Enno de Boer: When did we had the last general purpose technology like electricity? Electricity was the last one. That's a, a century ago. And I think we need to understand and acknowledge that AI is not IoT, is not 3D printing, is not email, is not internet. It's really a GPT. It's a general purpose technology which will change everything Narrator: You're listening to Augmented Ops, where manufacturing meets innovation. We highlight the transformative ideas and technologies shaping the front lines of operations, helping you stay ahead of the curve in the rapidly evolving world of industrial tech. Natan Linder: All right. We are back again with Enno de Boer at Augmented Ops. Enno, good to see you again, my friend. How are you? Enno de Boer: Very good. It's really good to see you. Natan Linder: Yeah. And for those who are not familiar with, um, Enno, Enno's work, he's a senior partner in McKinsey, one of the key members, co-founder of the World Economic Forum Global Lighthouse Network. You know, we're almost kind of celebrating a pre-decade birthday, kind of good time to introspect, right? Enno de Boer: Yes. Natan Linder: Yeah, we go back, we wrote quite a few pieces together on, like, where Industry 4.0 was, was going or not going, and you've been here back in 2022, but it's time to revisit. So to kick it off, like, what's the state of the world? Like, why are we pegging it to February? What happened February from your perspective? Enno de Boer: With Opus coming out and entirely changing how we can use agentic- Mm-hmm ... um, and showing us the power, everything changed. But, but, like, to level set us, I think we're, like, almost 10 years into the fourth industrial revolution. We have seen outsized results out of over 200, 223 global lighthouses across more than 80 organizations, and, and I think the, the impact that we have seen Across the world, across industries, was truly remarkable. But we always talked about it, Natan, that it feels more like an evolution than a revolution. So, so I'm still looking at this and saying like, "Look, this was a great evolution. Good things happened. Industry benchmarks got reset, but it wasn't as disruptive, and it happened over 10 years." Now, fast-forward, what's happening now? So, so first of all, geopolitical tensions at a high, number one. Um, we're now, and seeing that in the data, that every supply chain has at least one major global disruption every year. That was like pre-pandemic when we talked about a decade for every global disruption, and then just before the pandemic it was around 3.7 years that it was coming, so it was accelerating. Now it's every year. Natan Linder: Every year. It's the new, it's the new normal. Enno de Boer: It's a new normal. Supply chains are in at- attack, and actually what I'm seeing is that in the good companies, supply chain leaders have a real spot on the table in the C-suite. Uh, they're shaping. They're determining whether you have the top line, whether you have the bottom line in the right place, and, and that has entirely changed. But still, a lot of companies are waking up to that, and they, some of the companies have still their supply chain organization a little bit like IT in the basement placed, and I'm very worried about those. But so what happened February 26? I think everything changed. Everything changed. Um, we see it in the software industry, and I think we should understand and take the learnings from what's happening in tech, what's happening in Silicon Valley, what happens with the tech companies, because, um, there's one study that they ran in '25, um, with software coding and agentic, and actually it turned out that the productivity went minus 19% using AI tools. Natan Linder: Mm-hmm. Enno de Boer: I don't know if you had seen that. Natan Linder: That, you know, I've seen, I've seen some of the MIT studies on this. It's definitely like- The jury is out to a degree, but at the same time, you're seeing the phenomena that many large tech companies, you know, you think about the Shopifys and the Intercoms- Yes ... are reporting, you know, we tripled our R&D co-productivity. Enno de Boer: So, so here's the interesting thing. Let me, let me finish this one thought. Yeah. So minus 19% in '25, they repeated the study now. Natan Linder: Mm-hmm. Enno de Boer: And there are two interesting results. Mm-hmm. Number one is they found now 20% productivity increase, so from minus 19 to roughly 20%. But the biggest so what was most of the software developers didn't wanna take part in the study as at all anymore because they would needed to work certain amounts of their workflows with AI tools and certain without, and they said, "I cannot work without it anymore." Natan Linder: Mm. Enno de Boer: So, so it's a total paradigm shift. Yeah. I think at the same time, you see the data is like Opus 4.6 is now, I think we're at 14 hours, 12 to 14 hours of independent tasks that it can handle. It's an explosion. We were like literally a year ago, we were like at, at minutes, uh, around it or an hour. Like it's, uh, it's truly amazing. So what does that mean for manufacturing? That means software is just the first domain. You can see that everything that we are seeing and experimenting in software will come to all the knowledge work, to all the white collar work. It will take probably a little longer because it's less structured, the work, it's more tribal knowledge, et cetera. So it's a little bit more complicated and it will go domain by domain. We see it already lawyered or the, the whole law domain is getting disrupted, but others, it will come. It's, it's only a question of time. Now, what does that mean for manufacturing? Nathan, I think the dream, I think a dream is coming through. A big dream is coming through. Tell me, Natan Linder: tell me about the dream. I mean, I'm, you know, I've been hallucinating about what, you know, we were, we were working here at Tulip like for a decade as well, like on democratizing effectively, you know, s- there were different names for this, but by and large is like saying like, "Hey, you know, software defined operations, you know, democratizing it and bringing the tools to the people, do the work." And there was also a bunch of AI in this before Gen AI became all the rage. So I'm absolutely ready to hear your dream. What, what's, what's the dream? Tell me. Enno de Boer: So in the first industrial revolution, when Ford put up the Model T line, 80% of the labor in the Piglin factory were labor, blue collar, 20% was indirect. We have inverted that. Natan Linder: Yeah. Enno de Boer: 2026, we're almost the opposite. We have, like, like 20% maybe blue collar workers, and then we have all these on top functions in the factory- Yeah ... we have the inbe- all planners, and then we have marketing, R&D, we have sales, all of that. It has exploded, and it got entirely out of check. So the value creation is only, like, 20% touch the product, and that's my dream is now we have the tools at hand that we can go after this, and we need to create a total different clock speed there, because it's slowing us down to create the value on the shop floor It's limiting us a lot because we have this big, big, big fixed cost block that is sitting on it and that has driven us to these monumental, s- huge scale sites because the fixed cost block didn't allow us to bring manufacturing where the consumer is. Natan Linder: There's like this tension I want to see what you think about when you're addressing this. So-- And by the way, I totally agree with your observation, to be clear, but here's the tension. It's at the same time, millions of hands in the Western world are missing from operations. Like if you would- Yes ... ask manufacturers, they don't have people to staff the shifts, and th- those people might be, you know, semiconductor engineers or welders or operators working in discrete. So that's like tension one, and then just one more arc to this. I think it's kind of similar, but it's a bit on the transition phase between how factories are set up today and the new ones that are coming. So obviously, this is, you know, what's the degree of automation and- Yes ... not getting to like some Elon or, or Jensen visions of humanoids. We can talk about that later, but just plain old automation that you're kind of building- Yes ... you know, green fields that are more automated, therefore, they need less people, and so on. How does that sit with like where AI is gonna play? In operation, I think we're really talking about physical AI. Enno de Boer: Yeah. Natan Linder: You know, which is very different. You know, the pipelines that we are seeing, like you mentioned, in, um, a legal pipeline, which is inherently digital. You know, maybe except the part that you print stuff and go to court. Yes. But other than that, these are people sitting in a office with a laptop and high bandwidth stream with an infinite token budget, theoretically, and, you know, converting their understanding of what they need to accomplish to tokens. But in the real world, somebody needs to make sure there's enough oil in the machine and, like, run parts to the line and unpack stuff in the loading dock and so on. So help me with the tension with people- Yes ... missing from the workforce and, like, this push for automation, greenfield, Enno de Boer: brownfield. So, like, the last century, we worked on getting the, the line workers at the value add, get them more productive. Natan Linder: Yes. Enno de Boer: And that will not stop, and with physical AI, we can do another leap there. But what I'm saying is, let's take a deep breath, and let's shift our focus for one minute on everything that's on top because it's slowing us down. And l- why don't we free up some resources to move them to the value add? Because we need them in the value add. Because what I wanna see then is, I wanna see not these kind of big, monolithic, several thousand operator sites. I wanna see distributed, hyper-local manufacturing. Natan Linder: Mm-hmm. Enno de Boer: Um, a lot of manufacturing that's happening in the communities, for the communities, near the consumer, if you can collapse the fixed cost. Yeah? I wanna see personalized manufacturing. I wanna go back from mass production to real personalized o- of the order of one. Yeah, I can see that, and I will need to move more to the value add with that, and I need-- and, and we're getting to that. We, we cannot forget the focus. The value add needs to run like a, like a oil machine. Natan Linder: Yeah. Enno de Boer: But at the moment, we have all this kind of sack sitting on top that is- Yeah ... slowing us down. Natan Linder: Some analysts who spend a lot of time in the space said, "You know, there's some companies, they're like, uh, dogs, and it's, like, hard to teach old dogs new tricks," as the saying goes, you know? Enno de Boer: Yes. Natan Linder: So those companies, kind of hard to transform and kind of guides their decisions. Some companies are like cats. They're like- uh, rambunctious. They, they b- can't be tamed. They go around, scratch and claw, and they're like, m- m- roll with their own vibe. And some companies are like lizards, you know? They're like, "This is mine. N- now I'm, uh, extending it." They're like, they need to feed, like, you know, sometimes once a season. They need the sun to warm them up until they move, you know? But they could be- Yes ... very ginormous and, uh, scary like, you know, the Nile crocodile or something like that. So like, you're seeing that in, like, de facto. It doesn't matter which one, which vendor. It's, like, clear as the daylight that AI is becoming de facto operating system for factories and o- otherwise- Yes other operational environment, and that means necessarily, you know, the, one of the core thing for an operating system using this metaphor, you know, it's like timing and coordination- Yes ... context switching, and that's what a operating system does. So let's talk about that. If that's really the case, how is that gonna work with all those different kinds of animals in the zoo that kind of be- have those different behaviors? Enno de Boer: Well, I hope that all these animals survive. I- Okay ... I, I don't know. Natan Linder: Well, let's help them survive. Uh, w- aren't we here for that? You know, we should help them. I mean, you do see the ones that, that did great work during Lighthouse and scaled up, you know, so maybe that's a Enno de Boer: good clue. Yeah, but, but just back to, to our journey that we have done over the last 10 years is we have 82 organizations that moved and are doing literally Fourth Industrial Revolution at scale, so that is a very, very f- small fraction of the world. Not enough. Yeah. Not enough. We said it over and over again, and we... I, I, I guess if we would have not set up the platform, it would be a third of that maybe because, um, I think we did a good job in bringing people along. But honestly, I, I think probably we will need some innovative disruption here, and we will need to have new companies, AI native, coming up and, uh, shaking the tree a little bit. Natan Linder: Yep. Enno de Boer: Because I think there's so much goodness that can come out of this is why do we ship the, the, the products all around the world several times until they reach the consumer? Uh, why are we doing this? It's not good for the planet. It's not good for anyone. Why are we not really creating products that are fully personalized and, and doing exactly what you need? Why do we have these mass products? It's just for the scale because it's in our old system. Yeah? There's something to it. And, and to your point, like maybe going to the operating system, I think you are onto something. I think it's all about creating systems. All these LLMs, they're fantastic, but models think. Natan Linder: Yep. Enno de Boer: Systems act. Natan Linder: Yeah. So I think that's a really key point. If we kind of step back for a second, I think, you know, the growth curve on the digital experience are sending a very clear message. You know, since the introduction of, uh, say, web browsers and mobile, you know, touchscreen, we have not seen such a pickup of UI interaction modality plus, you know, set of functionalities that kind of the world is adopting and not looking back. I'll give an example in operation. So we came back from Hanover, it was April, right? So, and Hava- Hanover was very weird. Like, you can feel the change in the industry just to... And, you know, it's like heart of Germany and super traditional kind of, uh, uh, tradition of, of the event, and you know the gravity of that, of course. But the show is, like, fundamentally smaller. Yep. A lot less halls, a lot more digital. It's as if one agent composed all the marketing for all the companies. Oh, my God. You know? Everything is, like, AI manufacturing, operations for, uh, you know, industrial AI for this and that. And you go to see the demos, the demos are like, "Okay, here is, like, some agent talking to some raw data for manufacturing contextualized." Enno de Boer: But we see this. It's this kind of putting a tool on top of the existing workflow. Natan Linder: Right. Enno de Boer: And the, the, the, the, the impact is zero. Natan Linder: Is that what you're hearing from, from- Enno de Boer: That, that's what it is ... from your partners Natan Linder: and customers, Enno de Boer: like? Yeah, that's what we are seeing. It's Amdahl's Law. Mm-hmm. Amdahl's Law says if you speed up one element in a system by 20% If the entire system is not upgraded and has a new workflow, you get zero results, and that's what we're seeing at the moment. That's why 2025 had zero impact, and that's why, uh, I mean, the MIT study, we can talk about it. It was right on one thing. I think there was no impact. It was wrong on another thing. I think it said, like, that all the value is in the back office. I think that's wrong. Um, the, the, the value is really in the front office. The value is on the shop floor. Where's the value? It's in manufacturing. Here we are talking. Yeah. So Natan Linder: in the Enno de Boer: supply chains. So Natan Linder: you know, so I... The metaphor is in my head is, like, it sometimes felt like horse carriages on highways. Enno de Boer: Yes. ' Natan Linder: Cause y- And the, and, and then the other way around, which is it's motorcycles on buggy trails. Yes. Those are the two th- Enno de Boer: That's how it feels at the Natan Linder: moment. It, that, that's how it feels, and, like, really what you need to do is design the metro for the first time fully. Enno de Boer: Here I will do a plug for my home country. Natan Linder: Let's go. Enno de Boer: Germany. Natan Linder: Yeah. Enno de Boer: I think Germany has a lot to bring here. Yeah. We have invented systems engineering. We are systems builders. So please, dear friends in Germany- ... focus on that. Natan Linder: Yeah. Enno de Boer: Build a new system. It's time. We have all the tools, but we need to build new systems. Again, I think l- a lot of people talk about, oh, we need to own the orchestration layer. Like, yeah. Yes, but only if you have below that orchestration layer a system, an operating system, an NI system, a data system, a full system that is working and acting in a new way with very different roles for the human, very different roles for the, for the technology. That all needs to be figured out. This is the moment for us supply chain people. Natan Linder: Yeah. Enno de Boer: Because we have done production systems at least since the 1990s, since the machine that changed the world, MIT came out with the Toyota production system. Everyone built their own production system, which by the way, if I go on the shop floor now, Natan, it's really stunning. It's gone. People have lost- In what, in what sense? In what- They have lost it because they lost the focus. So dec- I'd say it's Natan Linder: decayed, not gone. Enno de Boer: It's dec- It's, it's decayed, yes. It's decayed. Natan Linder: There's like, you gotta do a renaissance to this thing. Enno de Boer: Oh, Natan Linder: yes. And of course, I believe, like, the future would be many new types of factory, decentralized, distributed, more automated, but I also don't think that humans will be out. Enno de Boer: No, no. Humans play a big, big role. Natan Linder: And all this discussion on physical AI when it's, like, kind of focused on how a car uses, uh, world models to navigate and how a robot brain is, uh, wired up, 'cause now you can train them faster, better. But then I'm like, what about the humans? You know? They're gonna be in this physical world, and humans actually need... And I think there's this phrase, like we were talking about human in the loop for AI, and I also think about the other direction, so AI in the loop of humans. Because- Yes ... you know, a h- a human is gonna be there making all sorts of, uh, judgment call. And, and I think a lot of the trust and responsibility for orchestration is moving to humans. Enno de Boer: I mean, it's funny, Natan. I, I think we are always going over, over the edge on this. Natan Linder: Yeah. Enno de Boer: When we offshored work, uh, to low-cost countries, how much of the work of an organization did we really moved over, and how much did we retained? We were never able to do 100% over and then, like, just kind of one person stays behind and sells the stuff or something. Didn't happen. Natan Linder: Didn't happen. Enno de Boer: This is the same thing. This is like outsourcing. No one has outsourced 100%, yeah? They- where are the outsourcing rates? It's the same thing. We have a new player. That's the agent, yes? It's like someone who's maybe outsourced. It's someone who's offshored or something. We need to build a new system and, and yes, it will change the balance of who's doing what, and we're in the process of figuring out who can do, and then theoretically we know what we can do better as humans and what machines can do better. But finding that sweet spot and, and creating the system that has then a higher output- Natan Linder: Yeah ... Enno de Boer: is, is, is a question. And I, I'm a true believer, just to put it here, is all that technology is to augment the human and get better results and make it safer, make it more fun, and, and let us concentrate where we are really good at. Natan Linder: Yeah. I'm kind of trying to think broadly on this idea of, you know, continuous transformation that I've been pushing against the, the, the silly term digital transformation that I think it's like now people might, uh, say AI transformation. I just think it's like the designation is about continuity, so it's like never over. Enno de Boer: Yes. Natan Linder: You see an existing state of some organization and, you know, they all got the memo, and they all understand and I'm imagining it's like, you know, the town next door got electricity wired up, you know? Yes. And they're kind of doing okay and, like, maybe a couple cars have appeared and, you know, this thing, combustion engine is, uh, safe enough and like, you know, they can start doing... And so they see that in the town next door, but this town is not connected yet and, like, they're kind of debating, like, where do the horses need to move, you know, or something like that. Yes. So it's like, how do you help them start and, like, truly get there? 'Cause the cognitive dissonance that they are is that those people who make those decisions on the operation, which is very physical and, you know, heavy with equipment- Yes da, da, da, machines, all that, they sit in an office and all day long they get, "Oh, we've added this agent. The CIO brought this," uh... You know, in the digital world there are corporations, like, adopting, like, some sort of a wildfire, whether it's good- Yes ... or bad, I don't know, but they're certainly spending, you know. Anthropic didn't get to its revenue, uh, level without giant enterprises spending across the bo- It's not just all coding. It's, a lot of it is just, you know, people- Enno de Boer: 50% according to the figures is coding- Yeah ... and the rest is non-coding. Natan Linder: But the re- Yes. Yeah ... so the rest is very substantial, marketing and HR- Yes and this and that and that. Yes, yes, yes. So they're living with that cognitive dissonance and, like, how the hell do they start? Enno de Boer: So first of all, I think one important thing is that you're saying, and let's be very clear here, you, you're talking about this comparison with electricity. I think that's a very important one. Let's, let's take a deep breath on that. When did we had the last time a general purpose technology like electricity? Electricity was the last one. Natan Linder: Yeah. Enno de Boer: That's a, a century ago. And I think we need to understand and acknowledge that AI is not IoT, is not 3D printing, is not email, is not internet. It's really a GPT. It's a general purpose technology Which will change everything. And first of all, I think before we walk forward, we need to agree on that. We're thinking it's just like, because, Natan, we have seen it in our journey in the fourth industrial revolution, and then IoT came, and then this technology came, then the digital twins came, and, and, and. And it always added, but it didn't change fundamentally the use cases. The use cases got better, they got more value out of it, no question. But it was all new infrastructure that was added to it, and we got better results, but nothing was a true step change. I think what's happening now is very different. It's GPT moment. Natan Linder: And how do you tell people to grab it and just start? I think that's what, uh, McKinsey knows how to do best. Now they- Yes. So- ... it goes into organization and says like, "Hey, you gotta take the medicine and jump three times and go left four times, and Enno de Boer: maybe you will get there." So I will answer. Look, I think we're at the very beginning of this journey because it just happened three months ago. I mean, also, see the magnitude, but some say it will be 10 times bigger and 10 times faster, and I could believe that. Yeah? So now how do you do that is, first of all, we need to acknowledge we are back to 10 years ago. We're in pilot purgatory. Mm. Everyone is do AI pilots, everyone is doing POCs. They feel very good that they have some chat engine running where they can ask smart questions and they get zero results. Mm. Zero is translating to the bottom line. I think a lot of CEOs are waking up to that, and I think there's, there, there's a group of CEOs that are seeing and saying like, "Look, this is a GPT moment and we need to do something fundamentally different." And there's another group that say like, "Yeah, but maybe I'll wait a little bit, huh, until others have figured this out, and because I don't want to rock the boat now." Mm-hmm. Now, what do we say is like, it's- Natan Linder: It's a strategy question, no? Enno de Boer: It's a strategy question. Yes. You need to start literally from your economic leverage points. We just brought out a new book, our Rewired book. We analyzed 20 AI transformations that were successful, and it starts with the C-suite looking at that and understanding what are our economic leverage points. Typically, a company, um, has something like five to, to six economic leverage points. Two or three are making 70% of the value. And when we are talking about the value, we ne- need to talk about these 20 transformations. On, on average, they made 20% uplift of their EBITDA. So, so we are talking here about for, let's say, a $20 billion company, we are talking easily about a billion of profit that we are going after that. So, so now you have CEO attention, now you have attention of the business leaders. By the way, also, like, that's a new thing. It's not anymore about use cases. You could do a couple of use ca- but you know nobody get benefit of, out of doing one use case. The lighthouses had on average- 10, 10, 20, yeah ... 20 to 40 use cases, and it's here the same. So you need to pick a big domain or an end-to-end process around an economic leverage point, and you need to figure out what are the, the 20 use cases you need to deploy, but more important, what is the operating system you need to enable that? What is the AI system, and what is the data system? And that you need to build, and that costs money. You get a good return. We see the returns are past two times, yeah? And they can be higher, and I think as we are getting better, they will get again back to the four times that we had in, in 4IR. But, but you need to start there, and if you're not willing to start there, probably you should first slow down before you accelerate. You need to know what are your economic leverage points, what can they bring, and what are you investing? I mean, Natan Linder: I'm seeing, I'm seeing a lot of, uh, uh, you know, big proclamations on, you know, billions of dollars of ROI unlocked and things like that, and at the same time, you know, we're talking about purgatory. So what is the coolest, like, r- uh, kind of AI use case in operation that you've seen that you're like, you went like, "Oh, there's a light bulb. This thing is good"? Enno de Boer: Well, now you're asking me a trick question here because you're asking after after use case. There's not one use case. So, so I can give you a use case, but I'm- Yeah, Natan Linder: but, but I'm intentional about this because I think- Yes like to get the 40 or 20 or whatever it is, we need to help people understand at least one that works. Yeah, so, so, so no- I, I'll tell you one after you tell me. So- Enno de Boer: Yeah, Natan Linder: tell Enno de Boer: me- How, make a deal ... tell me one. I'll, I'll tell you a few, but like, tell me one. Natan Linder: The one I've seen in one of the most strict, like, highly regulated, uh, you know, tier one pharmas, that they're doing AI-assisted human-based review. So to be very clear, this is a GxP environment. The control is the human. But a human needs to read and validate, like, literally hundreds of pages of EBRs. It doesn't mean that the humans stop doing that v- validation according to the SOP and exactly like it means, but at least in a first pass, an AI can mark the human like, "Hey, I pre-read this for you, human. Here are the places where..." And then the human goes in and say, "Actually, correct, correct, correct, correct, correct." And then it says, "Oh, all those things are correct," so now the model can get tuned and understand that. And so I think this is very real. Why? Because this is an example where, you know, this thing is very good at processing large amounts of text that is structured, and if constrained correctly, meaning there's a platform, there are the backstops, and the right connectors, tools, then you could get from, uh, something that could behave in a non-deterministic fashion, which is very scary, of course, in this environment, to deterministic or close to deterministic- Yeah at, uh, 99%-type a- accuracy. Even after that, it's not good enough, so human is checking. But what it does is it saves a batch release tens and tens and tens of hours because, like, it's actually right. So if you have, uh, someone who can do this instead of a 10-hour pass on the full EBR, then that, that... I think that's pretty real, and I think that some- Yes people will be like, "I'm not going back," you know, to doing it without it. What do you think? I Enno de Boer: think it's a good one, but, but so, so it's a good one, and no but, but- Then still a but because I'm German. Natan Linder: Okay. Because Enno de Boer: you're a Natan. Because- Natan Linder: Yeah ... Enno de Boer: it's, it's a use case. Natan Linder: Yes, it's a Enno de Boer: use case. You're right. So, so tell me how you transform an entire domain, and you talk about speeds. How do you, how do you create- Natan Linder: I, I'll tell you because I think it's built on fundamentals that allow stuff to scale in, in my view. I'm very biased- Yeah ... and so on. Why? Because it can be packaged in an agent or a skill, so it's now reusable. And, you know, in Tulip we talk a lot about composability. I think AI and agents are inherently composable. Yes. The bar to create computation that is repeatable and could be used routinely now is natural language. But the output of process engineers, uh, industrial control, safety, lean, operational excellence, all those folks, is not a GitHub with code. It is action. You talked about earlier about the action layer, right? So for us, it, like, lives in, you know, automation workflows, apps, what have you. So if they can make one, they can make hundreds. And I think- Right ... that is scalable. But, but I- I think it is cool ... always hear this, like, "Oh, where is the cursor moment for operation?" I'm just like, "Stop looking for the freaking cursor. There's no cursor for operation." You know, it's not a s- straightforward IDE, you know, digital first pipeline. It's something else. And again- It is ... I don't think we have all the answers, but that- that's why I think this can scale. Enno de Boer: Look, yes, I agree. So, so first of all, look, we know what the use cases are. In the Global Lighthouse Network, we have found, like, 150 use cases that are used over and over again, and have already delivered outsized value. So why don't we start with those? And if you bring agentic in- Natan Linder: Mm-hmm ... Enno de Boer: they can even excel at what they're doing, because now you take tedious tasks out that you delegate to the agent, you can get to something really phenomenal. I give you one example. I was, um, I, uh, I told you I was this week on Sapphire. Natan Linder: Yeah, how was Sapphire? It Enno de Boer: was pretty exciting. Sapphire was great. So, so I had a keynote and a panel that I, uh, moderated on agentic supply chain. Natan Linder: Mm-hmm. Enno de Boer: And, um, a real good conversation and, and good conversations all around. So I had on the panel Jeff Thiel from John Deere, and he talked about something he had deployed at John Deere, and I think that one is, is a pretty cool one. It's, it's nothing out of the ordinary, but so important. So- What Natan Linder: did they do? Enno de Boer: They created this agentic system around, uh, material availability and parts finding. Mm-hmm. Because you can imagine when you put a tractor together, there's probably 20,000 parts that go into a tractor. If one part is missing- At least, Natan Linder: yeah Enno de Boer: you're not getting it to the finish of the line. So, so getting the parts and finding the parts and anticipating that and looking into your supplier systems, et cetera, they have created an agent for that. And I think what was interesting is, and that's why it's a, uh, for me, a good example, is he, he was talking about the, how, how faster they are now finding time and how much less resources they're using, but then he said like, "Look, if we can't find the part, what are we doing? We're triggering our planning system to replan." So now you hand it over to another agent, and there you have the planning agent. And we just worked with a company that is one of the top supply chains, um, in terms of electronics, and they have put in place fully agentic supply chain planning, which I would call a system, and they created a billion dollar of impact with that. And what did they do is they, they brought their clock speed up ten times in terms of planning. They are now re-planning literally on the fly every day. Natan Linder: Yes. Enno de Boer: And so something happens, Joe comes not to the shift. I have not the part there. I have a supplier that goes down- So Natan Linder: now they adjust ... Enno de Boer: immediately. Not like everyone else that takes- Yep um, a four-week cycle for their S&OP. And you ask yourself why in twenty twenty-six we still need for an S&OP cycle a full month? Natan Linder: Yeah. It's crazy. You know, this is, this is actually a really important point, and, um, I'll let you have the last word, uh- ... because we're coming out on time. But, but what you just said, I think highlighted what I started to feel and also kind of hearing from people, 'cause a lot of time people talk to you about all this AI adoption at scale, whether it's, you know, everybody understands it's coming, like we discussed today. And the first sort of it's, "Oh, we're, we're gonna gain more productivity, and it's gonna be productivity." And they're not wrong, by the way. But what I really understood is that, like, for this, like, big non-evolutionary or revolutionary shift that we are discussing here, the people who think about this the best, first of all, they're trying to predict where the models will be in a year, not, not where the models are right now. And the second thing is that they're not thinking just in terms of productivity. For them, productivity, as soon as they made the mind shift, is a given. They're thinking about growth and speed. So it's like beyond productivity. So getting to the point that I'm not actually can just meet my current plan. It's like, how can I quadruple my plan or some crazy number like that. And, and I think honestly, I know these are, like, big words, but I think a lot of it has to do with culture and mindset of leaders. Enno de Boer: Yes. So look, if I have the last word, the last word will be clock speed, and I tell you why. I think it's not about productivity, and we all know this. In the end, it's, it's about doing totally different things. But if you wanna put one metric to it, my advice to everyone would be look at speed. Look at clock speed. How do you increase a clock speed of your company by 10X? And we're seeing it in software. I have an example where my colleagues, uh, put an entire agentic system into a bank, and literally the night shift was done by the agent system, and in the day shift, the, the, the developers would check and do that. And with that, they got already, like, 10 times. But then they figured out that literally the programmers couldn't keep up with what was produced in the night shift. So they started to create an entire looping autonomous system where they would bring the problem-solving. The human would come in to problem-solve. The human would come in to validate, to check, but would not stand in the way, and they got to 100X. So here's what I want. I want manufacturing to increase its clock speed by 100 times. That's my ambition for this next 10 years. Let's get there, and then wonders will happen. What can we do? We can innovate faster. We can bring stuff faster. We take politics out of the system, um, because there's one source of truth. Th- this you gotta, you Natan Linder: gotta send this memo to the EU as well, you know? Enno de Boer: So Yes. But clock speed will solve a lot of things. You, you eradicate waste. You, you eradicate kind of all these kind of back and forth. It will be liberating. So my last word is clock speed. Natan Linder: Enno, it was really awesome. Thank you for coming again for Augmented Ops and going deep on this stuff. Really appreciate it. Always a pleasure. Narrator: Thank you for listening to the Augmented Ops Podcast from Tulip Interfaces. We hope you found this week's episode informative and inspiring. You can find the show on LinkedIn and YouTube or at tulip.co/podcast. If you enjoyed this episode, please leave us a rating or review on iTunes or wherever you listen to your podcasts. Until next time.