The value of process industry AI: shatter the one-size-fits-all myth–transcript
John Nixon: Hello and welcome to today’s episode of the Industry Forward Podcast, where we explore the key trends, transformative technologies and real-world innovations that are reshaping fields from consumer-packaged goods (CPG), life sciences, energy and beyond. I’m your host, John Nixon, global vice president of Process Industries at Siemens Digital Industry Software. My guest today is Bill Hahn, he’s the director of Solutions Consulting alongside me here at Siemens Digital Industry Software. Welcome again to the show Bill!
William Hahn: Thanks, John.
John Nixon: Today, we are going to continue discussing the impact of digitalization on process industries, Bill. Specifically, we will speak about the role of [Industry AI] and how, with specialization, it affects the end users throughout a product’s lifecycle. Those users can be employees, customers, specialists, you know, really anyone that interacts with the industry in question.
Bill,what roles do you see digital twins, simulation and [Industrial AI] playing from, let’s say, the bench in the laboratory to at-scale manufacturing operations? Maybe tease out a little more on that?
William Hahn: Yeah, I think throughout the lifecycle, John, we’re going to see these technologies more-and-more at play, right? At the lab scale, we’re going to see simulation, and use of [industry AI], down to the molecular level. We talk a lot about molecule to manufacturing, molecule to market, right? How do we apply these principles all the way at the molecule level in the lab, and then be able to scale up the data, the process, and the analysis to the pilot scale? And then [again] to the full plant and commercial scale?
So, we have to start looking … throughout these different lifecycle phases. And as you know, the types of data and the tools and the technologies, they’re different as you move through the life cycle phase. What tools scientists are using in research are different than the tools we’re using in product development versus manufacturing. So, we need to make sure that A) the data can seamlessly flow through the digital thread from the lab, from R to D to M, right? If you think about research to development to manufacturing.
Not only does the data have to flow, but the tools need to enable the digital twin and the physics-based analysis, and the [industry AI] at each one of those lifecycle stages as well, right? And then understand the differences between those lifecycle stages. So, when [industry AI] looks at data in the lab, versus data in the commercial manufacturing phase, the [industry AI] needs to know the difference, right? And our digital twins need to know the difference so that we get appropriate responses when we’re trying to answer a critical question like: an issue on the shop floor, an issue in manufacturing that we’re trying to figure out what went wrong with that batch. I need to understand what was the context of what we were doing in manufacturing versus what was the research performed in the lab.
John Nixon: As you talk about that, it’s clear that there’s not just one [industry AI] for all. As you’re indicating, there’s [industry AI] for example, a design engineer. And the constant optimization of their experience may be anticipating, as part of their design needs: “what they’ll need is the next step, and the next step, and to ensure requirements, adherence and alignment.” And then there’s an [industry AI] on the manufacturing floor where we might be optimizing material flow.
So, to your point, just like humans, there’s expertise in [industry AI] along the lifecycle from stage gate, to stage gate, to stage gate. And it’s still, though, dependent upon a strong data strategy. And ensuring, as you were just indicating, that there’s this retrieval augmented generation (this RAG) enablement as part of [industry AI] in every stage along the way. And that’s coupled together with the laws of physics, right, the laws of thermodynamics, as closely mimicking the real world we’re in today. All of that comes together to provide a digital twin and a digital twin at scale.
John changes the topic
Beyond what we just talked about now, I want to give some very specific examples of what I see are benefits with comprehensive digital twins and [Industry AI]. Let me share a couple of stories for you, Bill. And I’d love for you to share some with me.
One example, just a very basic one, for example. I was in the Corps of Engineers, and I remember [being] under the 1992 Federal Facilities Compliance Act. I joined the military in ‘93. After my time as a platoon leader, I became a facilities engineer for the 1st Cavalry Division. And what was exciting was we were going through and we were, as per this compliance act, we had to go through it and look at our own [environmental] footprint, right? The military and federal facilities now all had to abide by the same EPA (Environmental Protection Agency), Department of Defense, Department of Energy, Department of Transportation, all of the requirements and regulations that existed.
And I remember in my office, I stacked up just to get a sense of what I was dealing with, all of the regulations and laws, both federal and state, that I would have to deal with. And it literally, Bill, went from the floor to the ceiling of my office! And I thought, “how in the world are we going to get through this?”
So, I spent weeks and weeks reading through and extracting from those regulations all of the requirements. We had to do things, for example, like keep our motor pools compliant. And we weren’t contaminating the environment with oil or fuel or what have you.
And so, you had all these requirements. And I won’t bury the lede. Fortunately, we were in compliance. But the way you did that was you had to go through all of these regulations. And Bill, you’ll appreciate this. I had regulations that referred to other regulations, and those regulations referred back to the regulations I just read about. So, it was this very challenging environment to pull all of that together. And I was able to boil it down to a handful of check sheets, eventually, so that we could very effectively go through, validate compliance, and that we, like private industry, were not impacting the environment in an adverse way.
Here’s the exciting part. If I had [industry AI] back then, what would have taken me weeks, I could have done in minutes to be able to extract the mission critical regulations and requirements to do my job. And that’s what’s exciting about the technology today.
I’ll give you another example, Bill, from industry. So, I at one time was working for a very large oil and gas customer. And part of my responsibility was to go out in the field as part of our asset integrity planning. And I would go out and have to inspect high pressure equipment, steam supply systems, for example. And you would be looking for things like your electrical heat trace wiring: did you have any brakes, insulation, was it solid? Looking at coatings, were we having any holidays? Looking at your ground bed installation. And looking at your sacrificial anodes and remaining usable life. Looking at crack propagation with welds potentially.
And on that last example, when you have something like crack propagation, you start to ask questions like, is this a material issue? Is it a workmanship issue? Was the weld procedure certified? Were the materials maintained correctly so that they didn’t become damaged before you actually did the weld? Did you hydrotest? Did you x-ray? Were there ultrasonics? Was there post-weld heat treatment? I mean, the list goes on and on.
And so, as you’re talking about the digital twin, as I think about data, I think about the real-world applications. everything from requirements adherence, like I was talking about from my time in the Army, to more recent days where we’re looking at asset integrity in the field. For me, a digital twin has its greatest value when I have a true forensic ability to understand everything from concept, to the moment I’m standing in front of an asset and I’m looking at it and I understand the challenges or opportunities in front of me with that asset.
What about you, Bill? Do you have an interesting anecdote of where you see value in the digital twin? I know you worked at NASA at one time. You’ve worked in oil and gas. There’s got to be something there that excites you about the digital twins and [Industry AI] of today. [Something] that you might have made use of in the past?
William Hahn: Yeah, John, I wish I had this technology in the past, honestly. I can only imagine the possibilities.
I mean, looking at my work in asset integrity. I mean, the amount of data that we had on pipelines, for example, from inline inspection data. Imagine, you can’t do a CFD (computational fluid dynamics) analysis of a 200-mile pipeline, right? It would just take an infinite amount of time. We know how computationally intensive something like CFD is.
But imagine if you could feed [industry AI] all of the inspection data, the inline inspection data from years and years of runs into an AI. And [then] have the [industry AI] look at all of that data and say, “hey, here’s five areas you should probably do a physics-based analysis to see if there are potential corrosion effects, et cetera, based on analyzing all of this data at once.”
I can only imagine how different the analysis would have been, John. As opposed to, “hey, let me just sort the spreadsheet highest to lowest and figure out where my corrosion features are today,” right? It would have been a completely different type of predictive analysis that could be completed.
John Nixon: Bill, I want to talk about utility for a moment, the utility of [industry AI]. I want to talk about its value to the individual.
- That individual might be an engineer, a design engineer, specifically an electrical design engineer.
- The individual might be somebody on a factory floor responsible for the effectivity of equipment and ensuring that it has maximum uptime and is producing at its maximum potential.
- It might be an end consumer. Perhaps we introduced a new flavor of potato chip. And so, the end user’s experience [is recorded] with that.
- And it also could be someone who’s looking at, scale, everything I just described from concept all the way to the end user experience.
As I mention that, as I talk about these different roles and the role of [industry AI] with those various personas, what comes to mind for you on the value of [industry AI] within that context for that persona?
William Hahn: I think, John, it comes down to efficiency, right? I’ll go back to my engineering days, right? As an engineer, right, you’ve been in the same boat. You don’t have all the answers yourself, right? You might have answers in your particular discipline, your area of expertise, but what do you end up doing? You pick up the phone, and you call somebody, right?
I’m in asset integrity and I need to call operations, right? Or I need to call the control center, I need to call the plant, or I need to call another engineering discipline like electrical because that’s not my area of expertise. Because the problem that’s in front of me isn’t just my problem, right? It’s a system problem and it crosses multiple boundaries that require expertise from multiple different people.
Now let’s look at this today. Imagine I had an electrical engineer as an agent. Imagine I had an operations agent. Imagine I had a plant agent. And now instead of me hoping that, John, if you’re [not] out at the plant, and can’t answer your cell phone; I don’t have to wait three hours for you to call me back. I just ask the agent what’s going on in the plant.
So, to me, it’s making all of this knowledge that we have throughout the entire organization instantly available to everyone within the organization. It’s like having 10 people standing behind you, right? Just ready to answer any question you may have to solve that problem that’s in front of you. And that’s a ton of efficiency in my mind.
John Nixon: So, Bill, as you’ve talked about these efficiencies, I start to think of efficiencies at scale. Let’s look at the end customer.
In my earlier example, you’ve introduced a new flavor of potato chip. Now, in the past, this new flavor would have been introduced. It would have gone out. And you would have had a very manual way of trying to garner what has been customer sentiment. You would be looking at sales over a quarter—maybe over a year. You’d have to have a very protracted timeline to have a feedback loop on whether that was the right decision.
Versus today, when we all carry a smartphone, and you ask for ratings through an app, or let’s say through YelpTM or your own app as the manufacturer, and they’re giving three stars, four stars, five stars. They’re also providing comments. “This was a little too spicy.” “Oh, I like the new tanginess,” things like that. You have these agents that are now calling through tens, hundreds, thousands, millions of feedback points that you can now pull. And as you were saying, you’ve got this agentic [industry AI] at every stage. You can immediately call through all of that, [to] understand what is speaking to the customer: could be culturally, could be personally, whatever it might be that’s affecting their perception of the utility of that product. And in turn, you can now re-tune your recipe so that it meets more closely, in this case, a new potato chip and a new flavor, what that looks like and is interpreted by and experienced by customers.
To me, when you talk about effectiveness and efficiency, and I think about utility. For me, [industry AI] at each stage has helped to collapse the timeline that a feedback loop will turn into something that’s value added for end customers. Do you have that same perception?
William Hahn: And I’ll take it even further and make that perception bi-directional, John. It’s also from a marketing perspective. Understanding, you know, for my product, how the users are going to be looking for that information, how the end customers are going to be looking for that information.
I’ll give you an example. So, my wife and I were recently looking for a new couch, right? How would I have done that even five years ago? You go onto GoogleTM, or AmazonTM, or IKEATM‘s website, right? And you start searching for, okay, I want a new couch. Let me go filter it down, right? Do you think, John, I’ll ask this question. Do you think that’s what I did this time?
John Nixon: I don’t know. Is it what you did this time?
William Hahn: It is not. Laughs I went on ChatGPT and I said, “ChatGPT, I’m looking to buy a new couch. I have some extra room in my living room because I reconfigured where my TV was. That gave me an extra 5 feet that I think I can actually have, like a corner couch at this point: where I can lay down on one side and sit on the other side. And I just got some input from my wife, and she says she wants to have a linen material that looks a little more professional, but we want to keep it in a given price range. We don’t want to spend $5,000, but we want it to look like a $5,000 couch. And here are some color restrictions that we’re looking for,” and so on and so forth. And I fed it all these requirements, and it gave me three options that met all of my requirements across three different websites of where I could buy this couch.
Now think about that for a second if you’re the manufacturer of those couches. Was your couch one of the ones that was proposed by the AI? What marketing materials were available online to that LLM so that your couch showed up in that search as opposed to three other competitors’ couches, right?
So, I think it’s both an inbound perspective of understanding what your customers want, “voice of customer and feeding that into the ideation process.” But also, an outbound [perspective] of “how are the customers looking for my product?” And do I have the right information out there available to them [so] they can find my product from a post-market perspective?
John Nixon: And to pick on you just a little bit, Bill. Were one of your parameters how comfortable the couch was for when you got in trouble and had to sleep on the couch?
William Hahn: It was. Laughs It had to be soft. I did not want a firm couch. I wanted to kind of fall into the couch, and I did describe that to the AI.
Both laugh.
John Nixon: Well Bill, I hope that couch worked out for you.
William Hahn: Thanks for having me on, John.
John Nixon: Today we discussed [industry AI], its value to not just the process industry but also the customers, employees and specialists that interface with that industry. We discussed how [industry AI] isn’t a one-size-fits-all tool. It must be optimized for the needs of each end user and each task. We mentioned how [industry AI] shortens the time-to-answers and links users to the information they need when they need it. We also spoke of how this ability to link information could have helped us in our past. We then talked about how embracing these [industry AI] tools can not only save money but also make it by funneling new users to your product.
I want to thank our audience for joining us today on the Industry Forward Podcast. I hope this episode helped you process your process industry knowledge! Please join us next time as we continue to explore technologies and ideas that are shaping the future of industry. Have a great day.
John Nixon – Global Vice President of Process Industries at Siemens Digital Industry Software
As Group Vice President for Process Industries at Siemens, John leads a global team that helps process industries leverage digital solutions that enhance efficiency, accelerate innovation and achieve sustainability goals.
John has over three decades of experience in strategy, operations and technology deployment for energy, chemicals, life sciences and CPG. He is well versed in the operational and business pressures of industry, including regulatory demands, decarbonization, talent gaps and the push for innovation.
William Hahn, Director of Solutions Consulting at Siemens Digital Industry Software
As Director of Solutions Consulting at Siemens, Bill leads the digital transformation initiatives for large enterprise clients across life sciences, CPG, energy, chemical and infrastructure industries.
Bill has over 18 years of experience in digital strategy development, systems engineering, and business process optimization. He specializes in end-to-end software development, and implementations, that aligning technology solutions with business objectives like risk management, compliance and operational resilience.
All trademarks are property of their respective owners.
