“The end goal for a lot of clients is this dream of a lights-out factory. Before you get there, there’s a stepping stone in between — and it’s this connected factory.”

Autonomous mobile robots spent years confined to warehouse aisles. Now, driven by a persistent labor shortage and a wave of AI, they’re moving into the heart of the factory — handling the material movement between “islands of operation” that people increasingly don’t want to do, and taking on mission-critical roles once considered too risky to automate.

In this episode of Manufacturing Matters, TECH B2B Marketing’s Jimmy Carroll sits down with Rockwell Automation’s Sheldon Zimmerman (Global Head of AMR Sales) and Isabelle Durso (Business Manager, AMRs) at Automate 2026 in Chicago to unpack where mobile robotics is really headed and what it takes to get robots up and running on a plant floor.

The conversation covers how AI is reshaping AMRs from the sensor up to the orchestration layer, the rise of mobile manipulation (and where humanoids do and don’t fit), why simulation and digital twins are becoming essential for de-risking deployments, and the real hurdles to adoption: IT/OT integration, physical infrastructure, and the brownfield-versus-greenfield divide. Zimmerman and Durso also make the case for starting small with pilots, the evolving safety standards that let robots and people share a floor, and why the “connected factory” — MES and ERP talking through a single orchestration layer — is the practical stepping stone toward lights-out manufacturing.

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. We’re recording at Automate 2026 in Chicago. And I have the pleasure of being joined by Sheldon Zimmerman, who is global head of AMR sales at Rockwell Automation, and Isabelle Durso, who was business manager of AMRs at Rockwell Automation. Thank you both so much for taking the time. I really appreciate it. I know you just got here too, so you basically had to run here. So, I appreciate the time, and I’m looking forward to talking to you guys.

Sheldon Zimmerman: [00:00:28] Awesome. Yeah, thanks for having us. We’re excited to be here.

Jimmy Carroll: [00:00:31] Great. So, I’d like to start with some kind of like high level questions and just get a sense of where you think things are in the market. And, you know, in the last year or so, what would you say are some of the most notable developments in robotics?

Isabelle Durso: [00:00:43] So, I think the biggest trends we’ve seen are kind of twofold. So, with the emergence of AI and the IPO of all of these AI companies, AI has kind of taken a front and center. So, I think AI has affected AMRs, obviously. We see this in the orchestration layers. We see this in some of the autonomous perception map generation, right? Self-guided driving. So, all of that is impacted by AI. It has had a huge impact on robotics as a whole, but in particular on AMRs. The second trend that we’re really seeing is mobile manipulation. Within the market, from a physical perspective, moving out of the software realm, mobile manipulation has been a great trend. A lot of customers are interested. I think it’s the big battle between humanoids and mobile manipulation on the wheeled robots that is currently kind of going through the industry. But I think that a lot of people are looking to have the robots do more than just transport goods from point A to point B but also manipulate the goods.

Jimmy Carroll: [00:01:57] You touched on a lot of topics there. I want to dive in on. One thing I do want to ask, though, just to get a little bit deeper, is when you say mobile manipulation, what are some really prevalent common examples of things that these robots are being asked to do?

Isabelle Durso: [00:02:10] I think a big one that we see over and over again is machine tending. This is a repetitive task that, again, with the labor shortages is very difficult to fill from a human perspective. So, if you can get a robot to do that task for you and to do the machine tending for you with an arm on top of a robot is perfect. The other piece would be more in the goods transport. Picking and packing or picking goods from a shelf, from a bin, and transporting it is another good application for mobile manipulation.

Jimmy Carroll: [00:02:47] Yeah, that makes a lot of sense. And you kind of give me a perfect segue into another question I want to ask because I looked at a jobs report recently, and it was close to half a million jobs open. And this was as of April. So, not the most up-to-date stats, but the labor shortage is not going away. This thing is persisting. And there’s a lot of jobs nowadays that, and I know it’s cliche at this point, right, but dirty, dull, dangerous. But there’s a lot of jobs people just no longer want to do. And as a result of that, robots are perhaps is really obvious, but robots are becoming way more important with every passing month and report you see. So, what are some real-world examples of how robots are keeping businesses going? I know you guys have a lot of good case studies and obviously a big global reach. What are some notable ones on your side?

Sheldon Zimmerman: [00:03:39] I mean, that’s probably better answered by myself from the sales side. I think there isn’t a Sierra or a Silverado that comes off an assembly line today that isn’t delivered by autonomous mobile robots delivered by Rockwell Automation. So, that’s certainly, you know, what we’re seeing is a trend towards more mission critical operations. If you think back two or three years ago and Isabelle can contest as well, um, attest, sorry. I think the people who were early adopters kind of looking at the technology, understanding what it can do, now they’re starting to rely on it for those really critical manufacturing just-in-time deliveries that deliver goods to their clients. So, we’re just seeing more reliance and more mission critical operations and more at-scale deliverables from an AMR perspective.

Isabelle Durso: [00:04:35] Yeah, and I think generally one of the big trends we’re seeing is this production logistics, right? So, bringing it not just into the warehouse but also bringing it into the manufacturing location, right? Any material movement between these islands of operation can be performed by autonomous mobile robots. So, bringing it from the warehouse to the line site. Those are jobs, like you said, a lot of people don’t want to do anymore that companies are having a very difficult time filling material handling roles and the pure material handling that is sometimes not even very complex. It’s a handcart. It’s a dolly. It’s a fork truck. And those are the things that if you’re placing people, you probably want to place people into the more critical roles that a machine cannot do. And now that robots can fulfill the simpler roles, such as material movement within these factories, I think that is a logical place for AMRs.

Jimmy Carroll: [00:05:34] Yeah. And it’s interesting, right? Because, Sheldon, you mentioned the use of robots in automotive, which is certainly not new. I mean, that’s sort of like the bread and butter from the beginning. But but these companies now are now using automation to do new things, new tasks. And you’re starting to see automation spread throughout the warehouse, throughout the plant floor and find new ways to add value. But a question I want to ask that is sort of interesting to me is, How can smaller businesses like small-to-medium enterprise, what’s a good entry point for them for automation?

Sheldon Zimmerman: [00:06:09] Yeah, well, I think it does start with a pilot project. Identifying kind of that low-hanging fruit inside of your manufacturing or warehousing facility where, you know you can get some early wins, some early successes doing those pilots. There’s a portion of those pilots that are important for the companies to learn how the technology can augment and complement their operation. But then there’s also a socialization factor with the people that are working with these AMRs and the environment and learning how they interact with humans in that space and then making sure that the organization is ready to adopt the technology at scale. So, really finding some early successes. I think when people go too large, too complex, too soon, then it starts to head down this path of nonacceptance. And then it’s really tough to get the ball rolling again.

Jimmy Carroll: [00:07:03] Yeah. And I was kind of thinking about this on a large scale. A Lot of warehouses today, I don’t know the exact percentage, but are not using automation. And it’s surprising. All the big companies like the e-commerce companies are, but there’s so many factories and warehouses today that are not using automation. And it’s like the robotics and automation technology is still sort of lagging behind. Why do you think that is? And what does it take to get robots up and running and productive in a facility?

Isabelle Durso: [00:07:37] I think the biggest hurdle that people look at is the integration of robots, right? So, first of all, there is this fear that it’s too complex and that people need to be experts in robotics in order to handle the robot, which isn’t true. But it’s the integration piece of it that makes it much more complex. It’s not just a device that you buy and you put on your floor and it’s set right. It’s the integration into the IT/OT layer. It’s also the integration into the physical infrastructure. If this is an existing building, how do you physically get the robot to interact with everything that is in that factory? Or do you need to modify any infrastructure in the factory? So, it tends to become a bigger project, especially like Sheldon mentioned, if you’re starting out really large, if you’re starting out very small with a handcart, it makes life a lot easier to kind of get familiar, especially with the IT/OT side of things. And then scaling up makes it a lot easier to then attack the physical infrastructure.

Sheldon Zimmerman: [00:08:45] Yeah, maybe just to complement what Isabelle said as well. I think it’s a big part of it, this brownfield versus greenfield. When we start with greenfield plants where they’ve tested out the technology, they’re building up these facilities, and they’re planning for that automation upfront. Okay, now we can go at scale. Brownfield, you’re shoehorning some of this stuff into these existing facilities. Makes it much more challenging. I think the other thing that people . . . When you think about the AI world, the software world very, I wouldn’t say very easy, but more easily integrated into existing architecture. When you’re dealing with physical movement of physical goods in a hybrid environment with humans and safety and moving pieces, and material has to physically get from A to B and physically transfer, those physical constraints are still things that need to be thought through, engineered appropriately, and done with an ROI case that makes sense. And those are all the constraints. I think when you deal with a physical being that relies on software that make it a little bit more challenging to adopt that at the rate that I think people are expecting from robotics mixed with AI.

Jimmy Carroll: [00:09:52] I want to touch on that AI thing in a second, but I think that it’s one thing that was just like flashing in my brain while you’re both answering is, what you just said really speaks to this concept. It’s not really a concept, but you hear it a lot, especially in the last couple of years, is ease of use. And maybe it’s obvious, right? Things should be easy to use, but robots can be easier to use than people think. And for these smaller companies where they look at it as more of a risk, companies like yours have gone to great lengths to make things easier to use and to troubleshoot and deploy. And I think that probably AI comes into play there a bit. And that’s one of the things I wanted to ask. So, as technologies like 3D and SLAM and laser scanners and time of flight, whatever, helped AGVs go from AGV to AMR, you know, you’re not using fiducial markers or tracks anymore. What does the introduction of AI do for these robots? How much more flexible are these now? And how does it help the end user?

Isabelle Durso: [00:10:55] I think from a technology perspective, obviously it takes it to the next level, right? Because now you don’t just have a sensor, you have a sensor with AI in it. And then that is converging with AI on the software side. So, path planning, predictability, and so on increases significantly when it takes all of that information together and processes it with AI versus just black-and-white decision making. So, it’s a much more nuanced decision making. It allows for higher speeds, it allows for quicker reaction times, and so on. So, we have that piece. But then we also have the piece that we come to the orchestration layer and then we come to also simulation. Now you can have digital twin, you can have simulation where you do the layout, where you do the path planning, or the map planning right beforehand. And that can be done with AI where you know what your throughput will be. And you know the impact that the robot will have before the robot even gets unloaded out of the crate. So, having AI be able to do this and then go into preventative and predictive maintenance once robots are running, this is where AI really comes in handy.

Jimmy Carroll: [00:12:14] Yeah.

Sheldon Zimmerman: [00:12:15] Yeah. I think just to complement that, again, I think once you get through that simulation and that deployment and that deep integration, I think from an AI perspective–at least when we’re looking at material movement and some of the robotics that Rockwell is taking part in–you’re looking more like machine learning. Adapting to the environment, identifying at the OT layer, at the unit level, identifying an obstacle, and then making a smart decision about how to get around that obstacle and then sharing that information with the rest of the fleet in that operation. And then as time progresses with that machine learning, allowing the system to take that as an input into the data stream so that it can make decisions longer term about how the operation unfolds and identify bottlenecks in your material movement strategy so that you can unlock those without having to take a whole data dump, get a data scientist, and kind of go through the effort of kind of filtering through all of that information. It should be happening real time.

Jimmy Carroll: [00:13:17] Yeah, yeah. I mean, it just sounds like it over time just enables more efficient operation. And you brought up a good point, Isabelle. Digital twins and simulation. Like how important are these today and, and how close to real life are these tools now? I did a podcast at the A3 Business Forum last year, and I think somebody said, Well, they are about 70% accurate depending on the type of application. But and I guess part two of that question would be, for companies that are maybe hesitant to deploy, what can simulation kind of prove to them?

Isabelle Durso: [00:13:57] Yeah, I mean, it’s exactly that. If they’re hesitant to deploy, understanding where your bottlenecks are, like Sheldon was mentioning. Understanding where the robot can have the biggest impact without having the biggest change required, both with people as well as the physical infrastructure. So, easing into it is a perfect way if you’re doing it with AI first, getting people to understand buy-in and getting people to accept the technology and running the full simulation. And I do think that simulation has come a long way, but I do think that that’s also where, going forward, in the future we’re going to see more and more happening is through simulation and digital twin.

Sheldon Zimmerman: [00:14:40] And simulation, just to complement that, again, I think simulation is as good as the inputs you give it. And so you kind of get garbage in, garbage out if you don’t have accurate information coming into the system. I think what I’ve seen some clients start to do is leverage some of the vision technology and the AI technology that you were referencing earlier in 3D cameras to gather real-time information. So, having an AMR or a robot drive the facility and gather that information and feed that 3D data directly into the simulation so now you’ve got a live look at what your environment looks like today. That type of accuracy did not exist 10 years ago, and it’s a huge driver towards getting more out of the simulations than you were getting yesterday.

Jimmy Carroll: [00:15:24] And for me, and I hate to say the same thing on every podcast, but one thing for me that is interesting is watching all these different technologies advance together to create new opportunities, right? Companies like Nvidia or companies that do industrial, other companies that do industrial computing or FPGAs or whatever, enable these faster, quicker decision edge processing, and then 3D. 3D cameras have sort of come of age and become much less expensive and smaller and easier to use than they ever have. And the ability to relay in real time, as close to real time as possible, like 3D point clouds, that is pretty amazing. And it’s really, really cool to watch. And I think AMRs are a good example of that.

Isabelle Durso: [00:16:13] And the other part to this is it’s not just the AMR simulation, but it’s like if you can simulate your full process, every part of the machine can be simulated, then you get some real throughput data as well. Not just the AMR piece of it. And I think for customers, that’s important as well.

Jimmy Carroll: [00:16:32] Well, we mentioned it earlier on, and given what you do, I have to ask, and it’s the humanoid discussion. When it comes to a wheeled robot versus a bipedal robot, where do you see humanoids adding value, and where would you like to see them adding value? I know for me, it’s maybe cutting my lawn, cleaning the pool, things like that. But in the factory, what do you guys think?

Isabelle Durso: [00:16:57] Yeah, that’s still a big question. So, humanoids are very good with unpredictable workflows. Normally, when we look into industrial automation, when we look into production, we tend to be very predictable because it’s very repeatable. It’s very predictable. So, humanoids, I’m not convinced that they have a steadfast role within the production area of a factory. But obviously when it’s unpredictable or there’s obstacles or there’s stairs to move, that’s where the humanoids shine, right? You could probably have them unloading shelves and so on. That’s an application that humanoids are also good at. But it’s like transporting and so on, they’re not very fast. They’re not very efficient, right? So, I think that’s where the wheeled mobile manipulators will probably be because they’re more stable, they’re faster. And if it’s predictable and it’s repeatable, that’s where they shine.

Sheldon Zimmerman: [00:17:58] And I think they just haven’t found their home yet. Like any technology, if you think about where AMRs for material movement came from, they came from remote control robots that kids in university thought were really cool at the time. And then they found a home in terms of scaling, right? And I’m not sure that humanoids have found that home yet. Whether it’s to your point around the house or inside of the industrial operation, I think the jury’s out still on where that scaling is going to happen. I do think to Isabelle’s point, there’s probably less likelihood of a one-size-fits-all in the next kind of five to 10 years. A humanoid robot that can cut your grass and then climb under your sink and fix the plumbing and then go outside and clean the pool. I don’t think we’re looking down that lens. I think we’re seeing it more from a use case and tool perspective. So, a robot to cut your lawn, a robot to cut your . . . to clean your pool. One that might help around the house or in an industrial environment, kind of application specific, because the functions of a human hand are still really hard to figure out. And there’s functions inside of a manufacturing facility that are going to be really hard to repeat because of that physical piece that I mentioned earlier, this physical environment that you interact with and require dexterity and digital sorry, and digits, your digits, to complete. I think it’s going to be a tall order, but I do see a lot of different form factors coming together to make this operation function. And then to Isabelle’s point, the orchestration layer that allows all of those systems to speak and communicate and optimize, I think is what we will see in the next, call it five years.

Jimmy Carroll: [00:19:38] And one major consideration beyond cost because whether it’s for the home, well, especially for the home, but in industrial settings, I guess cost could be less of a factor eventually. But it’s safety. I know that there are some companies that are actively working on this, and some organizations as well, but mobile manipulation and AMRs are proven. They’ve been working for a while now. And if they stop working, they just stop where they are.

Isabelle Durso: [00:20:11] I do think that humanoid research is pivotal for us as well in industrial automation because the research that goes even just into the hands into the motors and the encoders that go into each digit, right? I think just forcing the industry to rethink and to think differently and smaller and being able to accomplish some of the tasks that we want human-like creatures to have I think will also benefit, you know, the grippers that will go onto a delta robot inside of a machine.

Jimmy Carroll: [00:20:44] A really good point.

Isabelle Durso: [00:20:45] So, I think that the humanoid research will probably have a much bigger impact outside of just the actual humanoid robots.

Jimmy Carroll: [00:20:53] Yeah, that’s a really good point. I hadn’t thought of that. I know. I mean, within five years would be great. And everyone wants to see it to some extent because because it’s cool. And that’s what you picture from science fiction. And but I’m sort of with you.

Sheldon Zimmerman: [00:21:10] But safety, like you mentioned, is a major driving force for our clients, right? Improved safety, like you said, certain technologies are proven before industrial robots were proven behind fences, no one had a problem with lockout/tagout going behind the fence and do your work. And then you got into the cobot kind of generation where now we could work collaboratively with these robots. Now you’re seeing AMRs and robots drive around the facility. But we’ve allowed the industry to create some of the safety standards that people rely on to make sure their environments are safe. And I think that type of evolution needs to happen with the humanoid space, so that you’ve got this sense of trust that everyone is safe around them, right? And it will come. It just, like the other industries we’ve all partaken in in our careers, it takes a bit of time.

Jimmy Carroll: [00:22:02] Yeah. I mean, obviously these get maybe bad press, but you’ve seen a number of videos recently, I’m sure, of humanoids accidentally hitting people in demonstrations. And it’s like the safety regulations are trying to keep up with the advancing technologies. And it’s like once they catch up together, then maybe there’ll be some really cool opportunities.

Isabelle Durso: [00:22:23] Absolutely. Yeah.

Jimmy Carroll: [00:22:25] Kind of looking forward toward the future, do you do either of you have any major predictions or even expectations in the next couple of years?

Sheldon Zimmerman: [00:22:34] Well, like I said, I just think we’re going to see a whole lot of awesome new use cases and form factors of robots, whether they’re wheeled or bipedal, whether they’re working around your home environment or in the industry, I think you’re going to see an enormous amount of different form factors. And then it’s the software behind them and the orchestration and the communication I think that Rockwell and Isabelle and I are placing our bets on, that we think that’s going to be a big part of the industrial environments for the next number of years. But certainly I’m excited to see what the human mind combined with AI can generate in terms of those form factors and excited to, still with some years left under my belt, to see that in my career. So excited for the future.

Isabelle Durso: [00:23:25] I think that there is going to be two areas in the immediate future that we’re going to see a lot of growth. I think the orchestration layer is going to be one of the biggest developments as far as AMRs are concerned, right? Because you now have manufacturers, and interoperability obviously goes along with this, you have manufacturers that have purchased early on some robots, and they want to expand maybe into different form factors, and being able to orchestrate all of those different form factors with one orchestration layer between your MES, WMS, and the fleet management piece is going to be very important. I think the other piece is, like you mentioned, mobile manipulation is going to be the other piece where we’re going to see a lot of growth in the next one to two years.

Jimmy Carroll: [00:24:09] Yeah. I mean, even if you look back and, I don’t want to say when they exploded on the scene, but when LLMs first emerged, everyone’s going, well, these are cool, but I can’t really see a good fit for them in the industrial world. And now if you’re not already using them, you’re behind. And the rate at which these technologies are advancing now, who knows what we’re going to see in two, three, four, or five years. But, Sheldon, to your point, I’m sure it’s going to be awesome, and I’m looking forward to seeing what it is. Is there anything else. I mentioned we’re at Automate, so is there anything else we haven’t talked about? What are you guys most excited about? Like what will attendees come away with?

Sheldon Zimmerman: [00:24:52] Well, this is my favorite show of the year. We do a bunch, and this is by far my favorite. It brings together all of the industries that are really trying to leverage automation to push their companies forward. I think from our perspective, the story we’re trying to show is that connected factory, because I think the connected factory, I think like the end game, the end goal for a lot of clients is this dream of a lights-out factory. And I think before you get there, there’s a stepping stone in between. And it’s this connected factory. I think if you stop by our booth, you’ll see the connected factory. You’ll see the systems that Isabelle was describing before MES and ERP, communicating through an orchestration layer to deliver missions to an AMR. Cobots interacting with palletizers and AMRs themselves. So, I think the story we’re trying to show the customer is that Rockwell has it all. And I think we we have a really important story to tell in that connected factory. And I think that is how we will get our clients towards that end goal of lights-out manufacturing.

Jimmy Carroll: [00:26:00] To that point, I mean, I was going to wrap it up, but I have to ask because that’s a really good answer. What needs to happen? What innovations need to happen to get to the point where lights-out manufacturing is more widespread and possible?

Sheldon Zimmerman: [00:26:14] Well, I think some of our clients, what they struggle with is setting a true strategy and having that strategy maintain throughout the generations of people that might come and go through that leadership team. And so setting a strategy as an organization and maintaining that goal and keeping steadfast towards it and having a plan, right, is really like making sure you take the time to do the engineering and planning. We’ve had customers that kind of jump in both feet and want to get going really fast. And it’s great. It creates a lot of activity and a lot of different projects that we’ve been able to work on. But it also, sometimes, isn’t fully baked out or thought out. So, my recommendation to our client base is always take the time to plan, take the time to test it out, and then start to move. It’s not go slow to go fast, but it kind of is. You kind of do have to think through the process. Do the simulations, make sure that things are going to be adopted by the employees at your company. And I think those are the things that make it successful. You do all of those steps, then you start moving at scale. Now you have a blueprint that you can adopt at, you know, you’ve done it at one factory, but you own 15. You can now do it at 15 in parallel, right? You don’t have to . . . You’ve piloted, you’ve tested, you’ve vetted, it works. It delivers the outcomes. Now go. But if you don’t do the planning upfront, then sometimes you stall out. And if you don’t keep a strategy for more than whatever leadership team comes and goes, then you end up in this this starlight phase.

Isabelle Durso: [00:27:47] Yeah. You’re in it for the long game. I think is kind of the key. And alongside with that, one of the big pieces I think that’s important for manufacturers is that it requires a lot of change management through all layers of the organization, if that is in the hourly staff in the factories all the way up to the C-suite, there is a lot of change management and acceptance that has to happen. And that does take time, right? Along with developing the strategy and keeping the course that allows people to come on board and to understand the implications and the changes that need to happen in order to make that work.

Jimmy Carroll: [00:28:24] Well, thank you both so much for the time. I know it’s a going to be a really busy week for you. I would encourage everybody to go learn more about the connected factory and what Rockwell Automation has been doing. If anybody has any questions for us, we’d be happy to pass them along at manufacturing-matters.com. And once again, thank you so much for the time. I really do appreciate it.

Isabelle Durso: [00:28:43] Thanks for having us.

Sheldon Zimmerman: [00:28:43] Yeah. Thanks so much.