NDE 4.0 Podcast Transcript
Episode 22 — AI, ML, and the evolving role of the NDT inspector
Our Guest: Lennart Schulenburg – CEO, VisiConsult
Nasrin Azari: [00:00:00] Hello and welcome to floodlight Software’s NDE 4.0 podcast. In this series, we interview various experts in Industry 4.0 concepts, ideas, issues, and technologies as they relate to non-destructive testing and inspections. This show is designed to explore the biggest challenges and opportunities for the future of NDT guided by some of the smartest people in the industry.
Nasrin Azari: So be prepared for a thought provoking discussion and to learn something new in the next 30 minutes. Hope you enjoy the episode.
Nasrin Azari: Hi everyone. I’m excited to welcome Leonard Schulenberg back to the podcast. Leonard is the CEO of Zi consult, a leader in advanced [00:01:00] x-ray and NDT inspection systems. He’s also one of the trailblazers in applying AI and machine learning to the world of non-destructive testing. Guided by his leadership. ZI Consult has become a strong advocate for innovation and safety and service delivery, pushing the boundaries of what’s possible in industrial inspection.
Nasrin Azari: He’s also an active contributor to the global NDT community, frequently speaking about inspection technology, automation, ai, and the evolving role of the NDT Inspector. In today’s discussion, we’re focusing on artificial intelligence and machine learning and NDT to areas of Leonard’s expertise and topics that always benefit from ongoing exploration.
Nasrin Azari: I’m excited to have you back with us, Leonard. It’s great to speak with you again.
Lennart Schulenburg: Thanks for the nice introduction and happy to be here again.
Nasrin Azari: Awesome. So the format of this podcast, as you probably recall, is that I’ll pose five questions to you designed to dig into some of the most meaningful and interesting aspects of NDE 4.0.
Nasrin Azari: Let’s [00:02:00] jump into our first question, which is, what do you see as the main drivers of AI and machine learning adoption in NDT?
Lennart Schulenburg: Uh, this is a great question. I’m glad you’re asking because many times if you look at AI coverage in the media, uh, you’re some somehow losing track. Why? Why is this happening? It seems like, oh, this technology is cool.
Lennart Schulenburg: Look what it can do, but what is actually the driver of this happening and. And the good thing is we, we do have a driver in our industry, so we have identified, um, three drivers. We talked to many of our customers, mainly in the aerospace segment, but also automotive, and asked them, what is the main drivers for automation and artificial intelligence?
Lennart Schulenburg: Because AI is basically just automation in the digital world. Same like a robot would be in the physical world. So the answers are. Mostly always the same. It’s, it’s three things. Um, first is a lack of skilled labor, and there’s this dystopian view of robots and AI systems taking over the world, and everybody’s unemployed.
Lennart Schulenburg: Highly unlikely. We can get into that a little bit more [00:03:00] later, but what companies really struggle with in the West is finding enough NDT technicians to actually satisfy their needs. And that is a big problem because these people are. Sometimes just not there because the demographics in the west are getting worse and worse.
Lennart Schulenburg: There’s more people retiring than new people coming in. And that AI can really can help to solve by augmenting the human and, and making as making assisting them. So the second driver. It’s increasing volume of inspection that is driven by increasing production volumes, but also by more complex inspections.
Lennart Schulenburg: So if you have an additively manufactured part, it needs to be much more inspected than a casting component. Mm-hmm. So you need to inspect more. And third is we are under a lot of cost pressure, especially in the West. There’s competition from Asia. There is, um, there is, there is a race to the bottom from, from all sites.
Lennart Schulenburg: So people need to, or companies need to catch up and, and get their cost and control. And this is where AI and automation also helps. So to [00:04:00] summarize it up, the three drivers, you need to do more with, less people at a lower cost, which is really, really hard. And this is where technology can come in and help companies to solve that puzzle.
Nasrin Azari: Yeah, that’s, I mean that’s, I feel like that is really, um, a driver for AI in really any industry. Do you think that there’s anything different about NDT in that respect? I.
Lennart Schulenburg: Well, NDT? Yes. I would say NDT is a special industry because we are in a reor, in a highly regulated environment, rightly so. Uh, we do quality control in the end, and this is the most and foremost, um, um, it has to be safe.
Lennart Schulenburg: Right? Right. So whatever we introduce. Cannot make the process inherently unsafe or decrease safety of the components you manufacture. So this regulatory aspect slows down innovation slightly because you’ll always have to make sure you’re on the safe side. And then also, NDT traditionally is a very conservative industry, [00:05:00] um, because it draws conservative people to that industry.
Lennart Schulenburg: Mm-hmm. Because if, if you are somebody drawn to quality, you are most likely more conservative than on the like, innovation startup. I’ll just try this. Right. And, and, and, and that adds complexity. Um, or that adds a challenge towards new innovations and technology, uh, compared to, let’s say, a marketing innovation where people are like, yeah, I’ll try it.
Lennart Schulenburg: And if it’s, if it’s delivering a strange picture, I’m not using it. You can’t do that in n entity.
Nasrin Azari: Yeah, that’s a good point. That’s a good point. Um, for our question number two. In your experience, how are artificial intelligence and machine learning helping to reimagine NDT processes rather than just automate them?
Nasrin Azari: Like what’s the difference between AI and automation in your mind?
Lennart Schulenburg: See, you can do either way. I mean, every tool you can, you can, um, uh, you can imagine whether it’s a robot or it’s an AI [00:06:00] system, you can just put in there to replace how work has been done before. And I mean, you come out of the world of business process automatization and digitization.
Lennart Schulenburg: So you, you know, best that if you take. Crappy analog process and digitize that process. All you have is a crappy digital process, but Right, right. But, but when you want to really improve things, uh, you have to look at these new tools and have to reimagine the process. Is this process still the right process if I use these new tools, or do I have to rethink it slightly?
Lennart Schulenburg: And of course, you can always start by just implementing the tool, the tools. To augment the current process or to assist the operators, that’s the easiest step because you don’t need big changes. You don’t need to involve the process people, you don’t need to go through a requalification, but you will not unlock the full potential.
Lennart Schulenburg: Mm-hmm. So when you bring an AI system, yes, you can just, you know, put that AI system in place. It helps the operator by pointing out [00:07:00] some. Assisting him in his decision. But then really once you, once you see that, look at the outcome, then rethinking. Okay, how should that process really be? Like, should, should there be first the operator and then the ai, or first the ai and then the operator?
Lennart Schulenburg: What do we do if you have a conflicting, a conflict between the AI and the human? Does a third person, like a level three supervise that? There’s a lot of questions that can come up and that you can use as chances for process improvements.
Nasrin Azari: And then, yeah, so that makes a lot of sense. Where we see AI contributing more to better processes than simple, simply creating an pro, automating a process that already exists in terms of the machine learning aspect.
Nasrin Azari: Um, you know, that’s in, in my. View machine learning is really all about taking data, um, from a process to use it to actually improve going forward. I mean, there might be a variety of different ways that can be [00:08:00] be done. How do you see that as, uh, the machine learning aspect of helping to improve these system processes?
Lennart Schulenburg: That, that would be the greatest outcome that we could ever have if we use these, this data to produce less, um, scrap or reject. Yeah. So what we currently do is we’re working on identifying the scrap produced out of the manufacturing pro process to identify it better, more precisely, faster, cheaper. But what would be even better is to not produce it in the first place.
Lennart Schulenburg: Right? So there, there we move from the analytical side to the predictive side, what we focus on a lot on this first stage of AI implementation is the analytical side. So we are analyzing X-ray images data to actually find indications to find. Issues, um, in those parts, what we are not yet doing is deriving prediction out of predictions out of that.
Lennart Schulenburg: But we can already see by the data we generate, we suddenly [00:09:00] generate machine readable, highly structured data on our customer’s components or, well, they are doing that so they know in what area of my parts do I have hotspots. I constantly get indications. What size are these indications? What classes are those indications of?
Lennart Schulenburg: Is it a porosity, is it a crack, is it a shrinkage? And obviously if you feed all that data back into your production management system, you can derive something out there. And as long as you have fat enough data in there, you can just change something in the production process and observe a change in the quality.
Lennart Schulenburg: And at some. Point that information will help you to correlate, turn up the heat by one degree or two degree. Get more porosities or less porosities and that is what will help tremendously to improve quality down the line.
Nasrin Azari: Yeah, that’s, that’s awesome. I, I feel like there’s still, we’ve gone, we’ve made a lot of progress, but I feel like there’s still a long way to go, which in some ways is, is exciting and it will probably happen a lot [00:10:00] faster.
Nasrin Azari: The, the the, the time it takes to ha for things to happen seems to speed up, um, quite a bit these days. Um. So let me, let
Lennart Schulenburg: me, let me, let me quickly interject that because I think you’re right. Yeah, I, I mean, I’m, I’m perfectly a technology, I’m a technology enthusiast and this is going to slow for me, but it is going incredibly fast.
Lennart Schulenburg: If you look at past innovations like the digital innovation going from let’s say analog x-ray film to a digital detector mm-hmm. It took 10, 20 years. Right. And this, what we are seeing right now is happening in months to to, to years maybe. And in that is an inherent. Risk because this thing is moving so fast that people will be lost along the way.
Lennart Schulenburg: Mm-hmm. People not implementing that will not reap the benefits and other people will pull forward because of the big efficiency gains this brings. And on the other end of the spectrum, you’ll have people moving too fast mm-hmm. And implementing something prematurely, causing damage to their quality procedures.
Lennart Schulenburg: [00:11:00] So it, it is a fine balance and there’s risk on either side of the, of the we of the edge.
Nasrin Azari: Yeah, that’s really interesting. I, I 100% agree and I, I already see that people are, people in the industry are, are not entirely sure. You know, what the future is gonna look like for them and, and what the, you know, what will, what their jobs will look like.
Nasrin Azari: And also the things that you’re talking about feed really nicely into our next question, which is, uh, what do you think is not getting enough attention in the industry? What would you say the limitations of AI are in NDT that, that you think maybe the industry isn’t talking enough about?
Lennart Schulenburg: Yeah, and you’ve already mentioned it, the people, right?
Lennart Schulenburg: It’s always about the humans, it’s about the people. So, um, yes, the required skills will probably change through this technology. Technology. Same, like the skills changed when we moved from horse carriages to [00:12:00] cars. I mean, that was a. And I’m using this because everybody can, has an, has a picture in their mind right now, right?
Lennart Schulenburg: Mm-hmm. So all the people riding horse carriages, at one point they maybe became drivers of cars, taxi drivers or something else, but it required a reskilling, a retooling of the skillset of these people. And I think we are on the verge of seeing that again, just it’s that it’s much faster. It’s much more, yeah.
Lennart Schulenburg: Brutal and violent because it’s, it’s happening so fast and it’s quite complex and hard to grasp and that makes it hard. So what we have to do, we, on the, on the solution provider side, but also people at the As SNT, at the societies, at the training institutes, we need to identify what’s the new skillset of NET technicians going forward?
Lennart Schulenburg: What do they need to learn? What’s the skill gap that we currently have and how do we. Train that fast because we don’t have years. This technology is already here today and it’s, [00:13:00] it’s moving forward at a, at a really fast pace. So we’ll have to, we’ll have to implement an answer and, uh, luckily all the societies that I’m aware of, the a s and T, the German Society, the Indian Society, they are already working on training programs to address that challenge and question.
Lennart Schulenburg: So there is hope. Yeah.
Nasrin Azari: Yeah, that’s good. Um, and you know, speaking of limitations, I’m kind of thinking about what you said earlier about, um, you know, maybe technology, some technology being adopted too quickly and, and before it’s ready per se, or even, um, you know, there, there might even be some problems with bias, right?
Nasrin Azari: In some of the tools that we’re using and being able to recognize that, like having the skill. As a, either a service provider or as a user of some of these, the tools that we’re developing with ai, being able to tell that something’s not right, like it’s not [00:14:00] performing the way it should, and I feel like that is something that’s gonna be really important because as we put more and more trust and confidence in ai, it’s really important that we get it right.
Nasrin Azari: Right.
Lennart Schulenburg: You’re, you’re right. And, and the, the hard part of that is that that evaluation requires a lot of math and statistics, um, that I’m not proficient in that most people are not, either because that’s what PhDs do, or our data scientists do. This is advanced mathematics that you do to test these AI systems.
Lennart Schulenburg: And the only thing that the industry can do as mandate through strict standards and regulation is that a certain level of quality control and. Proof is provided by the suppliers of these systems. So we, for example, we are putting in significant resources in generating very extensive correlation studies, performance reports that show the quality of an AI system compared to a cohort of operators.[00:15:00]
Lennart Schulenburg: Mm-hmm. Or inspectors. And I say cohort because, um, and this is maybe an interesting thought for you listeners. Um, if we look at. Something or somebody, uh, giving, um, an inspection decision or doing an inspection de decision. That could be a human, that could be a computerized system, that could be an AI algorithm.
Lennart Schulenburg: Um, that entity is a decision system. So a human is a decision system. Human decision system, same as an algorithm decision system. Both of them make a prediction about a part and the prediction because they don’t know. The second you would cut open the part and you would look inside. You would know for sure, but as long as you are just looking at your x-ray scan or your ultrasonic scan, you’re making a prediction about reality and both the AI system and the human can be wrong.
Lennart Schulenburg: And this is what we don’t, what we don’t think of sufficiently yet, and what we are discovering when we implement AI systems. In the beginning, the results look very poorly. And then we dig into and we find out that the underlying human [00:16:00] process was flawed in the first place because there were biases in place and certain types of defects were neglected for whatever reason.
Lennart Schulenburg: Either it was a policy decision or it was a blind spot, and that, that is really, really interesting. The second we are starting to quantify things by implementing these tools, these computerized tools, we are really starting to see the. Issues that have been there for 10 years. And that’s very, very revealing for, for our customers.
Lennart Schulenburg: So what we constantly get out of these implementation projects is that, yeah, the AI is amazing, but we learn so much about our inspection process. We’re gonna start to implement all these changes. Mm-hmm. And this is really interesting ’cause nobody expected that, including us.
Nasrin Azari: Right. Oh, that’s fascinating.
Nasrin Azari: I wouldn’t have expected that either. Um, and you’re also doing a really, really good job of weaving your answers right into the next question, because the next question I have for you is, is about safety. And with the stakes [00:17:00] being high for AI based applications in NDT, how should we think about safety and accountability in AI driven inspection systems?
Lennart Schulenburg: Yes, stakes are high. Where we are, so we have currently, we have a gold standard, and the gold standard is the human. So that’s the benchmark that’s, that’s out there. Mm-hmm. And one thing I could never, never recommend as a solution provider to anybody, just take an AI system, replace all the humans, and let it run alone.
Lennart Schulenburg: That’s a stupid idea. And that will not work. So you have this, the human in the loop, uh, at, at every time. And the human is currently augmented and assisted. Because the humans is really good at making judgment at, at doing ethical decisions, and in most cases, at least at, um, looking at complex problems, the human is not so good at repetitive work.
Lennart Schulenburg: So if you have somebody sitting in an x-ray booth from for eight hours in front of a black and white [00:18:00] screen in a dark chamber. Looking for small porosities. That is not what the humans are made for. We are not good at that. Our attention spans are not made for that. Um, we’ll get tired, we’ll get exhausted.
Lennart Schulenburg: And this is where an AI system really can come in and act as an assistant. Same like you would have in your car, your. Braking assistant, your lane keeping assistant, your distance Warner, like these are tools that save life. And same could an AI system augment the human compensating for some of the human weaknesses and providing more time for the human to focus on the human strength, solving the more complex problems, communicating the results to other stakeholders, to colleagues.
Lennart Schulenburg: So. Back to your question, yes. AI systems should be supervised in the beginning, especially in critical environments. Yes, we need to put guard rails in place. Yes, liability is a problem. And ultimately, um, ai, an AI system is just a tool and you need to use it responsibly.
Nasrin Azari: Yeah, that’s good. The other, [00:19:00] the other thing that I, I was thinking about as you were speaking is that, um, not only are humans and AI tools good at different things, but.
Nasrin Azari: You know, an AI tool is very, very predictable and very consistent, whereas a human isn’t, and actually different humans probably have different skills or different, different expertise, different capabilities. Like some people might be really good at certain. Parts of their job and, and not so good at others.
Nasrin Azari: And I can all see potentially a world where there are different types of AI tools that complement different types of humans, that have different types of skill sets, which might actually broaden our pool of human, um, you know. Technicians because if you’ve got a human that you can pair with an AI tool that sort of compliments their [00:20:00] deficiencies, that kind of gives you a little bit more flexibility.
Nasrin Azari: It would seem.
Lennart Schulenburg: Yes. And our listeners don’t see me nodding, uh, heavily. I totally agree. Uh, nazarin. So, um, currently if we hire an inspector, that inspector’s task is to, uh, do maybe some of the manual preparations at the part they are, uh, reading, um, the results and they’re communicating the results to colleagues.
Lennart Schulenburg: So if you now would have an AI system and you before had three people, now you could have one. Person that is really detail oriented, supervising the AI system, because that’s this person’s strength. It’s a, you know, it’s a super, it’s, it’s a person who is always trying to find the issue with something.
Lennart Schulenburg: Just want to make sure it’s perfect. So that person is the supervisor. Then you have a communicative person. You know, that person never is so detail oriented, but he’s a great communicator to colleagues. Everybody likes that person. So that person could be the, the, the person that, that explains the AI results.
Lennart Schulenburg: To the other departments. So that’s this person’s strength. So yes, I totally agree. I think if we are smart [00:21:00] in how we, how we do people development and, and skill development, we could really scout for people’s strengths and put them around smart systems, automated systems where they can really, really, um, excel on their strength.
Lennart Schulenburg: And we, we compensate some of the weaknesses through, through AI and other automation tools.
Nasrin Azari: Yeah. Yeah, that’s, that’s exactly what I was getting at. And I think that that’s a fascinating way to look at AI tools as a helper of individuals to help round out where, where they either are weak or where they just aren’t, don’t have as much interest, um, to ba basically make a complete, you know, a, you know, kind of a complete human AI system to working together nicely.
Nasrin Azari: And, um. And once again, you’ve done a great job of, of leading us into our last question, which is focusing on the role of the inspector in the context of our evolving field. As AI continues to advance and the [00:22:00] NDT industry advances along with it, how do you see the role of the inspector changing?
Lennart Schulenburg: Yeah, that’s, honestly, that’s a very hard question and I, I don’t, I’m not sure if I, if I fully have an answer, but what, what we are, what we are seeing is that once you have implemented these systems, um, the, the role of the inspector is shifting more on these, on the interface aspects.
Lennart Schulenburg: So we. Where results get communicated to other systems or fed into an MES system or EIP system. So somebody supervising the AI system, somebody supervising the training process and making policy decisions. What is an indication? What’s a defect? What is it? When is it rejectable? So driving, really making these.
Lennart Schulenburg: Policy decisions and then working with other departments like fracture mechanics or simulation people to actually set these thresholds, right. To not create unnecessary scrap. And then you have the communicators that are communicating these results to other departments, making sure we’re making most [00:23:00] use out of NDT data because sadly, probably 80% of the data we generated is just thrown away.
Lennart Schulenburg: Mm-hmm. This is valuable data we generated. But, but nobody’s using it because it’s just used for a pass fail decision. So, so definitely these, these areas people will work on will change at the, at the same time, we need to be really mindful that we don’t lose valuable talent and skill because, um, it’s, that’s a big risk with using these great smart systems that you will lose the capability to do the job in the first place.
Lennart Schulenburg: And I, I, I think everybody can relate when we drive our cars. And, um, I got told when I was young still how to, how to use a map. I think my kids will probably never learn that except if I, if I teach them deliberately. Um, because everybody uses a smartphone. And if my smartphone in the middle of a big city will die, I’m.
Lennart Schulenburg: Big trouble because I don’t even have a map in my car, and if I have one, I’m not sure if I can read it and while I’m in traffic. So there’s a risk that we lose [00:24:00] skills that are valuable and are really, really important. Um, if we overly, uh, if we overly on technologies, just a risk. So we’ll, we’ll have to think ahead and have to try to maintain this skill, still use these tools and then deploy them.
Lennart Schulenburg: In an ethical and reliable way.
Nasrin Azari: Yeah, I think of the same thing with the telephone. Like I used to have to re, you know, I used to have to punch numbers in to make a phone call, and so. You know what ends up happening is you kind of memorized certain people’s numbers that you call a lot, and now you’re just like clicking on somebody’s picture or even talking into your phone, Hey, call Jim, or call home or, and you don’t even know what the phone numbers are.
Nasrin Azari: And if you had to, you wouldn’t know how to, you know, how to share it with somebody. So I think you, I think you bring up a really good point, um, that. In some ways, as we build the, the capabilities of AI [00:25:00] agents and AI algorithms, we’re sort of getting losing touch with how it all actually works. You know, we’re, we’re, we’re losing touch with the skills, um, ourselves, and that’s definitely.
Nasrin Azari: It’s, it’s definitely something to consider because those are valuable skills that can always ha you know, it, it helps us. It’s really important to understand what we’re doing and why we’re doing it, in addition to just actually getting it done, you know? Yeah.
Lennart Schulenburg: Yes. And in the end we have to, uh, keep in mind it’s a tool, right?
Lennart Schulenburg: And the tool doesn’t control us. We control the tool. Um, and it’s like a car. I mean, you can run somebody over over the car, you can drive from A to B with a car. It can be dangerous or it can be highly productive. And what matters is how we use it. And really we have to see it as a tool. And this is coming back all the way to the beginning of the discussion where the media’s doing it.
Lennart Schulenburg: Bad job with portraying AI as this self-serving dangerous. [00:26:00] Um, it’s not that it’s a tool, it’s mathematics, it’s statistics that will help you in a very narrow task to get that task done better. And then it’s our job to build the policies, the regulation, the procedures, the training around this tool to make sure it serves us as the master, because the human is the master in this, not the ai.
Nasrin Azari: Let’s hope it stays that way.
Lennart Schulenburg: Yes,
Nasrin Azari: yes. Yeah. I, I 100% agree. And, um, thank you so much Leonard, for being with us today. That was a very, very interesting discussion and I’m sure it’s peaked the interest of a lot of our listeners. Um, so for the listeners on the podcast, if you’d like to learn more about Leonard’s work and more about him, you can visit his website at Busy Consult.
Nasrin Azari: De and you could also follow him on LinkedIn where he’s very active and provides a lot of great insights on a lot of different topics. Um, so thank you so much again, Leonard. Is there anything else you’d like to add before we close [00:27:00] the session today?
Lennart Schulenburg: Well, I, I would like to ask you a question, Nazarre, what are you most excited, uh, when it comes to AI and NDT?
Lennart Schulenburg: What’s your take on that?
Nasrin Azari: So I can picture a world where. These catastrophic structural failures just don’t really happen anymore because smart machines are analyzing so much data so quickly that potential problems are avoided altogether. And I feel like that is like the nirvana for this industry and something that, that we’re all striving towards.
Nasrin Azari: Um. So I see that as kind of like the light at the end of the tunnel, at least the current tunnel. Um, but in the immediate term, I, as a software provider to the NDT industry, I am really excited to incorporate AI functionality into our platform, uh, so that we can enable our customers to just work that much faster and that [00:28:00] much easier and provide even more value to their organizations.
Nasrin Azari: Their customers. So I mean, that’s our goal at FLOODLIGHT is to improve the effectiveness of NDT inspection processes, and we’re already doing some of that. We’re already automating many of the manual tasks that are commonly being performed, but incorporating. Generative and agentic AI capabilities into our platform will just make it significantly more powerful, um, for our customers.
Nasrin Azari: You know, taking advantage of all of that data to provide trending analyses and just so many more. Capabilities that we are currently designing into our platform. And as we get there and as other products out there do the same, we’re all kind of helping each other get closer to that nirvana state that we all wanna get to.
Lennart Schulenburg: No, it’s, I, I think it’s really important, and I loved what you say, you’re saying the [00:29:00] same thing, right? You want, yeah. You want people to be more productive. You want people to use that tool and. And, and this I think is the thing I want to give all, um, listeners on, on take, take this away with you. Just try, uh, make a small project, try something, whether it’s generative AI or analytical ai, um, we are, for example, offering these feasibility studies where with very little cost, you can evaluate what could AI do for you?
Lennart Schulenburg: What’s the performance on my parts? And this is what we all need to do. We need to drive to small experiments. We need to see how does it work for me and what’s the value I get from that? Because at some point, somebody’s there with the money and that person wants to hear, what’s my return of invest? What does it do for our company?
Lennart Schulenburg: And so we need to learn this and we need to lower the, the fear and the, the, the barrier of entry. And you can only do that by trying, right, right. Yeah. And the project.
Nasrin Azari: Yeah. I think it’s important for people to. Gain as much understanding as they can and stay on top [00:30:00] of what’s happening, even if they’re not ready to, to do a full implementation.
Nasrin Azari: Like, I like your, your suggestion of just trying things. There’s so many tools out there that are free that you can just pick up and try and just get a feel for how, how generative AI works, for example. In any case, thank you so much for being part of the podcast, Leonard. It’s always, always a pleasure to talk with you.
Nasrin Azari: Um, in terms of, of our listeners, we welcome feedback as well as nominations for future guests. So to do that, you can send a message to us through the contact us form on our website, www.floodlightsoft.com. Um, thanks again Leonard, for being here today. Thanks to the, our listeners for joining us and see you next time.
Nasrin Azari: And that’s a wrap for today’s discussion. Head on over to our NDE 4.0 podcast page for more interviews like this one, [00:31:00] and reach out if you have any questions, feedback, or ideas that you’d like to share. Thanks very much and have a great day.
For more expert views on NDE 4.0, subscribe to the Floodlight Software blog at floodlightsoft.com.