Driving the adoption of generative AI in industry Podcast – Transcript
New technology can offer a lot of promise for improving on long-standing techniques or developing innovative approaches to different challenges but, when it comes to adopting these new technologies into complex industries, there are always challenges to overcome. With advancements as complex as artificial intelligence and agentic AI, unlocking their full potential across the design and manufacturing process will take time, yet in so doing, achieve new heights of what is possible.
Check out the full episode here or keep reading for a transcript of that conversation.
Spencer: Hello, and welcome to the AI Spectrum podcast. I’m your host, Spencer Acain. In this series, we explore a wide range AI topics from all across Siemens and how they apply to different technologies. Today, I’m joined by Shirish More, Senior Product Manager for NX, and Michael Taesch, Senior Director of Product Management for NX Manufacturing. So we’ve talked a lot about the foundation models, AI agents, where the industry is going involved in terms of AI and what the benefits are going to be. But I think we can’t get there without understanding some of the major challenges and the risks that are involved with bringing something like generative AI into industry. There’s obviously the risks of hallucination and the general concerns people have around that. And I’m sure there’s other things as well that most of us don’t even think about. So maybe, Shirish, you could start by walking us through a couple of those.
Shirish: Right. I think the first one that comes to my mind is user trust. So when I talk to customers, they normally are interested in understanding the roadmap. And the way I position my roadmap to them is, yes, we have all these nice capabilities coming out for you to use. But what are some of those application areas where I see fit for foundational model or even the agentic framework? The first thing that comes in my mind to answer this question is user trust. How can I gain the user’s trust? And engineers must trust the suggestions that, for example, a co-pilot is giving him based on the context in which he’s performing some task. And it’s a simple question, but it’s a very important one because I need to gain the user’s trust, making sure that the framework that we have implemented, the underlying foundational models that we have, not only it understands the engineering language, but the suggestion that the user is getting, the engineering content is generating or editing, it’s also industry grade, meaning that, yes, it’s vetted. You can trust it, and we will give you choices and options to verify what we are providing. So user trust comes in my mind. We need to gain user’s trust. We can, from a roadmap perspective, we can start with use cases wherein they can see what AI is doing, how the agentic framework is implemented, the information it’s providing. And then the second one that comes to my mind is the data privacy and the IP protection. So all of this foundation model, the AI agent will need some historical data or things that are approved in production or proven out in production. And another question that I’m going to get is, all right, how are you leveraging this continuous learning or continuous learning architecture using both the customer IP and as well as the Siemens Out-of-Box Framework? So data privacy and IP protection is another thing that comes in mind. And those are the two key things. And then the regulatory compliance and safety based on the industry for which we are developing these foundational models or AI agents. And that’s the third thing is we’ll make sure that working with both the industry experts as well as the regional regulations, we want to make sure that we meet the certification and all other auditing standards before we implement anything. So those are the three key things that come to my mind that we should keep in– I’m aware of, at least, before we start implementing and rolling out any solution to our customers, Spencer.
Michael: I mean, I heard exactly the same three from different customers on the manufacturing side, especially in data security. One, for me, was the most important one when I talked to a customer who was really distrust. And I think there’s really two types of trust. One is this hallucination. And I think you covered it very well, Shirish. Another one is really the trust, the general trust in AI and not just about hallucination, that this AI is not going to take my job. I think in the– especially in manufacturing industry, we’re in a very conservative industry. Everybody’s afraid of looking at jobs, so they’re looking at job security. And most of the customers I talked to, honestly, they were not receptive at all with this idea of AI. And no one was why. I mean, I don’t want to lose my job. Why would I help you? Well, why would I use this tool? It’s never going to replace me. I have 25 years of experience. And I think it’s time to address the elephant in the room here. We’re really not here to replace those guys in any way. We’re really here to help and empower them, right, and really help the next generation of camp programmer. They know it’s hard to bring new people on board. They know they need to do something. They know they’re going to retire. So when you start explaining, it’s not a tool. It’s going to replace them. It’s a tool that’s going to help them. Then you can start up a conversation. And then it became a lot more receptive. And I’ve seen a lot of people nodding like, yes, I could use this now. So I think for me, at best, we’re really looking at making their job more relevant and increase their creativity. They are really smart people. And right now, I see them focusing on just simple tasks. I want them to focus on high volume. I want them to focus on complex part, program more part, less time. Why not using the talent that those guys have to find to maybe, in my case, a camp process when spending time creating operations from scratch? And I think this is where we will help them out building this trust. That’s why we also have an approach of not, hey, we’re going to replace you right away. We’re going to have a level five autonomy. We’re going to do this gradually. And we’re going to bring the customer along this journey. And we increase the level of autonomy gradually with them, not against them. And again, this trust. And when you start explaining the vision, when you start explaining how we want to do it, I can see in the eyes, change the tone and be much more receptive of this idea of AI. So of course, like Shirish mentioned, we need to make sure the results are good. But this trust is more like human trust, which is quite interesting when you talk about AI. But it’s there. We need to have them to trust the AI.
Spencer: Yeah, definitely. I think you touched on a really interesting point there that cuts across all AI in industry, which is, we don’t want to be replacing people. We want to be replacing them doing these very small, menial tasks and let them move up to that high level, that high value work that only they can do. You’re not replacing people with AI for that kind of work. But like you said, the highly manual CAM stuff, if you can build an AI agent or an AI model, that can do that instead. Makes all the difference. Yeah, so I guess to round things out, after this great conversation, do either of you have any kind of closing thoughts about what we just talked about or the broader impact of AI in going forward?
Shirish: Yeah, I think just to summarize, and thanks for setting this up, Spencer. These are really exciting times. And what excites me most is how AI is becoming genuinely useful. It’s no longer a hype. Talking with a number of customers, they are coming up with so good use cases and requirements, wherein you can clearly see how genuinely useful it’s going to be as we move forward. And with some of the workflows that we discussed today and domain-specific activities that we are putting together, leveraging AI, what you are going to see in your future is we are building tools which will start solving some real world problems for both designers, engineers, and manufacturers. And the key is to keep AI practical, responsible, ethical, and deeply embedded in the workflows where people already work, going back to that human in loop discussion. We want to get AI and we want to get it deeply embedded in the workflows where they’re already working, where they are already using tools that we have provided. And to be honest, as someone who has been in engineering for more than 20-plus years now, I have never seen a moment quite like this. AI is no longer an experiment or a research project. It’s becoming a trusted partner in design and manufacturing processes. So I believe honestly, our job going forward now is to shape it responsibly. It reflects the expertise, values, and creativity of our engineers. And then they can start seeing the value that they get out of our AI tools.
Michael: Could not say more. I mean, 100% with usurers. And let me tell you a little bit of a secret. When I started two years ago with this initiative, a boss told me, just, hey, we’re going to use AI for CAM. I was really, really skeptical. I mean, I’ve been doing this like you should realize 20 years. And we talked about this. I’ve seen it. I’ve seen the attempt. I really didn’t think AI could replace what we do, especially in the CAM. I mean, we all special in CAM. We all know that. It didn’t take me long to realize that I was wrong. And really, every day I’ve been working on this project and with the Industrial Foundation team and the AI team, this is going really, really fast. Now, this improvement every day. I mean, we see even our personal life as much as you want to fight it. You need to use this tool. You need to be– all the kids need to learn about the AI. So really, along the same line as you, I share, my advice is just don’t fight it. Really embrace it. Because it’s coming, whether we want it or not, really. And we know manufacturing is a key component of the economy right now. And there’s really nowhere around adopting technology that will boost our manufacturing capability. Why would we fight it? So out there, just embrace it. It’s going to change our world. I know that. I agree. This is not a hype.
Spencer: Well, what a fantastic note to end things on here then. Like you said, it’s going to change the world, right? So I’d like to thank you both once again for joining me here.
Michael: Thanks, Spencer. This was a really, really, really good podcast. I really love this. I think you’ll be both of us already passionate about this.
Shirish: And it’s always nice talking to you both, Spencer and Michael. In fact, me and Michael, we shared some nice ideas when we made a Realize Live.
Spencer: That’s great to hear. But that’s all the time we have for today. So once again, I’ve been your host, Spencer Acain, on the AI Spectrum Podcast. See you again next time as we continue learning more about the exciting world of AI.
Siemens Digital Industries Software helps organizations of all sizes digitally transform using software, hardware and services from the Siemens Xcelerator business platform. Siemens’ software and the comprehensive digital twin enable companies to optimize their design, engineering and manufacturing processes to turn today’s ideas into the sustainable products of the future. From chips to entire systems, from product to process, across all industries. Siemens Digital Industries Software – Accelerating transformation.
