“We’re at a perfect storm, finally, of being able to see automation really grow again.”

Automation is more approachable than ever before, with more experts in the field — even for very specific applications — available to help walk you through the process of understanding whether certain applications can or should be automated. If you don’t have much experience, start simple and take a crawl, walk, run approach to implementation. Work with your supplier or integrator to set clear expectations that can help you start building up your capabilities and getting your return on investment sooner, suggested Stu Shepherd, president of Shepherd Solutions, at the beginning of this special episode of the Manufacturing Matters podcast.

In the episode, hosted by TECH B2B Marketing’s Jimmy Carroll, Shepherd discusses this topic along with several others, including risk assessment, ease of use, and the importance of defining clear use cases before jumping into robotics, AI, or humanoids. In addition, the podcast covers lights-out warehousing and manufacturing, the concept and reality of physical AI, workforce problems, marketing hype cycles, and what the next decade may hold for robotics, warehousing and logistics, and human-machine collaboration.

Jimmy Carroll: [00:00:33] 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. I have the pleasure of being joined by Stu Shepherd. Stu, thank you so much for taking your time. I really appreciate it.

Stu Shepherd: [00:00:46] Thanks, Jimmy.

Jimmy Carroll: [00:00:47] Stu is the president of Shepherd Solutions, and he’s got a vast background in automation. I’m really excited to talk to you. I should say continue to talk to you. We’ve been chatting for a while, while we get set up here, but we’re here at the A3 Business Forum 2026. And for you, what are the things you’re most excited about in automation today?

Stu Shepherd: [00:01:06] So, I think the thing I’m most excited about is the fact that for the first time in my lifetime, automation is accepted by more and more companies as being a solution. In the past, people would rather use manpower and other things than consider automation. But with the lack of manpower availability and the desire to work in warehouses 24/7 or factories that are dull, dirty, and dangerous like in the old days, that kind of stuff, it’s caused people to consider automation more strongly. And there’s more high value, really high-functioning solutions available than there’s ever been before. So, we’re kind of at a perfect storm, finally, of being able to see automation really grow again.

Jimmy Carroll: [00:01:44] Yeah, it’s almost like it’s no longer just a nice thing to have. It’s required especially for larger companies, but right, for the smaller companies, these small- and medium-sized companies, how can they adopt automation? What are some. It’s like I saw a really good clip of you recently. And you’re talking about how adopting automation is easier than it’s ever been before. Can you explain that a little bit?

Stu Shepherd: [00:02:06] Yeah. So, there’s more automation experts available that have solutions that really work in very specific applications. And if your company happens to match up with the applications that are available, there’s a lot of expertise out there to coach you on how to get through that. Now, you may not necessarily know which jobs are the ones that can be or should be automated, but there’s people out there like our company and others that can help you walk through that automation planning process. The other thing that’s kind of important is if you don’t have very much experience with robotics, just kind of take a crawl-walk-run approach to implementation. Start simple with something that’s very realistic. Set very clear expectations both between you and the supplier and get started. You know, don’t wait until perfection. Don’t wait until Nirvana. Don’t wait for humanoids because there’s things you can do now and start making money now and get the return on investment and then build your capabilities up. Kind of like exercise, not that I look like I exercise, but, you know, start doing reps now to start getting stronger and smarter about how to do automation. It doesn’t take a PhD to do it. It doesn’t take, you know, college degrees and all kinds of other things to get it done. There’s a lot of automation available today that anyone can deploy if it’s the right application for what you’re trying to do.

Jimmy Carroll: [00:03:24] And that’s a really good point. right? And once it’s deployed, so many companies focus on this–it’s not really a buzzword, it’s kind of something that’s become ubiquitous–is ease of use, right? So, now there’s technologies that are deployed and whether it’s on a big factory floor or a small shop or whatever, folks that are not robotics engineers or systems integrators can use that. What are some key examples of technologies becoming easier to use that you’ve seen in the last couple of years?

Stu Shepherd: [00:03:52] Yeah, it’s a very interesting question, because ease of use is a relative term. If you’re a programmer and know how to do things and ease of use is easy to program from a language point of view. In the case of a normal person, is it easy to use in the level of knowledge that you have for the job you’re doing? Like, if I know how to weld, can I use the robot to weld? And the language and the programming that takes place in, is it a language and terms and setups that I’m familiar with, that I didn’t have to go to additional school to learn how to use the robot, or in the case of a machine tool, is it programmed through the machine tools operating panel, and I don’t have to pick up the robot, teach it, and it is programs like the machine tool. That’s the beginning of ease of use. But even with that, you have to be at least knowledgeable enough about how to weld or how to run the CNC machine. True ease of use is still a ways off, but true ease of use is literally the robot and the person being able to work together with hand gestures or language or things without any specialized programming.

Stu Shepherd: [00:04:54] We’re still a long way from that. So, again, start with something simple that you’re very familiar with. If you’re doing palletizing, start with palletizing-specific software and language and software, and the same thing with the other applications. And don’t try to do a moonshot right out of the gate, but start with something that’s proven, that you’ve got clear measures of what success and failure look like and how to avoid the failures. And then it’s not just the ease of use, but it’s also the ease of maintenance. You know, what happens when things go wrong? What happens when it stops? How do I get it back in the business again? That’s part of that ease of use, too, because you don’t want to have a situation where your brand new piece of automation stops, and you’re on the telephone or on email or whatever, trying to get somebody’s attention to come help you out. You need to be able to get it back in business, because every second that it’s down, it’s not productive. And if it’s not productive, it’s not making money.

Jimmy Carroll: [00:05:46] And that’s interesting too because in that same clip that I saw of you, I believe it was an A3 clip maybe from Automate or something, but you also talked about the importance of a risk assessment and then really going through and understanding what the specifications are, right? Because like you mentioned humanoids, but another kind of buzzword or inevitable topic that I always have to bring up is AI, right? So, often you hear about companies just wanting to . . . I’ve legitimately heard companies that have, from C-suite,a mandate to deploy AI. It’s like, well, for what? So, can you talk to me a little bit more about the risk assessment and ensuring that your application is actually going to be a fit for AI or automation in general?

Stu Shepherd: [00:06:28] Well, it’s a good point because just because somebody wants AI doesn’t mean that AI is the right thing to use. You know, it may not be the right use case, or it may add too much expense because part of the thing about AI is you may have to add additional processors and software that you didn’t have to have before that takes the cost of automation up. In some cases, there’s automation out there. It’s simple enough. You don’t need all the AI stuff. It’s nice to have, but do you really need it? Not necessarily. So, there’s two pieces. The first thing is to write a very detailed sequence of operation. What is this job doing, and what are you trying to automate? What parts can be automated? What parts need to still be manual operations? So, it’s literally writing a list. You know, I pick up this part out of this barrel, I’m going to this machine, I’m doing X with that machine, or whatever and write it out in detail. This is what I’m trying to do, so that you can explain to the supplier this is what this job is. The risk assessment part of it. I recommend doing the risk assessment three times. The first time is to look at the general application and say, what’s dangerous, what’s happening in this application? Are there things like abutments on the floor or other things that I have to worry about tripping over, even some of the simplest stuff? Does the machine produce chips or sparks or hazardous chemicals or whatever? What are the things that a person is being exposed to that when you do the automation, you want to mitigate that and get that out of the way.

Stu Shepherd: [00:07:49] So, you take those two specifications, the sequence of operation and the first pass of the risk mitigation, and maybe even some injury data, you know, did you have carpal tunnel syndrome issues or other injuries that happen to occur in that particular area? We’ll document all that. Share that with the integrator and say, I need the job, I need the robot or the system to do this, this sequence of operations, and these are the hazards I need to avoid. And then likewise, the automation person should also be taking a look at, if they put a robot or other mechanized device in, what do you do to protect that system so that when people are working with it, there’s no danger to them for flying debris or pinch points or collisions or whatever else. The second time that you do the risk assessment after those kind of first two passes is when you actually deploy it and get it on the floor, and you look for the things you missed, and you tune that in a little bit because you may not have caught everything the first time around.

Stu Shepherd: [00:08:47] And that’s also an important thing to do to make sure your operators are safe. They’re familiar with what’s going on. You do have to test the safety equipment periodically, whether it’s once a month or whatever, to make sure it’s still working. Just like you don’t want to find out the brakes are missing from your car the first time you put on the brakes, so you need to check to make sure it’s okay. And then after you’ve got some runtime experience, what are the things people learned as they’re running it that they didn’t think about when they first started doing it? And how do you continue to make that safe? So, safety is kind of an ongoing thing, and some people don’t like to talk about it. But frankly, as you have this continuous improvement scenario, you can also continuously improve the performance of the system, make your operators safer, and then you’re that much smarter the next time you do the next automation job. So, it becomes much simpler going forward. And it doesn’t matter if it’s a robot, a CNC device, an AI-enabled device, a humanoid . . . It doesn’t matter. You still need to start with what is the job in detail so you have your expectation set, and then how do you make that safe through the risk assessment? Those are the first two steps for every automation project, regardless of the form factor of the automation.

Jimmy Carroll: [00:09:56] There’s a lot of different questions I want to ask based on that excellent answer. But one thing I do want to ask is earlier you mentioned the warehouse, and I’m just thinking about a risk assessment for the warehouse, because if you talk about injuries, potential injuries, there’s a lot of jobs in the warehouse that, not just injuries, but they’re, it’s a lot of work. It’s a lot of work.

Stu Shepherd: [00:10:15] A lot of heavy lifting.

Jimmy Carroll: [00:10:16] Yeah, a lot of heavy lifting. So, I guess my question is lights-out warehousing or how close are we for that? And, and with lights out, where do you, how do you, what’s your take on right now, lights-out manufacturing. Are we seeing it? Is it prevalent? If not, what will it take to get there? And then kind of same question for lights-out warehousing.

Stu Shepherd: [00:10:37] So, we’re closer to lights out than we’ve ever been before. But lights out is a bit of a myth because there still takes people for intervention. They’re still unexpected details, but the more consistent and well-planned the warehouse is, the more likely you can have a very high degree of automation. And there are some warehouses, I would argue, are very close to lights out. But there may be certain things that are still in development that haven’t been deployed yet. And I think it’s safe to say that it’ll be rare to see high-performance warehouses in the future that are not built with already having robots and all that kind of stuff planned inside. And what that also means is warehouses are being designed for automation. One of the challenges of the old-school warehouses that are already out there is they weren’t really designed with automation in mind. So, you’re retrofitting, and we did the same thing in the automobile industry back in the ’70s. The cars of the ’70s, while some of them I love, and even the cars in the ’60s, they weren’t designed for automation. And you were lucky if you could figure out how to put 70 robots in the body shop. Today, it’s not uncommon to see 8 or 900 robots in the car shop, but the car has been designed for automation.

Stu Shepherd: [00:11:46] The plant’s been designed for automation. The layouts have been set up to be able to have AMRs versus fork trucks, and racking systems are designed with interchange and I/O and these kinds of things. So, as we have learned and we’ve evolved and we’ve figured out the gaps, then you’re going to see a higher and higher degree. But true lights out is still tough because there’s variables. You know, what happens when your boxes get wet because you had a leaky box truck? What happens if a box tears? There’s all these different things that can happen in a warehouse scenario. And the product you’re handling is very complex. You know, when you look at fulfillment centers like for  CVS, Amazon, all these other companies, you look at the form factors of all the different things they’re trying to handle. And then a manufacturer changes something. You know, how easy is it for them to have their system adapt to that change that was caused by the supplier? Or if the demand for that product is significantly higher, did they plan the right surge capacity and the right locations in the warehouse to handle those surges, whether they’re seasonal or not?

Stu Shepherd: [00:12:50] Those are all lessons learned that you don’t know until you fire up a big warehouse and get it running. So, the advantage of the big operators in the world, like the Walmarts and Amazons and these kind of guys, they have so much experience and so much data that they can now do that work because AI doesn’t work without data. But once you have the data, you can analyze it using AI, and you can improve the process. You can improve the automation, you can make it more adaptable, but it takes large language models, large data models, large experience use cases to actually generate that. But the other side of that is for the small- to mid-size manufacturer, can you get lights out? Kind of depends on the application. But is lights out really the right answer? Is that the most economical approach to getting it done? So, in some cases, I encourage people as an example to make it so that it’s human enabled during the day for the weird stuff that you got to do. But the standard stuff that can run automatically, run it on second or third shift. Turn the lights off for second or third shift.

Jimmy Carroll: [00:13:51] Yeah, yeah. You almost read my mind a little bit there because you’re talking about product variability and the ability to handle whatever it might be–flatpacks or boxes or whatever. And they all come into my house on a weekly basis. My wife is a, you know, she keeps them all in business. But does AI, I mean, you could say, I suppose, that AI adds a new layer of flexibility. I know it’s like kind of marketing hype, but it does, right?

Stu Shepherd: [00:14:17] It does.

Jimmy Carroll: [00:14:18] In terms of being able to handle those . . .

Stu Shepherd: [00:14:20] And we’re seeing real use cases where AI is really improving the performance of systems, especially in the vision world, is probably where we’ve got the most traction with AI. You can take these large image models of these very complex situations and dissect all these images at an extremely high rate of speed with these new processors, and be able to tell, is the fruit ripe or not? You know, is it the right size? You know, does it go to this particular supplier or that particular supplier because of grading? There’s all kinds of fantastic things that are AI enabled, but there’s other ones that maybe AI hasn’t quite gotten to yet. But over time, as we collect the data, and we find what the gaps are, you can start incrementally making things better. Now, what AI does is allows you to program faster and make corrections faster. It may not solve the problem the first crack, but it makes it faster and easier for you to replicate what’s gone wrong and then get a solution going until you get to the point where you can do digital twins. Digital twins is like the plateau. That’s the top end of the scale. If you can truly simulate everything with a high degree of accuracy, yeah, your go-to market and speed of getting automation done–significantly faster.

Jimmy Carroll: [00:15:28] We mentioned it, and I have to circle back. And then it’s kind of a part two to this question too. But humanoids. right. What’s your take on them? We talked briefly before this a little bit about where they are, where they might be going. But where do you see them finding any traction? Are they finding any traction now? Any predictions there?

Stu Shepherd: [00:15:49] I wouldn’t call them predictions. Just the observations I have is there’s been a couple of successful deployments on a big scale. So, Digit was deployed in a warehouse and doing its job, and it has a use case. There were some robots that were deployed at BMW that ran for a period of time, and there was a lot of lessons learned, and they knew that those applications were going to be short term, because you need to take those lessons learned, bring them back in house, make the next generation, and get back out there. But there’s a lot of hype. There’s a lot of overpromises and under delivery of some of the solutions that are out there. It’s a bold experiment, and somebody is going to make it work. But right now it’s pretty early because it’s one thing to say you’ve got software-enabled AI. Physical AI is really hard because just because the robot has the intelligence to do something, doesn’t mean that the grippers, the end effectors, the other things that they have to work with are ready for automation or capable. And a lot of people hope that humanoids can be general purpose. The fact is, the more general purpose it is, the more expensive it becomes. In some cases, you want to simplify and get a very case specific, very singular application to be cost-effective. And then there’s a question of reliability, durability, and safety, and these kinds of things on the safety front to the risk assessment. You know, you need to ask the manufacturer of the humanoids what happens when you drop the power on this thing?

Stu Shepherd: [00:17:16] Does it fall over on Granny? And how much does it weigh? Does it crush Granny or crush Granny’s cat? You know, that’s a real live issue. So, we’re creating standards now for humanoids on dynamic stability. We’re doing interoperabilities. What happens when it walks to an elevator? How is the elevator know it’s a robot. How does the robot know what to do as far as pushing buttons, or should it do? Because with the software that’s available, the camera systems and the elevator should be able to sense it’s a humanoid. Does it need to push the button or not, have that extra function? Who knows? So, you know what is the use case? And is that the most economical thing because no matter how great the idea and how great the humanoid is, can you afford it? You know, does it make good business sense? The greatest ideas in the world? Great. But not everybody can afford it. So, what’s that scale? And not everybody can afford a car. So, can they afford a self-driving car? Well, if they can’t afford a car, probably not. Can they afford a home? Well, you can’t have a robot in a home if you don’t have a home. And is the home designed for automation? Well, maybe not. So, how much money do you spend on the house to make it ready for automation? And oh, by the way, buy a humanoid. It’s not easy.

Jimmy Carroll: [00:18:33] No, I mean, it’s a really good point. I have three little boys, so if I had a humanoid, I would say within 90 seconds that something would go wrong, you know? So, and a dog.

Stu Shepherd: [00:18:44] And you see some great videos, there’s some fantastic stuff, but you don’t know what they went through to create the video. You look at some of that equipment from a safety perspective. There’s pinch points. There’s all these things that you don’t want to have in your equipment, around your cat or your kids or whatever else, but it’s a journey. You know, when Joseph Engelberger and his colleagues deployed the first robot in 1961, it’s not anything like the robots we’re doing today. So, you have to start and to that end, I applaud the work that’s going on. And I do think there’s going to be some spin-offs from the humanoid research to help humankind in another way. You know, kind of like the NASA program generated Velcro and a whole bunch of other stuff. I think humanoid technology will help us with prosthetics and other devices that are useful, that maybe are not a full humanoid, but an arm, a leg, or, you know, other human augmentation. And that’s pretty exciting because there’s a lot of people in the world that need that kind of assistance, and it’s more cost-effective.

Jimmy Carroll: [00:19:47] That’s really interesting. Yeah, I hadn’t thought about that. I want to come back to a question on physical AI, but the first, the part two of that question for me was kind of the opposite of my first question is like, what are you most excited about? And what are the things that you see a lot, maybe that are overhyped, that you say that, you know, be careful.

Stu Shepherd: [00:20:06] The biggest thing is setting expectations and having validation that things work. So, part of the, as we talked about with risk assessments and sequence of operation, what do you expect the robot to do? And have a very clear specification that it has to do these things in order to purchase and acquire the product. Because if you don’t have clear expectations, and you don’t have specifications for what you’re trying to buy, don’t expect them to just randomly fulfill that. You know, they’re going to do what they can. Then the other question is, was the robot designed for what you’re trying to get done? As an example, if it’s going into a food or hygienic-type situation, is that robot designed and form factored, designed to be around food and chemicals or pharmaceuticals, those kinds of things without causing cross-contamination or creating bacterias or creating a listeria outbreak or whatever? So, you’ve got to be cautious about, is it the right application at that point in time, or do you dial back your expectations as something that’s more pragmatic, more reasonable? Because there’s a lot of hype, but some of the hype is for attracting investors and what the future can look like. And that’s fine. But at the end of the day, it’s got to work. And it’s got to work well because otherwise what you don’t want to do is take this expensive device, buy it, and then park it in the corner because it’s not working right.

Jimmy Carroll: [00:21:31] Yeah, totally. Yeah, back to physical AI. I wanted to ask you about that. It’s a term that I’m seeing more of these days, and I have my own internal definition of what that means. But what does it mean to you? And how would a company go about assessing whether or not some sort of physical AI solution would be for them?

Stu Shepherd: [00:21:48] My opinion is, physical AI is a new marketing term to help companies differentiate what they’re doing compared to other traditional application solutions. So, physical AI is the physical embodiment of an AI-enabled application. It’s really applications with the AI twist. So, the application has to work first and then the AI has to make that application work better. One of the litmus tests to find out is it really AI or not is it just fill-in-the-blank menus that you’re doing, or is there actually intelligence and decision making that’s happening inside the AI code? And how does that cause the robot or the automatic device to work? By the way, it doesn’t have to be a robot to be an AI, you know, a physical AI element. The physical AI is the bridge that goes between the computational capabilities of the robot from an AI perspective, to the flexibility of real-world physical contact with tables and chairs or food or people or whatever in such a way that it allows the application to work. So, it’s one thing to have a robot that is AI, but it’s another thing to have that robot actually do something physically. And how does it make that happen? How does it work? That’s where physical AI starts to come in. So, that’s where you have to start looking at fingers, designs, grippers, and effectors in general, whether it’s welding or mechanical handling or that kind of thing. But also, you know, vision systems and other things. How does all that come together to cause the application to actually be done? Because Dancing Robot may sell tickets for a dance, but it doesn’t necessarily solve the problem that the customer is trying to get done.

Jimmy Carroll: [00:23:35] What are some examples? I mean, you mentioned food handling, but things like painting, grinding, polishing, you know.

Stu Shepherd: [00:23:43] So, those are tough applications. Any physical contact application where you’re producing a finish or changing the quality of something, there’s a certain degree of subjectivity. So, the challenge is, are there other devices that can be done that produce that degree of quality better, faster, easier than a humanoid? But AI can’t help because, as an example, like we were talking about earlier with being able to use large image models, large language models, when you’re polishing, there’s a whole bunch of data that you get. There’s feeds and speeds and heats, aware of the media that you’re grinding or polishing with, there’s these things. But then there’s also the finish, and you can use all that data and now calculate and predict application performance. But at the end of the day, you still have to have a person take a look at it. Unless your AI-enabled vision system has been trained well enough that it can look and see where the surface flaws are, and then likewise, if it does detect a surface flaw, can it go back and polish that surface flaw. That’s true intelligence and decision making. And there are applications where that’s happening. There’s automobile finishes and a paint shop where that’s happening. There’s people polishing knee joints and other femoral emplacements and these kinds of things. Those applications are starting to emerge out there, but they’re on the high end. They’re like leading edge. And the question is, is that in the economics range of where people can can afford it to get things done?

Stu Shepherd: [00:25:11] And then there’s other things that aren’t even that . . . Paint finishing is another one. I think AI and paint finishing will come together pretty quickly. There’s a number of applications that will be faster to adopt than others, but there’s other simple ones that may work you know just fine sooner than later. But the key is the predictability and understanding how does the system run? Because all systems degrade over time, and AI can help you predict that rate of degradation and either call for help–I need maintenance–or change the adjustment. I need to grind a little bit longer or, you know, handle a little bit faster. And then we’re also seeing, I think, really good applications and path performance. As robots or devices are moving, AI can sense that path, that motion, and say, am I using my mechanical joints the best way possible? Or is it just the way I’ve been programmed? And some programmers can actually cause damage to the product, just like some people drive their cars and are really hard on brakes and really hard on suspension because they don’t really know how to drive it. Well, the AI aspect of driving a robot allows that robot to have a higher degree of performance, faster speed, optimized over time by using AI to tune the performance of the path and make it better. So, you’ll see faster cycle times, you’ll see more predictable outcomes, and therefore faster paybacks on the system.

Jimmy Carroll: [00:26:39] That’s really interesting. I want to ask a lot more questions about that, but I want to be respectful of your time, too. And I know this based on talking to you now and before that, that you are a pragmatic man, but do you have any predictions or maybe not predictions? If you don’t want to go there, but hopes, things that you might see in the next one to three years in the automation space?

Stu Shepherd: [00:27:00] Yeah, one to three years is pretty narrow, because if it’s one to three years, somebody’s already got it incubating in the lab now, and they’ve already got it going. I think in ten years’ time, we’ll see a lot of very well-developed, experienced applications because the rate of change has gotten much faster. We’re still short on end effectors, we’re still short on actuators and some other things. But there’s people making great progress in those areas. And as they experiment and fail, in some cases, they’re learning from that, and they’ll continuously get better. But just because it works doesn’t mean it’s commercially ready. So, once you get the commercial functions down, now you have to figure out how to manufacture. It’s one thing to design a robot. It’s a whole other thing to design a robot company and deliver that product. So, there’s a lot of business development work, and there’s got to be a lot of work with end users that are early adopters to create that flow and get some successes out there that people can trust and say, yep, I’ve seen it work. It works like I need it to work, and I can trust the outcome. And my hope is, if I want to stretch it out even further, I’m 65 and not a young guy, but I’m hoping in my lifetime that there are humanoids that can help me when I need it as I get older.

Stu Shepherd: [00:28:17] That’s a stretch because humans are complex. Working with humans is even more complex, but I’m hoping that through the work that’s going on today, there’s things we can do for the elderly because the elderly are around us all the time, and that’s a constantly growing and changing group. In the last several years of your life, not easy. You know, my parents have passed, and we’ve learned a lot from our parents and their passing that if we could make their life easier for them and better, better quality of life, then that would be great. And frankly, Joseph Engelberger, who I received the award named after him, he and his daughter Gay, who’s running the foundation now, one of their primary roles in the marketplace is trying to figure out how to make automation useful for making human beings have healthier lives, whether it be in drug delivery or other things. And I think that’s an area that has been underfunded, and it’s a massive market because we’re all getting older, and the population continues to grow. So, there’s the total addressable market is there. The question is, how do you address it and monetize it in such a way that everybody is happy, and it’s low risk? There’s not a safety issue.

Jimmy Carroll: [00:29:27] I really love that answer just from a human perspective, not just, you know, in the business side of things, like, I hope you’re right. And all these companies that are out there producing these products, I hope they’re right too, you know? Anything else we haven’t talked about that you want to get out there? Where can people go to learn more about Shepherd Solutions?

Stu Shepherd: [00:29:45] Well, you can contact us directly at shepherdsolutionsinc.com. And you’ll find that the website’s, you’ll see our addresses. The best way to get us is through LinkedIn, actually. But also in general, not just plugging my company, but A3 is the voice of automation for our marketplace. It’s the trade association. A3, by the way, is the Association for Advancing Automation if you didn’t know that. It’s represented, embodied in 1,500 member companies that are really good at this type of work. And we’re all out there trying to make the industry better, all trying to grow it, all trying to create standards, and implement automation. There’s a lot of expertise within the organization. It’s kind of one-stop shopping, frankly, for a lot of people in our industry. And it’s a very approachable organization. There’s a lot of fantastic people that work in this group, and I’ve learned a lot from them. And every time I think I’ve learned something, I find out how much more I don’t know. And I find people inside of A3 that can help me figure out the rest of it.

Jimmy Carroll: [00:30:45] Absolutely, yeah. I couldn’t agree more. Anytime we work with a company that’s not a part of A3, one of the first things I do is recommend, you should join A3, and you should come to this event.

Stu Shepherd: [00:30:52] Yeah, absolutely. So, we’re at A3 Business Forum in Orlando, Florida. Six-hundred and fifty of some of the smartest people on earth in robotics business are here, as well as users, that kind of thing. And you know, the website for Automate or for A3 may not be intuitively obvious, but it’s automate.org is their website, and there’s just tons of resources in there, as well as a directory of all the members like our company and a lot of information on how to get started. So, you know, I think the parting shot would be is, if you haven’t automated, the best time to start is now. It’s kind of like planting a tree. The best time to plant a tree was 40 years ago. But the best time, the second best time is to do it now. So, same thing with automation. Start simple, get coaching, get people that have done it before to work with you and take very small pragmatic steps and jump in. The water’s fine.

Jimmy Carroll: [00:31:43] I love it. Perfect. Absolutely perfect close. Stu, thank you so much for taking the time. I really appreciate it. If anybody has questions, I would encourage you to reach out to Stu on LinkedIn or we’d be happy to field them for you. You can find us also on LinkedIn or at manufacturing-matters.com, and we really appreciate you tuning in. Thanks, everybody.