interview draws to a close, Joannes and Philip contemplate enduring skills, with Joannes advocating for elevated thinking aligned with managerial challenges, and Philip fortifying the pertinence of relational capabilities—asserting human connection as an irreplaceable asset amidst burgeoning technological change. Conor concludes, recognizing the valuable contributions of both guests and inviting them and the audience back to their pursuits.
Full Transcript
Conor Doherty: Welcome back to LokadTV. Joannes and I are joined by Philip Auinger. He’s the founder of Career Companion and today he joins us to share his insight on a very important question, one that sits right at the crossroads between personal development and hardcore analytics, and that question is: what exactly makes someone good at supply chain? Now, before we get started, you know the drill: subscribe to the YouTube channel and follow us on LinkedIn. And with that promotion out of the way, I give you today’s conversation with Philip Auinger.
Well, Philip, thank you very much for joining us.
Philip Auinger: Big pleasure, thanks for having me.
Conor Doherty: It’s been a while in the making, but before we get into the conversation proper between you and Joannes, could you please introduce yourself to the audience and explain what it is that Career Companion does?
Philip Auinger: Sure, so my name is Philip Auinger. I worked in supply chain for something like eight years, out of which four years as a team leader for a regional team. At one point in my career, I realized that supply chain isn’t only about numbers; it’s really also about people. Then, I discovered that many people in supply chain are great with numbers but not that great with people, and that’s where I definitely saw for myself, “Hey, maybe this is a sweet spot for me, maybe this is something that I can dive into.”
That’s back in 2019 when I founded my own company, Career Companion, which focuses on communication skills for people who are specifically working in supply chain. I love working mostly with young people because they used to have all of their careers ahead of them, and it’s kind of me thinking, “Boy, I wish I had known this when I started out.” That whole motivation is always in the back of my head when I work for the supply chain community, especially on LinkedIn also.
Conor Doherty: Well, actually, that’s how I became aware of you in the first place. I saw some of your posts. In fact, just yesterday, I was responding to some of the things you posted on LinkedIn, completely unrelated to setting rapport for today, I assure you. But you mentioned that you like to work with a lot of young professionals. So what is it that you think young professionals require in terms of mindset, in terms of soft skills that are not being provided adequately?
Philip Auinger: I really think you need to distinguish between are these young people who are learning theory at a university, or are they at some kind of an academy or university of applied science where they learn the real-life supply chain, because there’s a world between it. Just recently, I was working with one company, and they asked me to review their materials, and it’s pure theory.
I said, “Yes, it could be that I don’t know if you want to calculate safety stock, this is the correct formula, but if you have a sales rep yelling at you that you have to have everything on stock, you won’t convince them by quoting the formula.” And I think that’s the key point really for young people also to understand that there’s theory and there’s practice.
The best way you can start off into your career is understanding that both are important and kind of drawing the connections between the two, because many people who work in supply chain don’t have the theoretical knowledge but kind of learned it on the job.
But if you, especially as young students, learn the theory and then during your studies you already work with companies, you already have real-life data, real-life case studies, that’s where you can make that connection, and that’s where young people definitely have a starting advantage.
Conor Doherty: Well, thank you, Philip. Joannes, you’ve been waiting again. I know that in the past you’ve described some skills as being sort of fluffy; you want to be more focused on the hard skills. So I’m just curious, when you hear what Philip has to say, what are your thoughts?
Joannes Vermorel: I mean, yes, both hard skills and soft skills are, I believe, critical. But I would say what passes as theory in most universities for supply chain is fairly outdated and fairly useless. So that’s, I would say, the safety stock formulas, for example, are great as nice mathematical puzzles because they have analytical expressions. You use normal distribution, so you can actually write it down, you can actually do a calculation.
Another example would be the EOQ, economic order quantity, which can be framed as a second-degree polynomial which has a nice analytical solution, so you can literally write it down and you will have a square root in the middle. Oh yeah, great. So I think most of it is things that are first. I mean, it is technical, but it is also a little bit trivial rather than deep technical knowledge. It is relatively shallow; it is not things that are very, very useful.
And then on the other hand, when it comes to soft skills, I would say it’s mostly absent. For me, what I see as one of the areas that is most lacking is proper writing skills. The supply chains are complex; they are complex beasts. For example, just framing what are the trade-offs that we’re trying to address because it’s critically important. If you don’t have that, then yes, people will be complaining. The CFO would say, “Oh, that’s way too much working capital.” Sales would say, “Oh, it’s way too many stockouts,” and then other people, etc., etc.
So, but the reality is that you only have trade-offs. Before stepping into the technicalities of the trade-off, you should be, as a supply chain practitioner, able to convey that in a meaningful way, ideally in writing. I mean, it’s best if you can then support those writings with probably with meetings, as many meetings it takes. But here when we are touching this sort of soft skills, which is highly structured, very high-quality communication, this is completely absent.
Completely absent. That would be my take of the sort of most glaring gap in the education of young professionals for supply chain jobs.
Philip Auinger: An example of exactly what you said, of how you write it—this was stupid young Philip who did this. So I used to collect forecasts for very important products from sales. Obviously, if you’re in demand planning, you don’t want to overdo this, because if you ask too much, you won’t get any answers anymore.
So what I did was I used to send out monthly emails to ask for forecasts, and then I thought it would motivate them to say, “Hey, if you had been more precise with this one estimation we would have been at 80% forecast accuracy.” So what does a sales rep care about my forecast accuracy? Of course, it was important for me, but now looking back, that was a stupid way of phrasing it.
If I say, “Hey, thanks for your inputs because based on this we were able to purchase the raw materials and we were able to deliver on time and it was better now than it was before when you didn’t give us any forecasts.” That’s the way to sell it to sales. If you kind of always have in the back of your head what does this mean for other departments, what does it mean for that one person that I’m trying to convince, that’s where you have that magic wand where you can get a project off the ground which normally would just be, “Yeah, it’s a nice to have, but who cares if you do it or not.”
But if they see the value in it and they trust you with it, then you can really make a difference because suddenly you have people working with you and not against you.
Conor Doherty: Because it occurs to me that quite literally everyone involved in this conversation is a professor of something. So we’re all in third-level education. As peers, may I ask the following question: We’ve set up the topic of education and skills and what students are learning, and I don’t want to focus purely on, okay, what are they lacking. Let’s be charitable at first. You talked about young people entering education, so whether they’re at trade schools, whether they’re at universities, what are the things in terms of the theory that they are getting right? So what’s the valid theory that they are getting, and then what is some theory that maybe they’re missing?
Philip Auinger: Go back one step because now we’re speaking about people who get university education or whatever higher education to work in supply chain. Many people don’t. So I did this research once, I asked around in my LinkedIn network, I asked, “Okay, how many of you actually studied supply chain?” And it was less than half, something like 55% actually come from a different field. The most extreme I ever encountered was marine biology, and they worked in supply chain. They were excellent supply planners.
So maybe to put that comment first: You don’t have to study supply chain to be able to work in supply chain. If you do study supply chain, or if you do study that and if you do want to work in the field, then what I see very much in the students I’m working with is very strong analytical skills. It’s like that Sherlock Holmes kind of approach to things. You find a problem and you want to dig down, you want to find a solution. Maybe it’s asking five times the question why to really get to a root cause, but it’s that skill that I see with young people. It’s that skill that ideally you keep your entire career because that’s what enables you to not make the same mistakes again. I think that’s a critical way of thinking that’s valuable for students from day one if they want to work in supply chain.
Joannes Vermorel: So I mean, clearly yes, I’m even surprised that in your survey you had 50, you had 45 percent of people who have formal, you know, supply chain background.
My experience would have been, if I had to guess, that was just a perception, if I had to guess a number, I did not do any survey, I would have just said one-third, but you know, so even lower than that.
For me, it is a testament that most of what passes for supply chain theory is broken, deeply broken. The reason is that if you look at, for example, people with and without formal training in, let’s say, mechanical engineering, we are not even in the same realm. I mean, the sort of capabilities that you have, capacity to execute work, do things, are just orders of magnitude better if you are, you know, trained.
Same thing for surgery, at no point would you say, “Oh you know what, half of our surgeons have zero training in surgery but we’re doing just fine.” I mean, if it was that, that would mean that what passes as training is just nothing, it’s just nothing.
And here what you’re describing, and I very much agree with that, those people in fact who succeed without any training, what they have are skills that carry over, analytical skills, I would say capacity to be very diligent, organized in their work schedule, etc.
So you have plenty of things that carry over and that’s very nice, but again, that is something that reflects really poorly on what passes in academia and in many professional training institutions on supply chain.
Because again, you would expect people with formal training to be, if we had true, genuine, very efficient theories, to be an order of magnitude better. I mean, again just think of how many people are self-taught playing violin and being excellent at it, it’s like almost nobody.
Even for basic sports like playing football and whatnot, again, the people who have undergone specific coaching programs and mentoring are just not even in the same ladder than people who don’t.
So that’s, I would say, my perception of this field and this percentage that you’ve surveyed, which is very interesting for me, kind of confirms this intuition.
Conor Doherty: Philip, if I can just come back on that, so when you said that, I think it was you said 55% so one out of two approximately of your audience had formal supply chain training? Or did you mean that they had studied that at university or they had taken courses, trade training, so for example, APICS, Six Sigma, things like that, you know, like certifications like trade school?
Or was that formal university?
Philip Auinger: It’s a while ago, I think I had worded it by saying, do you have formal education, for example a university degree, I think that’s how I put it.
Because obviously, having a diploma from a university and really working on this for two years, three years, five years, that’s something else than obtaining a certificate for example.
Conor Doherty: That leads to the next question because, again, it’s entirely possible that someone might say, well, does one really need, actually it’s self-evidently the case that one might not necessarily need to have formal tertiary level education to excel in supply chain.
One could take shorter form courses like Six Sigma, like APICS, and again then the question becomes, is that sufficient in your opinion?
Philip Auinger: Let me rephrase the question again: not only is it sufficient, is it necessary?
Many people contradict me when I say this, but I’m not a big fan of certificates. I have one, five levels, very annoying, lots of work to pass that exam. I remember like when I was at school the pass mark was 50%, and there the pass mark was 93%.
If you had 92.9, I had that once, I missed it by one point, then you failed. So it was very, very strict but the learnings honestly weren’t all that great.
Especially, it was, I won’t say the name now, but it was a big American institute that imposed its way of seeing the world on the rest of the world.
If you’re in Europe, you realize many of these things just don’t make sense because it’s not the same market, it’s not the same way we transport goods, for example.
We had to learn things that we knew were wrong for our own business, but we had to learn them this way to pass the exam, and that’s why I think sometimes they’re overrated.
At the same time, they come with a hefty price tag. If I had paid for that myself, I would have been very angry about the value I get for money.
If it’s paid by a company, we were something like 10, 15, 20 people, I mean that costs dozens of thousands of euros. That’s a big number to then say that most people afterwards thought, “Yeah I learned this, I passed the exam but I don’t feel it was that relevant for my job.”
If you have no experience whatsoever, that’s a good way kind of to catch up the missing education that you had before, if that’s the right path for you. It could be right.
But if you already studied that, if you already have years of experience on the field, you’re just kind of learning the basics and the basics based on your experience aren’t even correct anymore.
As you pointed out before Joannes, sometimes it’s just very theoretical and even the theory is outdated.
So again, many people contradict me on that.
Conor Doherty: Well Joannes, are you one of those people who will contradict Philip on this take?
Joannes Vermorel: No, no, no, I you know I’ve been teaching at university for seven years and I came to the realization that having to grade students is a massive constraint on what you can teach and how you teach it. I had immense privilege so I could do pretty much whatever I wanted. I had an administrative staff that was very lenient on me, so they didn’t really care if the way I would grade everybody was canonical or not.
But the reality is that if you want to do it by the book then you end up with things that exactly what you point out. Which is if you want to have your grading system, you will focus on some sort of checking whether people can absorb trivia as opposed to skills that are much more interesting but so much more elusive.
For example, I was discussing about capacity to write, to present a very complex muddy trade-off into something that creates clarity for non-specialist parties. This is difficult. How do you assess such a skill? As a professor, you can ask your student to produce an essay or something, but it takes so much time to review and grade that we’re not even in the same category. It’s the sort of things where it takes one hour for the student to produce the essay and it takes one hour for you, the professor, to grade the essay. It is brutal.
Philip Auinger: Do you really think they’re still writing it themselves these days?
Conor Doherty: That’s what I was about to ask.
Joannes Vermorel: Obviously, that’s the interesting thing. ChatGPT is a fantastic tool and those LLMs are fantastic tools. I’m using it and I’m using that also but the interesting thing is that if you’re crap at thinking, you do not think clearly, you will be prompting nonsense and ChatGPT will happily play along with you. Yes, you will get a text that is superficially very consistent, well-written and what not because thank you ChatGPT, but it might be completely off your target.
So again, for me, the interesting thing is that the availability of LLMs puts even more pressure on those skills in a sense. Because then you need to be able to very swiftly assess whether your LLM has produced something that was aligned with your intent. You can skip the part that was very tedious and slow, which was just writing it down painfully word by word, this thing is mechanized by ChatGPT. Now what remains is the thinking, and I would say bad news for many students. Thinking might not be the part where they truly excel, you know that. So again, it’s kind of a brutal and humbling experience to use those tools because you are confronted to your own limitations suddenly. You realize that the limit is not your capacity to write because the LLM is taking care of that, it’s your capacity to think.
Philip Auinger: May I add here, of course, I think that many people then have the bias to think that if they just scan through what ChatGPT wrote, they understand it and they grasp it and it’s deeply rooted in their mind. That’s not how our brains work. So if you just let ChatGPT do this and I ask you one month later, “What did ChatGPT write for you?” Completely gone. If you wrote that on a computer, you might remember parts of it. If you wrote that with your own hand, you’ll probably remember most of it. And that’s a given and that won’t change just because we have more technology around.
Around and that’s why for me, just I’m doing a leadership course now for my old alma mater, the university where I studied. And of course, there is also a written part which is basically three reflection questions. But I say if you want to write this with ChatGPT, be my guest. I won’t be a policeman to check if you did it or not. But we will have a call afterwards in a group which will be the final exam, and I will notice very quickly if you really thought into this or if you just let an LLM do this.
At the end of the day, even if you cheat through that or if you’re good at passing that oral exam, great for you. But did you really learn something from this or did you just learn to work around the system and get through this one course with as little effort as possible? And I keep telling them the more you really think about this and the more I can inspire them to use their own brain cells to come up with solutions or even come up with questions where they say, “I can’t solve this. Hey professor, how do you see this?” Then they learned a lot more than having ChatGPT write, I don’t know, five pages about a subject. That’s just not the way you learn.
Joannes Vermorel: My take is that, you know, those tools are fantastic. I’m using them all day long, and I think that’s where I’m not teaching right now at university like I used to a full master course. But the way I think it creates, it asks for deep questions about how do you want to even modify the teaching, taking that into account? And for example, I’ve been reading a programming book very recently that was very nice.
It’s, I think the title is like “Python for Thinking Like a Computer Scientist.” It’s a very short book, and what I found very interesting that the author literally was saying, “Oh if you have any further questions” as a footnote on this thing, “just take your favorite virtual assistant and probe the assistant on those keywords.” But the interesting thing was it was a way to keep the discussion very short while giving hooks on how you can expand, and it was carefully executed. It was well done.
So, you see the way I’m thinking of it is that and the way I’m using those LLM tools is that when I tackle a subject and that I’m writing it down, I will constantly probe the tools on stuff that I don’t really know, things that are at the limit of my understanding, limit of my knowledge. “Okay, tell me more about that.” And if I’m suspicious that the tool is just giving me bogus information, I can double-check. But really, usually the problem is just sheer ignorance of mine on fields I know nothing about.
So, it’s not I’m not pushing the limits of the tool onto very fancy tricky reasoning. It’s just that my specialty is supply chain software, and if you ask me very specific questions about, let’s say, shipbuilding, the various stages and whatnot, I’m not a complete expert on that, so very ignorant. And an LLM can really very quickly help you to fill those gaps.
Conor Doherty: Well Philip, if I can just jump in because and I don’t want to weigh the conversation down in pedagogy too much, but you were drawing a distinction earlier between like formative and summative education. And again, I would agree with you that summative assessment, the idea of like just write an essay, most people are just going to have ChatGPT do that. The more formative aspect of assessment like here are a series of small assessments, I want you to do group work, problem solving in real time around me. That’s how I execute my course also, a master’s course. So I’m just curious, when you’re teaching, how do you gauge that people are learning the skills that you think they need to possess?
Philip Auinger: Obviously, I can’t look in everybody’s minds. I can’t even look into one person’s mind. If you have a group, I mean I’m lucky it’s a group of maybe it’s 18 people, 20 people, so it’s a fairly small group. So it’s fairly easy to have a look in the audience and see is there like skeptical eyebrows, or are they all okay? Big difference; you can read very much from eyebrows, just by the way.
But what I did at one point was I thought we were talking about lateral leadership, so if nobody reports to you but for this one project you’re the project leader. That’s very tricky. So what can I do now to teach them how tricky it is? I can create a beautiful PowerPoint and write five points and say, “You need to watch out if you ever are a lateral leader.” Or I ask them, “Hey, work out what do you think are the critical paths if you’re a lateral leader for this project.” No, what I did instead was an experiment.
Of course, you have to moderate, you have to facilitate when you walk in saying, “This is an experiment, you will hate me now, but let’s try this.” So I basically gave them very rudimentary, very bad instructions of what to do and gave them half an hour, and then I shut up. I did not say a word, and they were getting frustrated; they didn’t know what to do. And then in between, as I gave them some cues to make it even worse, and this is stuff like, “One of your team members now goes to the next group, and when they arrive there, they’re opposing whatever has been done so far,” or one was, “One of your team members now suddenly goes quiet or says, ‘This is nonsense, I don’t get it, why are we doing this?’”
At the last few minutes, I walked around and kind of criticized everything, “That’s what you’re writing on? Seriously?” Like really putting bad energy into that. And then after half an hour, I ended that experiment, and then I collected, “Okay, what did you learn and what did you notice in this?” And all of the points they brought up here would have been on my PowerPoint, but they came up with them. And not just because they asked ChatGPT or because they discussed it in theory but they experienced it. They noticed how important it is to have clear leadership, a clear goal, structured goal, that you don’t have people jumping back and forth and everybody’s involved, and so on.
So they learned this, and I told them, I mean this is a prophecy, I can’t prove it yet, but I told them the first time you’re in this position of having lateral leadership in a project, you will think back to exactly this moment. So I hope that I made an impact here to let people grasp something by experiencing it and not just by reading it or writing it or talking about it, but by experiencing it. And that’s where if you look at how you learn and how you teach, that’s obviously state-of-the-art. If you can let people experience it, they’re definitely going to have more learnings than with any other method.
Conor Doherty: That actually sets up the transition here because again, we’ve discussed quite a bit how we prepare people, particularly young people, for careers. In fact, you mentioned leadership problems and then that actually sets up the next point, which is Philip, in your experience, what are the biggest problems that you’ve noted young people have when they transition from studying supply chain concepts to the real world of actually applying them in a corporate environment?
Philip Auinger: Expectations are a major subject, so first of all, kind of judging that yourself. You’re so smart now because you have a master’s degree in supply chain, so you know it all. And then you have your first day of planning in fast-moving consumer goods. I plan milk products, for example. I had a panic attack in my first week. I ran to the bathroom and splashed cold water on my face and tried to stop crying. I thought I knew it all, and then I hit the wall of realizing none of what I learned in theory there applies here today. With your boss being angry, the truck driver wanting to know where to load the goods, your sales rep being angry—there were so many things just falling on you which I wasn’t prepared for.
At the same time, I think that if you listen to the overall agreement on LinkedIn, for example, everybody is bashing Excel: “Oh, it’s such an old tool and it’s so terrible.” Yeah, we’ll come to that, but still, most companies are running their supply chains on Excel. And at the same time, most of the time, you’re using basic mathematics which apply also for the next 100 years. And it doesn’t matter if you do that with a tool called Excel or with a different tool; you’re applying mathematics.
I think that many students walk out and think, “Oh, I’m going to a big company and everything’s going to be AI driven.” No, it won’t be. You’re going to struggle between having messy data somewhere in an ERP, trying to get that out, and if you’re lucky, you can build that into an Excel and try to draw conclusions.
So I think being able to do that and to realize, “Okay, this is how we’re doing it now. Is there a way to get smarter? Is there a way to automate this? Is there a way that AI can do this report for me instead of me spending eight hours creating this every week?” That’s where students can make a difference because they’re digital natives by now.
If you finish your studies now in 2025, it means you were probably born after 2000. So you’re the kind of future for supply chains that can bring in the best of both worlds to say, “Okay, understand how the math works, I learned that, but at the same time, there are tools that are so much smarter than a human brain is.” I think that’s where students can really make a difference.
If you’re not good at either, you’re in trouble. So if you can’t do any math in your head, if you can’t double-check if what ChatGPT or whatever tool is presenting to you, if you can’t plausibilize—if that’s a word—if you can’t sense check that these numbers make sense, you’re in trouble because you will blindly report, “Look what my tool just said, my tool is so great.”
One more little anecdote, if you allow: we had an intern once and she was so proud because she did a big fancy analysis. At the time, she came to the conclusion or had the calculation saying that the warehouse cost for a small country, Slovakia, was at 17 billion euros at a time when global sales of the company were at 18 billion euros.
I asked, “Are you absolutely sure that your numbers are correct?” “Yeah, the analysis said it.” “Are you absolutely sure we’re wasting 90% of our profitability or 90% of our sales on one warehouse in one country? Are you absolutely sure?” “Oh, maybe I forgot a zero somewhere.” To which I said, “No, you’ve probably forgot about 27 zeros.”
I mean, I did have that actual conversation, but just at the back of my head I was thinking you need to be better at sense-checking numbers.
Conor Doherty: Joannes, I mean, you’ve been auditing companies for a long time. You’ve been running one for 16 years. I’m sure you have similar stories or challenges where people who studied suddenly transition.
Joannes Vermorel: Yeah, although you see the statement about digital natives, I would say yes and no. In a sense of, yes, the younger generations are not hostile to a computer, so we are indeed. But nowadays, you have to go for a very ancient, you know, like a 65-plus years supply chain director to find someone who is not at ease with a PC, dealing with spreadsheets and whatnot.
Maybe you will find some younger, but overall, I would say yes, but I find a lot of digital ignorance nevertheless, even among very young generation. And when I say digital ignorance, say if you go outside—let’s say FMCG, because FMCG is like the easiest in terms of data management, the number of products is very limited.
So you have high volumes, few products. If you go into really challenging supply chains, let’s say aviation, you’re going to have millions of parts around. And most parts—and you have parts that go from screws to aircraft engines. So the diversity is like crazy. Some stuff will be liquid; some stuff will be cables.
I mean, it’s the diversity of stuff that can be absolutely mind-blowing. Same thing, for example, for oil and gas. If you want to have the list of stuff that you need to keep an oil rig running, it is just mind-blowingly complicated.
Okay, now you realize that the data is not in one ERP; it’s like 10 ERPs. And because three of them are ancient were never phased out, and then there was merger and acquisitions, and so you have an applicative landscape that is I would say a consolidation of plenty of old ancient enterprise software products that can be dating from the 90s, sometimes earlier. And they are still around and they still work.
And as you see, where I say, yes, people are digital natives, but are they capable of navigating a complicated applicative landscape? That’s really a lot more demanding than just being able to fire an Excel spreadsheet and copy-paste a few numbers.
What I see is that too frequently, as soon as people face difficulties, they drop the ball and pass it to IT. And then IT has a four-year backlog. So, you see, yes, IT can compose this SQL query for you; they will be doing that in two years from now or something.
So that’s the sort of things where, if you’re truly—I would say what I would qualify as truly digital native—is there is a hodgepodge mess of applicative landscape where you just manage over day after days to gain knowledge about that, get familiarity, and be able to query those different systems.
Find help where you can and not use IT as a crutch, as a replacement for the skill that you should have, but as potentially a mentor to give you some help when you’re really facing something where you’re stuck and you just don’t know how to move forward.
But again, as a mentor, not as a sort of colleague that will do your work for you just two years from now.
Conor Doherty: Well, again, we’ve mentioned skills a little bit, and I just want to try something slightly different this time, given the present company. So what I’ve done is I’ve curated a list of a few skills, and what I’d like to do is pitch them to each of you.
You simply tell me if they are a must-have or a nice-to-have but not necessarily critical, and then you can quickly explain why. So I’m going to start with one of the most obvious and I will start with you, Philip. Must-have or nice-to-have? Excel.
Philip Auinger: For the past, absolute must-have. For the future, nice-to-have.
Conor Doherty: Okay, why?
Philip Auinger: Because, obviously, if you look at when Excel was introduced, I would guess that was in the ’90s. Ever since then, it’s been basically the only tool if you went to a company that was around. But nowadays, just look at the changes we had in the last few years.
Even look at something that many people consider very new, like ChatGPT. Look at the progress that’s been made in the last months. So I believe that in the next few years, functions that Excel can do will be done by other tools with clicking one button and making a good prompt.
Still, it’s good to, as I mentioned before, it’s good to be able to understand if that tool is now producing something correct. Basic math and basic logic and plausibility checks still make sense, but you won’t need Excel for that in the future. But these things may take a while, so I’m not saying how far in the future.
Conor Doherty: Well, Joannes, no cursing please! But Excel, must-have or nice-to-have?
Joannes Vermorel: Yeah, spreadsheets have been around forever. In fact, they were introduced in the late ’70s, and then Microsoft Excel itself was in the mid-’80s, but yeah, it’s ancient. I would say must-have mostly because, frankly, if you spend 10 days on Excel, you will already be very good.
I mean, if you work diligently, we’re not talking of investing two years of your life. You know, 10 days, if you do the effort, you will be already solidly good at Excel.
Conor Doherty: All right, thank you. Next…
Joannes Vermorel: Honestly, you only need something like 30 functions to run a supply chain. That’s all you need.
Conor Doherty: Yeah, yeah. Well, next one, Philip again. Must-have or nice-to-have? Python.
Philip Auinger: Nice-to-have. I would say that was a big trend when I left the corporate world, so that’s seven years ago almost back then. That was the new thing: Python, R.
Suddenly, ChatGPT came around and I believe that they just have so much better functionalities and so much better options of doing so many things that Python, R, and others kind of were in between future hopes but aren’t that relevant in the future anymore. So I would say nice to have.
Conor Doherty: Joannes, same question: Python must-have, nice to have?
Joannes Vermorel: I would say must-have but with a caveat that it doesn’t matter which programming language it is. So what matters is to master one programming language, it doesn’t matter which one, because it gives you the sort of mindset of what a computer can actually execute.
So suddenly, you can differentiate between “Oh, this thing can have an analytical resolution that an algorithm will execute and I will get this calculation” versus a problem where no, it’s like super fuzzy. Maybe an LLM can give me an answer, but fundamentally this is not something that belongs to the realm of things that are computable.
So I think knowing at least one programming language, it doesn’t matter which one, it doesn’t have to be fancy, but I think it’s very important to master programming, at least the basics of it in any language. It doesn’t matter which one.
Conor Doherty: Thank you. Philip, must-have or nice to have: Power BI or any data visualization tools?
Philip Auinger: Data visualization, absolutely very, very important. We were mentioning before, if you want to convince other people, a number graveyard like a big spreadsheet will not convince people. If you put that into a graph, suddenly you see things.
I don’t think Power BI is necessary for, let’s say, any person working in supply chain, but I do think it’s good to have a department that deals with this sort of stuff; to have those geeks who are able to create amazing Excel sheets that standard users cannot and kind of treat them as like a decision support department.
I think one of the companies calls their, I think it’s Nestle, their data analytics department, decision support. If you put Power BI into a team of people who understand supply chain and who create reports that then the rest of the company uses, that’s very, very valuable.
But again, I don’t think everybody working in supply chain needs that. So it’s a nice to have.
Conor Doherty: Joannes, same question: must-have or nice to have Power BI?
Joannes Vermorel: Yeah, I would say mostly nice to have. Again, here that’s the sort of thing where not having the concepts in mind won’t really prevent you from getting the proper guidance from ChatGPT.
See, that’s where, when I was comparing Python nice to have or whatnot, if you can’t even think the way programming is working, how computer instructions unfold one at a time with branches, loops, and whatnot, if those things are absent, you will struggle like hell to even be able to get an answer from ChatGPT that makes sense.
For visualization, I would say if you don’t have the sort of slice and dice primitives in your mind or if you don’t know that things are called, you know, bar charts, line charts, pie charts, and what, it’s kind of fine. The tool because it’s very visual will manage to guide you.
So I would say nice to have. I do not see, I don’t think that the concept, if you’re too ignorant, if you’re somewhat ignorant of those concepts, you should be able to muddle through. That would be my take.
Conor Doherty: So side note, just to clarify, Joannes, you are a native French speaker. A pie chart in French is called a camembert, correct?
Joannes Vermorel: Yes, exactly, we call it a camembert.
Conor Doherty: Sorry, that is so… just… I’ve waited two years to drop that nugget. Finally, it has presented itself.
Absolutely, that is a deep cut for people who know me. Anyway, sorry, the last one of this segment. So, Philip, must-have or nice to have: agentic AI?
Philip Auinger: I’m curious to hear what Joannes has to say about that.
Joannes Vermorel: Must, but it will very quickly devolve into remaining relevant as a digital native. So what do I mean by agents? It’s, for example, OpenAI’s deep research capabilities giving you an LLM that has the capacity to look at about 200 pages to give you a primer on a subject.
So you have this agent doing a loop, and there are plenty of situations where an LLM can fundamentally call itself to perform an iteration to complete a task that could not be completed with just one completion, one LLM completion.
But I believe that’s the sort of thing that is progressing very fast. Knowing when and how to leverage where and how, you want to leverage those sorts of capabilities will just be part of, “Do you know how to use ChatGPT efficiently or whatever competitor of ChatGPT there is?”
Again, it’s like for me, this is part of being a relevant digital native, just like being able to use Google Maps, Uber, and whatnot. There are so many apps where if you don’t know how to use Google Maps, people would think, “Oh, that’s? Yeah, it’s not very difficult.”
But again, if you don’t even know that the thing exists, then you are at a disadvantage compared to people who are readily using that. But acquiring the skill is totally straightforward.
What people think about, that will be the real future version. But think of it as the future version of ChatGPT where you delegate control over files that are on your computer.
Where suddenly this LLM right now is kind of boxed into a web page that is completely isolated, that was until recently completely isolated from the rest of the universe. Now, ChatGPT is very good at looking at the web and asking questions before it starts as well, like clarifying tasks.
That would help; that’s a way to break the boundaries. And the next boundary that will be broken will be, “Oh, I grant ChatGPT access to my local environment and then check files that are on my machine.”
That will come, and that’s where people start thinking about those agents because you may ask a question that requires more, I would say, iterations such as: “I have 200 Word documents lying in a folder. It’s a mess. Create ten folders or so, a dozen folders, and rearrange appropriately the documents that are on my desk.”
We are not there yet, but I would be willing to bet you quite a few dollars that one or two years from now those capabilities will just be mainstream. It’s not even very complicated to do, and thus again people will be expected to be able to use these sorts of things.
But again, that will fall under the general umbrella of being a digital native.
Conor Doherty: Well actually, Philip, on that note, how do you see things like AI, be it agentic AI or otherwise, but how do you see AI shaping the future of supply chain positions in general?
Philip Auinger: It’s exhilarating and very exciting, and at the same time I’m very afraid that we’re going to lose, especially in supply chain and analytical jobs, we’re going to lose 50, 60, 70% of available jobs. AI fans will always say, “Yes, but there will be new jobs that will be created.”
Yes, that’s true, but I doubt that they will be enough to fill these 60% who are now basically out of the job. So obviously there are amazing opportunities that you have with this, but no matter which, there are few jobs.
So really manual jobs, things like, I know, taking care of the elderly or chopping wood and so on, you will not be able to do that with AI. But things when it’s about dealing with numbers and making a decision—how much I buy, how much will we sell—that’s going to be replaced very quickly.
I think the key will be to be a person who embraces that change and still is kind of here to coordinate all this. So these roles will be very sought after, and that’s a fairly new concept. So that’s why if you’re looking at the next years in supply chain, that’s what you need to level up when it comes to skills.
Overall, generally speaking, for the planet and for humans living and humans who actually want to have jobs, AI is very tricky. So, I don’t want to paint everything black here, but there’s a big risk that comes with this.
Conor Doherty: Let’s hear your take on this. I’ve heard it before obviously.
Joannes Vermorel: Lokad is very much—I mean, we are part of the people who are actually pushing super strongly for a massive degree of automation. My take is that, yes, fundamentally back office jobs will be massively, massively automated.
I believe that supply chain back office jobs that are strictly back office—I mean, Lokad, what we deliver for our clients is a massive degree of automation, so that is coming and coming fast. But, to the audience, I believe that, and that’s what we typically recommend to our clients. I believe that the old school demand and supply planners was something that was incredibly looking inward. People were trying to look inside the company to find information to project demand and organize purchases and whatnot.
What we’re saying is that if you mechanize all this work that is inside the company, it’s just mechanized, but it paves the way for doing more. This more is connecting with the clients and connecting with the suppliers. So you see, if all the tedious clerical work, which is to consolidate the information and have the proper numerical recipe in place so that you can do the forecast, replenishment schedule—all of that is automated—then it means, okay, this is taken care of. But that means that suddenly you can start a game that is the next level, which is better cooperation with the clients and better cooperation with the supplier.
Here, we are not anymore a back office job because if you have to interface with the rest of the world, you become front office. And that becomes much, much trickier to delegate to an AI because fundamentally you are—what you’re seeking is just not pure analytical capabilities. What you’re seeking is establish a dialogue, get commitments from your clients, get commitments from your suppliers, etc.
So, my specific take on that is, yes, mechanization is coming. Lokad is part of this wave. But for young practitioners, there is a way, and the way is to make sure that they connect with either suppliers or transporters or clients. Any would work as long as you have this attachment to something that is not strictly analytical and inside the company, you’re safe. Because then your value will not disappear just because some people like Lokad just mechanize those numerical recipes.
Conor Doherty: Philip, I’m mindful that it’s now four, so I have one last question, and I’ll go in reverse order. Joannes, in the world of increasing automation, bearing in mind everything that we’ve just discussed, what remains the one skill or mindset that for a supply chain practitioner will not go out of fashion?
Joannes Vermorel: For now, I believe that high-level thinking is still very much outside what LLM delivers. So if you can really think, have crystal clear thinking about very complicated situations, we are not even close with LLM to be able to mimic that. LLMs are super good when it comes to linguistic patterns.
With numerical recipes like Lokad, we are very good into things that are risk estimation, quantifying risk, and whatnot. But in both cases, you need to have a very clear line of thinking about what needs to be done, why the architecture of the execution—these sorts of things. So my suggestion would be to cultivate high-level skills, and very frequently, the proxy of that is to just try to think as if you were the supply chain director of your company.
Try to think if you were the CEO of the company, try to have empathy with what their problems are. How can I—yes, my job is, right now, to deal with replenishment on this segment, but try to every day elevate a little bit what you’re doing to embrace the sort of high-level problems that your management might be facing. That, I think, will be a sure path to cultivate the sort of high-level thinking that is not going to be automated anytime soon, at least if we look at the sort of software technologies that are being pushed to the market right now.
Conor Doherty: Thank you, Joannes. And Philip, same question.
Philip Auinger: I would say, as long as you look at companies and their org charts, so as names of humans written—as long as we still have that and there’s no robots, there’s no R2-D2 and HAL, and I don’t know, KITT written in there, interpersonal skills will be front and center. Because that’s the only thing that differentiates us and that can make things possible where machines will say, “Sorry, you can’t have this because this rule didn’t apply here.”
But if you then, as a human, make a call and say, “Yeah, I know, but it’s really urgent. This is our most important customer,” they can make something possible. So, it’s these interpersonal skills that you definitely need to keep, and they, I feel, are part of a larger—not so much a skill set but really a mindset of staying curious your entire career. You talk to other people, you read industry magazines, you listen, you follow people on LinkedIn.
You understand what’s happening these days, what’s the future trends because it takes a while for these things to happen, but if they happen then you were able to prepare for them. Obviously, soft skills, they won’t change too much honestly. But everything else, everything around technology, look at what changed in the last five years and now project that into the future in five years.
So, yeah, as long as you stay curious and you’re willing to learn new things and forget old things that you thought were true but aren’t true anymore, you’re probably going to have the right mindset.
Conor Doherty: Yeah, well, I don’t have any further questions, Philip. Thank you very much for joining us. It’s really been a pleasure.
Philip Auinger: Thanks a lot.
Conor Doherty: Right, well on that note, gentlemen, I’ll draw things to a close. Thank you both, and to everyone else, get back to work.
