Rethinking the course to manufacturing's future with the Software Defined Factory (SDF) as the North Star

In this episode of People B4 Machines, Amanda Cupido speaks with Prasad Satyavolu, Accenture's Americas Lead for Industry X Digital Manufacturing and Operations, about the transformative potential of the Software Defined Factory (SDF). Prasad explores how SDFs, powered by digital twins, physical AI, and data-driven orchestration, can redefine manufacturing processes while addressing cultural and technical challenges. He shares insights on balancing human augmentation with intelligent automation, fostering workforce transformation, and creating a roadmap for future growth. Tune in for a forward-thinking discussion on how manufacturers can embrace innovation to stay competitive in a rapidly evolving industry. 

For more bold questions and sharp insights, visit www.peopleb4machines.com. Remember, the future isn’t fully automated—it’s people-powered.

Transcript

Prasad Satyavolu: 0:00

What is the problem that you're trying to solve for? And which I equate to pain. So what is the pain you're trying to solve for, which is your existing plot? But I also introduce one more dimension when you are estimating for technology, and that's what I call it growth. That means you just cannot solve for pain. Those are your aspirants that you need to take. Right. But then you have to think ahead to say, hey, what are my vitamins that I need for my future growth? And that is important because that's when you can see the future, that in future I may make different types of products. And in future, my product portfolio may increase fivefold. Or I may be, if today I'm making 1000 SKUs, I may make 50,000 SKUs. So how do I look balance the two? And what is my overall business strategy for doing that?

Amanda Cupido: 1:05

People B4 Machines.

Prasad Satyavolu: 1:07

Conversations on the chaos of factory automation from Monday morning to the very near future.

Amanda Cupido: 1:15

Welcome to People B4 Machines, conversations on the chaos of factory automation from Monday morning to the very near future. I'm Amanda Cupido, a speaker, author, and entrepreneur with a passion for the intersection of technology and humanity. Today we're talking about the future of manufacturing and specifically the software-defined factory, or SDF. This is where essential software controls machines, processes, and entire workflows. This can be a North Star for many, but we'll talk about how to get there with today's guest, Prasad Satyavolu. He is the America's lead for Industry X digital manufacturing and operations at Accenture. Prasad has over 30 years of experience in the manufacturing industry, specializing in digital transformation, innovation, and operations. He also has a proven track record of incubating and scaling new offerings, solutions, and businesses in the areas of industrial Internet of Things, data monetization, and artificial intelligence. Prasad holds a life cycle assessment certification from MIT Professional Education and sits on the board for CESME, the Smart Manufacturing Institute. Welcome.

Prasad Satyavolu: 2:30

Thank you, Amanda. Glad to be here.

Amanda Cupido: 2:32

All right, let's dive in. Right now, would you say the physical factory is still the heart of manufacturing?

Prasad Satyavolu: 2:40

It will always be. It'll never go away. After all, we're making things. And you need raw material, the raw material to be processed, parts to be processed, and made it to end products which end up on our desks, in our homes, or we use it in industrial environments. So physical factory will, was, and shall always remain the heart of manufacturing and heart of our society. It is the economic value addition transformation that is happening. So that'll remain.

Amanda Cupido: 3:15

Those are big words.

Prasad Satyavolu: 3:16

They are. It'll remain the core of who we are as humanity because in society we are we are rethinking how we are going to use materials, how we are going to move from one point to the other, how we are going to be served in the future in a restaurant or whatever. We are all about experiences. And for having those experiences, physical spaces will need to be created in a way. And for creating physical spaces, we need physical things. Therefore, they need to be produced in a manner that are creating truly that experience. So how will that production process work better, faster, cheaper? And in a way that is producing high quality. That is at the heart of this discussion to me is how can we help manufacturers, companies that are producing different goods, can create high-quality products with competitiveness so that they can withstand the global competition, be differentiated, and ultimately these products are delivering experiences that we, you and I, desire in society.

Amanda Cupido: 4:36

Yeah. Well, okay, so you talk about raw material there, and I'm just curious, some say that data is the new raw material. What do you say to that?

Prasad Satyavolu: 4:46

Is a raw material that will help products evolve or create better experiences. So how does this work? So if you take a factory today, you have a lot going on. For example, let me start with parts. So in a factory, you are not producing, most factories don't, of course, produce everything that is required to produce parts. You're making a car, you're making a body panel, requires sheets of steel, which are coming from a steel mill, uh, which needs to be received, then go to stamping and from stamping, welding to painting, and from painting to final assembly. So a lot of movement of these parts. And with that, each movement, you need to make determination of when do I do what? And how do I optimize my workflows? How do I optimize my labor on the shot road? So there is already a lot of software within the factory. So we have level zero to level four systems which are providing hardware integration, um, actual machines, equipment, PLC level, all the way to ERP and supply chain systems that are processing information, taking data in, and creating more data because you need to track the products, you need to know the batches, etc. Now the question is how do you make all the robots, conveyors, automated material handling systems, folkless? All of these work in unison as an orchestra. And the number of products that we are introducing into the line is getting more and more complex. Larger number of products we are introducing. So this leads to some very interesting uh dimensions of how to use software. For example, how do you create twins of physical twins of the factories to have all these pieces work together nicely in an environment? And not just creating those twins, now how do you orchestrate them? How do you plan for such a factory? All of these physical orchestration now falls under what we call it as physical AI. And the whole science and a little bit art of physical AI is part of this new wave of what is going to help uh define the software or create the software-defined factories. Now, if you step back and see what is really happening as atoms are moving in the physical space in the form of parts through various mechanisms, uh, conveyors and others, and being handled with robots or human beings, data is getting generated and produced. So you are tracking that data and applying intelligence to that data. That is what is uh the next currency. So it truly is a enabling raw material in the next generation of how uh and the future of manufacturing.

Amanda Cupido: 8:01

Great answer. I want to pick on the tactic of creating a twin. Do you think that that's one of the tactics that can help legacy factories become software-defined? Aren't there other tactics that you'd recommend for these legacy factories?

Prasad Satyavolu: 8:15

Yes, I do. Uh I'll tell you the biggest challenge in creating twins. Um, it was, and we are we are solving for it in very interesting ways, I should say. So the problem is in order to create a digital twin, you need a 2D or 3D drawing at a minimum of the plant and equipment that that you already have. Now you have thousands of pieces of equipment already there in the plant. And you are looking at, oh, where are my critical elements? Uh critical machines, which are potential bottlenecks or capacity constraints. You definitely want digital twins of those to be built. But uh then you're looking at an existing plant and you are saying, ah, this is too much effort to build a digital twin of an existing plant. So there were those barriers. So one substitute was, which got very popular as a digital twin, is to take a schematic or a 2D drawing or a 3D drawing, and you are instrumenting many of these equipments with sensors and PLC data, and then using industrial Internet of Things uh core concepts of sensing, collecting, analyzing, and presenting those data. Uh, we were kind of building digital twins that were uh we'll give you on a schematic, okay, this particular wall is malfunctioning and create a predictive alert on that. And so that you can dispatch a maintenance crew in advance, take care of that, etc., etc. And now we have taken it to a whole new level of using techniques like Gaussian splatting, uh, which allows you to take videos of uh existing plants. It's for a half a million square feet facility, it's a half-day uh exercise. You can fly a drone or you can walk the plant uh with your iPhone and you can record videos. And voila, you use those videos to create digital twins of a star. Now, they may not have very high fidelity, but they are far superior to what we were doing earlier. So this has a leaf fraught, and this is one of the key steps that'll enable creation of digital twins for existing facilities. And that'll help solve many other problems that we have and we couldn't which we couldn't solve with effectiveness earlier.

Amanda Cupido: 10:53

All right. So we're talking about the technical elements here, but I think it's important to also talk about the cultural side of things that's holding plants back from becoming a software-defined factory. So so why don't you speak to that a bit?

Prasad Satyavolu: 11:06

I think um you touched upon something that is the most critical element. You and I started with the question of will there be a need for a physical factory? You know, people, culture are an essential part of that transformation. And again, it will always be that way. Even if it is a fully autonomous factory, uh, we will still have human beings to monitor it and make sure that many of the AIs that are running in the factory are are ethically working unbiased way, producing things the right way, etc. Cultural transformation has three key aspects. Number one is how do you make sure that the concept of human augmentation is understood by people who are working in the plants? That means any new technology, it could be a physical automation or an AI that is working for a business process, or together a convergent solution with uh a physical automation, with an automated order processing or a bad change process comes together and executes a business process. So people at the supervisory level, people and the plant manager level, or the tugger who's bringing material or stocking the supermarkets, they understand this and that it is about human augmentation. And that's my personal philosophy that we should use technology for human augmentation. So that's one. Creating awareness, education of how this new exciting word of convergence can help them. The second thing is allowing them to be part of the design of those solutions and users of those as the design develops. I have seen adoption becomes a big challenge in factories. Anytime you want to introduce automation, even if it is a physical automation, people resist. Why? Because it suddenly creates a learning gap that, hey, I don't understand it, therefore I can't operate it. So you need to do training, you need to make them understand that this is beneficial for them. But making them part of the design process of the solution is a critical element of that cultural change. And then finally, when the solution is ready, how do you make sure that A, people who are not just part of the solution are adopting and using it in their day-to-day operations minute by minute, as the need may be, and they become ambassadors of that change. That is what will create a bigger impact on the organization, the scaling impact. So that means if we did a small proof of concept in a small cell in a particular manufacturing environment line, how do I make sure that that cell has adopted it? Well, has gotten a business result, and now is ready to scale and is ready to share with the rest of the teams in the factory to ensure that everybody is actually excited to use it and becomes an inspiration and creates a little bit of internal competition. So those are the dimensions of culture change that I feel are essential, particularly in this era when there is huge fear that automation and software defined is actually going to take away jobs.

Amanda Cupido: 14:52

Well said. And you know, this is a theme that comes up in this entire series where we're the disconnect between the top floor and the shop floor, and how are we getting teams on board? And so it makes sense that it comes up here too with you. But I'll also encourage listeners to check out we have a full episode dedicated on that cultural shift with Jeff Bernstein. So feel free to check that out also in this series. But I want to bring it back to the technical side with you, Prasad. So historically, information technology and operational technology have operated in separate silos and then having different priorities, risk profiles, et cetera. Do you think executives are equipped to lead in a world where IT and OT are inseparable?

Prasad Satyavolu: 15:33

They are. Convergence is no longer between just IT and OT data. It's about the total environment that you have because IT systems have been straitjacketed into hey, you are above L2 level, and then that's what it is. And OT systems are remaining at L0, L1 level. So those classifications vary based on the plant and who took control of the plant as soon as a plant was launched. So it's not just about IT, it's about ITOT, PT, and ET, as I call it. So PT is product technology. And uh what happens is the ability to produce different products on the same line in a flexible, agile manner is going to be the true differentiator for the future factories. You may build a factory today for producing a certain type of product, but if you are not able to pivot that factory for the future demand of newer products at the least cost possible, then you will be not competitive or you will become incompetitive. So I think that competitiveness is critical in this case. And I expanded your question from just ITOT convergence. So product technology is important. Why environment is important? The technology that captures the environment, the physical environment, is also going to be equally critical. There are what is the temperature, humidity related parameters, many of them affect quality. Your ability to bring all of that data and the data around orders and what's going on in your supply chain to make it work in convergence for the factory optimization is going to be the key next. So it's it's an exciting journey. Many organizations are embarking on it as we speak.

Amanda Cupido: 19:03

I love that you've put these new acronyms out there. And it feels strange to think of a world where we're not even using ITOT anymore, but this is part of that mind shift cultural change that people have to start imagining.

Prasad Satyavolu: 19:16

Absolutely. Absolutely.

Amanda Cupido: 19:18

And as we talk about the potential that this technology has and how we're gonna have to just really rethink so much, do you think that manufacturers on average are underestimating or overestimating the potential that the tech and software can bring to help redefining production as a whole?

Prasad Satyavolu: 19:35

The first issue is like estimating itself. Because what happens is you are you are in day-to-day, you are in operations, you feel with brute force execution you are able to uh manage the plant. And I don't blame them. Plant managers are responsible for getting product out the door at the right quality and at uh right cost. And that's what they are accountable for, that's what they are held, I mean they are answerable for. So what happens is their focus shifts to producing those products. But if you really look at the dimensions of how do you do workforce transformation, how do you build the future for the factory and plan, it becomes a critical next level of thinking. Sometimes you're looking at organizations with multiple plots. You need to look at the network. That how do I manage my network of factories effectively and efficiently? And the related dimension of that, which um I wanted to mention was how important is the workforce transformation in this case? So every plant manager, and one of our studies revealed this that almost 70% plus managers operating the plants felt that they need a workforce transformation strategy going on. Uh, and from a technology standpoint, uh, you are really looking at next generation of workforce coming in. So onboarding that workforce, actually first inspiring them to join manufacturing is a task. That leads you to another dimension of how do you use the tools that this generation is used to now, gaming, virtualization, devices that allows them to carry that experience uh of having played with that into their working environment. So some very interesting uh dimensions are propping up on how do you actually work on the these dimensions of engaging the society into manufacturing.

Amanda Cupido: 21:46

And when it comes to the technology side of things, I think one of the fears is overinvesting in technology that doesn't deliver. And so what would you say to those who are hesitant around figuring out how much to invest? Because the other side is, of course, underinvesting, and then they're not keeping up with a transformation that's long overdue. So, how do you balance that?

Prasad Satyavolu: 22:10

I think you need to create a value case. And that means very where you're counting cents per unit cost of the product. It's important. If you're making aluminum cans, you're talking about 10 cents. And then you're talking about a major investment on that in technology, then you have to think twice. How much cost can I shave off of this product? So the reason I took that extreme example was of a very low cost item is to illustrate the point around value and in estimating what are your right strategies that you should put in technology investment and where should you deploy it. So the first thing is to ensure that you have a very strong value case for technology. That is your starting point. And then you look at okay, what are the key problems and challenges that you will solve as part of that value case? And what will be the elements of technology that you will need to work together to solve that problem. And this is where systems thinking also comes into play. You look at the problem more holistically and not solve a problem uh in a very narrow way, otherwise, you will end up spending a lot of money. So what we advise is there is no right or wrong answer. The answer lies in uh what's your problem? What is the problem that you're trying to solve for, and which I equate to pain. So, what is the pain you're trying to solve for, which is your existing plan? But I also introduce one more dimension when you are estimating for technology, and that's what I call it growth. That means you just cannot solve for pain. That those are your aspirants that you need to take. Right. But then you have to think ahead to say, hey, what are my vitamins that I need for my future growth? And that is important because that's when you can see the future, that in future I may make different types of products. And in future, my product portfolio may increase fivefold. Or I may be, if today I'm making 1000 SKUs, I may make 50,000 SKUs. So, how do I look balance the two? And what is my overall business strategy for doing that? So these are some of the dimensions, not all of them, which should guide the plant managers or capital program managers, incorporate or anybody running their own businesses or large corporations to see how much they should invest in technology and how should they pace it? What is the core infrastructure required? What are the modular ways in which they can build on it without throwing away the past investments? So something to think about. But there is no right or wrong answer to that. It's about what's the problem? What's my growth imperative? Where is the value? What is the right technology? What is my roadmap? How will people be? What is my future business strategy? Something to consider.

Amanda Cupido: 25:20

I love the analogy of aspirin and vitamins because I could already see people saying, well, I don't have any pain points, so we're good. But it's about looking ahead and then the vitamin mix, it's like there's so many vitamins out there. Which ones do I take? There's no one that makes you a superior being. It's really great analogy. So well, well done. Before we wrap up here, time's flying. Um, but I love to throw a curveball question at you as we talk about the future. So if you could ask one question to the future version of yourself as a manufacturing executive in 2035, so 10 years from now, what would you ask?

Prasad Satyavolu: 26:02

Wow, you had me there. But let me think this through. I I think the key would be how am I balancing the role of human beings in a plant alongside the intelligent objects which have self-learning and which are able to execute many, many tasks by themselves? And how am I balancing it? How am I able to maintain the motivation of the human beings to operate in such an environment? That'll be my question.

Amanda Cupido: 26:40

That's a good one. Okay. Well, we'll leave it there for now. Prasad, thank you so much for joining me. And thank you for tuning in to People Before Machines Conversations on the Chaos of Factory Automation Powered by Eclipse. We're here to challenge the status quo in factory automation because machines don't build factories. People do. The technical producer for this podcast is Ryan Dentinger. I'm Amanda Cupido. If you learned something in today's episode, do us a favor, share it with a teammate, a plant leader, or anyone who's tired of the automation echo chamber. Be sure to follow this podcast for real talk, bold questions, and sharp insights. And remember, the future isn't fully automated. It's people powered. Thanks for listening to People B4 Machines. Dialogue from the Eclipse Factory Automation. For past podcast episodes, search people before machines on Spotify, YouTube, or visit eclipseautomation.com.