“AI needs data, and industrial data is not available on the general internet.”

“It’s not like you do either AI or manufacturing. If you’re smart, you optimize what you have. You would be amazed how much you can gain by just retrofitting your old asset.”

AI is finally moving “from PowerPoints to the real world” on the factory floor, but it can’t deliver when fed siloed, proprietary data. For Schneider Electric’s Andre Marino, the foundation comes first: open, software-defined automation that harmonizes industrial data, so manufacturers can train AI models on their own operations instead of buying technology and getting frustrated.

In this episode of Manufacturing Matters, TECH B2B Marketing’s Jimmy Carroll sits down with Marino at Automate 2026 in Chicago to talk through what it takes for manufacturers to modernize and compete — from the hyperscale “AI factories” racing to add capacity to the small and midsize companies that make up 85% of the U.S. manufacturing base.

The conversation covers the growing competition for energy between AI data centers and factories and why the answer isn’t adding capacity alone while digging into digital twins for upgrading lines with minimal downtime, closing the skills gap by drawing IT talent into the OT world, and a task-first (not form-first) take on humanoid robots and world models. Additionally, the discussion covers what Marino would like to see in U.S. robotics legislation, and his playbook for smaller manufacturers: know what you’re solving for, digitize, automate, optimize — and repeat.

Jimmy Carroll: [00:00:00] Hi, everybody. My name is Jimmy Carroll. I’m the vice president of operations at TECH B2B Marketing. And welcome to this episode of the “Manufacturing Matters” podcast, where we discuss the latest trends and technologies reshaping the manufacturing industry. Today, I have the pleasure of being joined by Andre Marino, who is senior vice president of industrial automation at Schneider Electric, at Automate 2026 in Chicago. Andre, thanks so much for joining me. I really appreciate the time. I know how busy you are.

Andre Marino: [00:00:23] Thanks for having me. It’s a pleasure.

Jimmy Carroll: [00:00:26] Awesome. So let’s just jump right in, with, I guess, a high-level question: What’s most exciting to you today in manufacturing and industrial automation?

Andre Marino: [00:00:33] Well, that’s a good question. You know, the pace and the advancements that we are seeing, the adoption of these new technologies… You know, you just compare a year ago to what you can see today? It’s super exciting, right? Because now we are seeing that these technologies are coming from the power points to the real world and the manufacturing side, and helping companies to be more productive to compete in the market. I think that’s a super nice to see.

Jimmy Carroll: [00:01:03] Yeah. We were talking about that on a podcast earlier. If you look at where LLMs were when they first came out, a lot of people were thinking, “well, these could be useful, but I’m not sure if we could deploy this on a factory floor.” And now it’s almost like if you’re not using it in some way, you’re behind. And the technology has advanced so much that it’s very exciting. And there’re a lot of technologies like that, [for instance] humanoids, which maybe we can touch on later. But anyway, I want to dive into some more specific questions with you. So, as we look towards the future of industrial automation, AI looms large, like we’re talking about here. A good analogy for me is, if you think about the way that AGVs went to AMRs. AGVs required  fiducial markers and tape and these kind of things. And then AGVs now incorporate things like 3D, time of flight and SLAM and laser sensors, and it allows these things to move autonomously. By introducing AI, these things become more flexible. But that’s sort of a good metaphor for industrial automation on the whole. What are some big ways that you’re seeing AI create new opportunities in the market today?

Andre Marino: [00:02:14] Yeah. Well, we all recognize the potential, right? And we are all exploring different applications where we can use AI. And I think it’s good to contextualize that. You rightly said the different AI models that you have from, you know, the old days, where we used machine learning and proprietary AI algorithms, that they were there in industrial applications for many years. Now you’re discussing large language models and what is next? Now, there’s one thing —— and I think it’s good for the audience to think about —— AI needs data, and industrial data is not available in the general internet. So we can train this model. So the first thing that we say is that, yes, AI is there, but there’s a fundamental thing that you need to structure first. So you can train your models based on your own industrial data, and then you can optimize on top of it. So the first thing about AI for us is the notion of open and software-defined automation, right? Because if you continue to have silos of data, proprietary systems, and you have one robot here, and then you have one machine there, and then you have one POC here, one POC there, different softwares will be super hard to harmonize this data, to give context to the data.

Andre Marino: [00:03:30] So then you can have your AI models, right? Either training your own large language models at the edge. Or, you know, moving ahead for the future. So I think that’s a good, an important thing to contextualize. This is a fundamental piece that we need to address. Otherwise people will buy technology just to get frustrated that, “well, you know, it’s not really trained on my data, so it’s not acting properly the way I expect.” Having said that, we see now after using proper tools to harmonize data, embedding AI, you know, in a diffuse manner in a plant. So we have our drives now, right? That used to be like intelligent drives, but programmed using intelligent models in a decentralized manner. Collecting data, feeding back the plant so I can orchestrate the entire plant. So it’s getting super exciting. And I think the gains and the productivity gains you can have, if you think about this properly, it’s enormous.

Jimmy Carroll: [00:04:40] Oh for sure. There’s one thing I want to ask about that I think that you, in particular, in your position at Schneider, would be almost the perfect person to ask. I went to the Schneider Electric Innovation Summit last November, and one of the topics that was most prevalent there was data center AI readiness. In one session, there was this company that had created this technology that helped cool data centers with plumbing. And I’m, like, I’m not thinking about plumbing innovations when I think about AI! But, it makes a lot of sense, right? For us to get ready —— when I say us, I mean the US, North America, or the world, really —— to be ready for AI to continue to grow, what has to happen at the data center level?

Andre Marino: [00:05:33] Well, the AI factories, or as they are called today, a data center. The hyperscalers, the AI factories, they are like huge facilities, right? We need them because they are actually the ones that enable us to run the AI models and to train all these models. So, one side presents an enormous challenge for society because they consume a lot of energy, they consume a lot of resources and a lot of money. So, there’s a lot going on. But they also bring a lot of good things for the economy because there’re multiple things that they activate in their value chain. But think about that. Because of these factories, we have one of the most powerful optimization engines that we’ve ever had, which is AI. So we need them to go super fast so we can continue to develop these large language models and other models that are coming, on the spatial world, the world models and everything that is being developed now on the physical AI, which is a good distinction to be made. And, you know, we also kind of compete for the same resources on the manufacturing side, because the same energy that I’m empowering in a data center. I need the same energy to power a factory. That’s why we keep saying, “hey, you know, the US has this unique opportunity to develop a lot on the data center and be the leader there. But also, we have the opportunity to reimagine manufacturing in the US and grow manufacturing.” But there’s competing resources here. So we need to be super smart in how we optimize our installations and our plants so we can do both at the same time.

Jimmy Carroll: [00:07:16] Yeah. It’s a great point. I talked to Congresswoman Jennifer McClellan, who’s part of a bipartisan group that introduced the robotics bill. And she said, “hey, it would be great if we won the AI race with China, but not at the risk of running out of food and water.” So we have to innovate not just on the AI side, but, you know, on the data center side and everything else as well. It’s something that doesn’t get talked about as much, that those facilities will need to innovate as well. And I heard a lot about that at your event and it was very interesting.

Andre Marino: [00:07:51] And if you’ll allow me, Jimmy, we talk about that and we implement things that we are talking about, but we believe that the US will not win adding capacity alone. Adding capacity, that is important. You know, there’re many things that we don’t have enough capacity, even in the manufacturing side. But we also need to get better in modernizing what we have and be more efficient. Because, you know, you are way more efficient in capital allocation if you take an old asset and modernize it. And we can do it. I’ll give you an example: we have a factory in Kentucky. In Lexington. It is a 66-year-old factory. There’s a program from the World Economic Forum that recognizes modern factories that are using technology, smart manufacturing technologies, that see a significant gain in productivity, in saving resources, etc.. And that [Lexington] factory was exactly recognized by the World Economic Forum as a “Lighthouse Factory.” So there is a very good example of taking an old asset, and modernizing the asset to reduce a lot of energy consumption. We reduced paper utilization by 90%. We reduced energy consumption by 24%, and significantly drove up the uptime of the plant. You can do these things, so it’s not like you do either AI or manufacturing. If you are smart, you optimize what you have. And that’s what we keep saying: It’s nice if you have a new plant, you can use the best of the technology or start from scratch, but also you need to pay attention to the 85% of small, midsize companies that are out there in the manufacturing space in the US. We need to help them to adopt technology and modernize what they have.

Jimmy Carroll: [00:09:37] Well, that’s a perfect segue into another question I wanted to ask you. In many lines of work, downtime is few and far between. So, you only have certain weekends throughout the year where you might be able to implement some new system. How can these companies modernize without too much downtime?

Andre Marino: [00:10:01] I think the answer for that is to create a digital basis of design. You can call it a digital twin, or you can call it different names. But we do have an effective way to create a digital model where we can simulate, we can test things, we can train operators, before we do any modification on the shop floor. So here, doing Automate, in our booth we are demonstrating that capability, which is: Imagine you have a line, a bottom line, or, you know, whatever other line and you have —— which is common today —— your customer, who is looking for different product. And then you need to change the machine, right? In the past, like you said, you need to ship the machine and then you change the machine, and put the machine back. Now we can simulate everything in our digital twin, in our digital model, and understand how it’s going to work, then ship only the parts to the shop floor, modify when we have a downtime opportunity and that’s it. So the time to address these changes is way less than what it would be if we didn’t have the capabilities of the digital model.

Jimmy Carroll: [00:11:15] So that’s a really good point. What are some other ways that manufacturers today are responding to this growing need for more energy efficiency and these green initiatives?

Andre Marino: [00:11:28] Well, I would say —— and it looks simple —— but everywhere you have an electrical motor, you have an opportunity to save energy. Just by doing basic edging drives and being more efficient there. And then you optimize the process. Because if you optimize the process, you use less energy. Sometimes your process is not too efficient. You’re running at a speed that you don’t need or you’re just wasting energy. And then you go for a complete optimization of the plant, which is also important. Every time that you reduce downtime or you reduce waste, you save energy. So the notion of being more productive is not just for the cost, but also you can be more efficient in saving energy. And sometimes the problem is not just having money to pay for the energy. The problem is not having the energy, even if you have the money. It doesn’t sound very common, but think about the amount of energy we just talked about going to data centers. It’s a competing environment, these manufacturing data centers. We just use more and more electricity. So we need to be more and more efficient. We talk about the supply side in the news, which is putting more energy into the grid. But we need to talk about the demand side, which is how we get more efficient. So your question is super important. And we have tools, we have software for that, etc.. So we just need to adopt this.

Jimmy Carroll: [00:13:12] So in terms of fully leveraging the advanced technologies that are available today, like AI and robotics and all the different innovations —— whether it’s AMRs or maybe eventually humanoids —— what workforce challenges exist and what changes need to be made for manufacturers to be able to fully take advantage of industrial automation today?

Andre Marino: [00:13:36] Yeah, that’s a super important question, because we need to recognize that we don’t have as many people excited about going to manufacturing as we should. So coupled with a generation that’s going to retire —— which is part of life, right? —— you have a skills gap. A scarcity of skilled people that can go into manufacturing. I think the good news is when you use open software defined automation and you decouple the hardware and the software, you start to attract a lot of the newer generation who are well versed in the IT world, to the OT world. So that helps a lot, getting new people into manufacturing and addressing the skill gap that we have. Also, we have tools that can help us to translate knowledge into models. And simulators that we can then use to train the new generation that’s coming to manufacturing. [Train them on] how to perform certain tasks in an industrial environment. And I think collectively as a society, we need to talk about manufacturing as well. You know, I talk to my son, “hey, look, I’ve been in this manufacturing world for many, many years. It’s super exciting! You have a lot of technologies, it’s not just what you think —— that’s it’s a dirty place and, you know, boring. Not at all! We are getting super sophisticated and super exciting. So I think universities, us parents, and everyone in the community needs to talk about and show how exciting the manufacturing world [really is].

Jimmy Carroll: [00:15:25] Yeah, absolutely. There’s a jobs report from, I think it was April —— nearly a half a million jobs still open in the US. So the jobs issue is persisting. It’s not going away. But it’s kind of funny that you talk about certain technologies helping to attract young talent. Whereas maybe eight or 10 years ago, people were fearing that robots were going to take jobs, but they’re actually creating new and more exciting jobs, because those other jobs that people were afraid robots were going to take? Nobody wants to do a lot of those jobs anymore.

Andre Marino: [00:15:58] Exactly. So, there is this notion where you get people to do certain tasks, certain jobs, that are way more attractive —— analyzing data and creating models, helping and training robots to perform these tasks that nobody wants, frankly, to do. So we need to address this gap because there’s a lot of jobs that nobody wants to do, and we don’t have people coming to do them. But then we need someone that understands cybersecurity. We need someone that understands how to train the robots to create digital models. So, we need to balance these things because it’s not like a magical thing that suddenly one day, you know, things will be fully autonomous and no human is involved. It’s not happening like that.

Jimmy Carroll: [00:16:41] You still need those people that understand that specific domain. I think of welding as one example. The stat’s a little old now, but at one point, the vast majority of people that worked in welding were approaching retirement age. And eventually, is that going to be a line of work that is extinct? Well, no, welding needs to happen. So you need people that can  work alongside those robots, and understand how to use them, at least at a high level. Maybe it’s not an engineer-type person. But as technology has become easier to use, that person that works on the factory floor can at least get the system running, maybe troubleshoot it. Ease of use comes into mind, in terms of these technologies. Anyway, I’m rambling about that because it’s a segue into a question I want to ask you: What does the future look like in terms of the current generation and the future generation being able to work alongside robots? What does that look like to you and how will that change over over time?

Andre Marino: [00:17:35] Yeah, I think that’s the question that everyone is asking now about humanoids and etc.. I think it’s, first, very important to understand the task and not the form. People are talking about the form, humanoids or this or this, but they’re not talking about the task. And I think this is important. Why? Because if you have a high-speed, low-volume type of problem, you can fix it [by using] a normal robot. That’s not a use for humanoids. Now, if you have a space that is low-volume, high-mix, with a  bit of variability, where probably humans are performing something today, that will be too costly to just move to a traditional robot. So, yeah, maybe it’s a good application for humanoids, but there is, I think, a catch here. And I’m not saying that cannot be used —— you know, there’re a lot of humanoids out there and they are evolving super fast —— but humanoids are not trained on spatial data today. Or, most of them, because I’m sure someone is doing this as we speak. But most of them are not. Some of these humanoids are based on large language models, which is a language model that is trained on written data. Right? And, you know, when WE interact, we know the consequences of our act. So we perform a task, but I know if I push you too hard, what’s going to happen with you? And I can adapt this.

Andre Marino: [00:19:12] The data that the humanoids are trained on, they cannot adapt to this. So they need to be trained with a spatial understanding. And there is a new frontier model that, you know, is what they call a “word model,” where you get sensorial data, vision, you know, all kinds of pressure and etc., and you build a model that they call world model that mimics our living space. So then you can train and THEN  you can move into, like, a new physical AI that can understand this, right? But this is a new development in terms of technology. And we are not yet there. Maybe in the future we will be able to do these things. And there’re many companies working on this as we speak. And then probably we will see in places with low-volume, high-mix, where I have some variability? I cannot do a fixed robot, let’s put a humanoid. But, even if you if you go there, you still have all the mechanics, like the hands. These are super advanced things that we still need to develop. So we are years ahead of a full replacement. You know, we can have a conversation with different people for hours, one would debate left or right, but that’s what we understand is happening so far.

Jimmy Carroll: [00:20:35] Having this conversation about humanoids —— I feel like I have to ask everybody about it at this point. But, it’s almost like the technology is advancing faster than the applications. The applications aren’t quite ready for it. But also safety. The safety standards aren’t quite there. People are actively working on it and it’s being developed, but safety is a major consideration for humanoids. Whereas, wheeled robots, AMRs are more established and well defined. And so that’ll be interesting to follow to see where the humanoids go, how safe they are, what certification they might get, both in the factory and beyond. Somebody I was talking to not long ago who works at a robotics company was saying, “hey, humanoids might have their place on the factory floor. But what if you had one at home that could just trim the hedges and mow the lawn and move things around and clean your pool?” And I’m like, “well, that’s great, but how safe is it? Is it going to step on my dog or kick my child?”

Andre Marino: [00:21:42] Yeah.

Jimmy Carroll: [00:21:43] But it is very exciting. It’s obviously a hot topic in the industry and something that everyone’s going to be keeping an eye on. Anyway, what’s exciting for you guys here, for Schneider? So for anybody that’s going into the Schneider booth here [at Automate], what’s the first thing you’re going to want to show them or talk about?

Andre Marino: [00:22:00] Well, I think definitely the concept of open and software=defined automation. You will see the orchestration of an entire plant. So you see the electrical side of the plant —— and people often forget, because they see the automation, but there’s a lot of electrical equipment and things going around that discipline. THEN you see the automation, the lines, the robots and etc., all connected and with the data layer orchestrated. So then we can run the AI models on top of it. And you will see that because you will see different companies that are competitors out there inside our booth, all talking about open and software-defined, because that enables an ecosystem to develop new applications. And sure, then we have, in our own core business, new motor control centers, we have new PLCs, new drives for different applications. We have drives that we are launching for HVAC and cooling, which goes very well with the current demand for data centers. So it’s super exciting to see that we have new motion control equipment for machines. So there’s a combination of software, hardware, and AI being demonstrated, something that we try to put in a context that’s more simple to consume and simple to understand how to apply in a manufacturing site. Because I think one of the the jobs that we need to all do is to demystify the use of technology and simplify the use of technology. And that’s what we are trying to demonstrate in our booth as well. We call it smooth automation, which is essentially translating to “easy to use.” ? Easy to consume. So I think it’s a super nice demonstration. We welcome you to go there and see that.

Jimmy Carroll: [00:23:57] Very cool. I do want to ask one more question. We were talking before we hit “record” here about a panel you were on at the A3 Business Forum in Orlando in January, which was really good. And, one of the topics that came up at that —— and I mentioned this earlier in the podcast —— was this robotics legislation. And I think it was Mike from FANUC that was talking about the importance of that. What would you like to see in a robotics bill? What things are most important to you?

Andre Marino: [00:24:24] I think, first, is one that recognizes the US’ place in the robotics industry, that incentivizes companies to develop here in the US. We were discussing companies that were very famous and then they disappeared because there’s competition out there. The competition, you know, is being incentivized to test new technology. There’s investment coming in and proper policy that removes a bit the uncertainty as well. So I think it’s a combination of predictability on the business side of the equation, plus incentive for the development of technology. And we discuss about universities as well. There’re big academia/private sector policy-makers working together to bring more skills to create a business environment that can flourish, where you can put capital and develop. And I think this is not “one or the other.” I think we need to contribute through the government and having these open discussions, ultimately to the benefit of the US, with a proper policy that incentivizes this ecosystem.

Jimmy Carroll: [00:25:38] Incentivize is a great word to use. Because it’s going to be difficult for the US to compete on a global scale with China otherwise. What else haven’t we covered? That’s exciting to you today that you might want to mention, a hot topic of yours, personal interests?

Andre Marino: [00:26:01] I think my personal interest is how we get the 85% of the manufacturing in the US that is small and midsize companies aware of this technology. And also that they get aware that they need to start adopting technology. Otherwise, it’s impossible to compete.

Jimmy Carroll: [00:26:23] Right.

Andre Marino: [00:26:25] And what is positive, but also challenging, is that these technologies are creating a gap between those that are adopting and the ones that are not adopting. And this can be putting you at risk, put you out of the business, for all the reasons we have just discussed here. So the the productivity gain, the cost reduction, the access to energy and etc., it’s kind of imperative to the business. So, that’s what excites us. But, also, it’s a big task for everyone: How can we help this manufacturing base to perform better? The big corporate companies, they have access, they have capital. They are making their choices with a lot more resources. Maybe good choice, bad choice, who knows? But they have the knowledge and understanding. So I think, Jimmy, your podcast has a tremendous opportunity to help these companies. Because you have a big base, an audience, to transmit that they NEED to learn about these new technologies. And not to be scared, because it’s not that difficult. We are trying as hard as we can to make it easy for you to apply and adopt technology.

Jimmy Carroll: [00:27:38] Yeah, there’re a lot of companies now that are —— and this is obvious —— talking about ease of use. And that sounds like a big thing that you’re trying to do, trying to make these technologies easier to adopt, easier to deploy, easier to troubleshoot, even for the non-engineering type people, for the smaller companies who look at automation adoption as as a risk. And in some respects it is a risk. But, with the proper application analysis and ROI calculations, it’s “automate now or get left behind.”

Andre Marino: [00:28:09] And you will like the example of our old plant in Lexington. You would be amazed how much you can gain even just retrofitting your old asset. Don’t think that you need to replace everything. There’s a lot you can do with what you have, especially in the software world. We can overlay that technology. And, yeah, we are trying to make it as smooth as possible. I think it’s a big test for all of us in the manufacturing world.

Jimmy Carroll: [00:28:38] Would you say that starting small, finding the right application to automate first, sort of getting your foot in the door or dipping your toes in the water or whatever comparison you want to use —— is that a good first step for these smaller companies?

Andre Marino: [00:28:56] Yeah. Think about the problem you’re solving. And probably you have a list. You want more productivity, you want to reduce costs, you want to tackle safety, you have cybersecurity concerns, or you just want to optimize an area. Think about, first, what you’re [trying to solve]. And then after you understand that, you go and digitize, you add sensors, you add equipment, you add software that allows you to capture this data. And then you go automate, and then you optimize later on. And then you repeat that cycle. The thing that I don’t recommend to people is to just buy the technology and figure out what you’re solving later. You know, if you go to a show like [Automate], you can get excited pretty quick. And then suddenly you buy a piece of technology that promises you a lot, but you haven’t gone through the process of what you’re solving for first. What do I need to capture in terms of data? So … automate. And then keep optimizing. Put together a plan. Think about it. There’re a lot of systems integrators, companies like us and others that can help you achieve that. When we have these engagements, we’re especially concerned to not push the technology for the technology’s sake, because we think it’s a waste of opportunity. Because if you see companies do this and they fail, then they immediately get scared, right? Because they lost a lot of money. And sometimes it’s not about solving the most difficult problem. Try to solve the problem that repeats many times across your plant.

Jimmy Carroll: [00:30:38] Sure.

Andre Marino: [00:30:38] Because you scale. You see a lot of people trying to test the technology, saying, “let’s see if I can test the technology on the most challenging thing.” And then that takes three years for you to solve eventually. But solving one small but important challenge, that multiplies many, many times and gains productivity, and then you gain confidence and then you go for even more challenging situations, right? That’s why you see a lot of people stuck in pilots because sometimes they’re not understanding that it’s not about [tackling] the hardest thing first. Try to see different problems that can multiply in many parts of the plant. It can be a controller of HVAC, it can be a motor you’re trying to optimize, it can be whatever you were trying to do where you can gain scale.

Jimmy Carroll: [00:31:36] I think that’s excellent advice. And I would encourage everybody to keep an eye on what Schneider’s doing. I think it’s just se.com, right?

Andre Marino: [00:31:45] That’s correct.

Jimmy Carroll: [00:31:46] Follow them on LinkedIn. There’s always a lot of good content there. And check out the Innovation Summit. Really worthwhile. Andre, thank you so much for the time. I really appreciate it. I’m looking forward to going upstairs in a little while and checking out the Schneider booth and learning more about what’s going on.

Andre Marino: [00:31:59] Thank you, Jimmy —— and I wish you all a smooth automation with Schneider Electric!

Jimmy Carroll: [00:32:04] Love it. Thank you.

Andre Marino: [00:32:06] Very good. Thank you Jimmy.