The All-Axis Podcast

The Reality of AI Adoption in Manufacturing

• Michael Thiessen, Sebastian Chedal

Share episode

Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.

0:00 | 48:46

Artificial intelligence is dominating headlines, but moving beyond browser-based sandboxes to achieve real-world productivity on the shop floor requires technical AI governance and deliberate strategy.

In this episode, host Michael Thiessen sits down with Sebastian Chedal, founder and CEO of Fountain City, to separate AI hype from practical industrial application. Sebastian shares direct insights on the current state of manufacturing AI adoption, explaining how custom autonomous agents are compressing development cycles, eliminating ERP reporting bottlenecks, and preserving critical institutional knowledge amid an aging workforce. Together, they explore how manufacturers can transition their team "behind the glass" to orchestrate high-level operations rather than getting trapped in digital stagnation.

Whether you are a manufacturing executive, plant manager, or industrial technology lead, this conversation offers a clear roadmap for safely and effectively integrating AI into modern production workflows.

In this episode: 
00:00 - Introduction: AI Hype vs. Real-World Manufacturing Impact 
05:15 - Moving Beyond ChatGPT: Moving from Sandboxes to Autonomous Agents
12:40 - Technical AI Governance & Protecting Proprietary Operational Data 
21:10 - Overcoming ERP Bottlenecks & Automating Industrial Reporting 
30:05 - Preserving Institutional Knowledge in an Aging Workforce 
38:50 - Moving "Behind the Glass": Empowering Operators as AI Orchestrators
44:20 - Key Action Steps for Manufacturing Leaders

Tell us what you think - send us a text message!

Thanks for listening!

Speaker - Voiceover

Welcome to the All Axis podcast from the Experts and Tebis. Each episode we'll talk about technology topics, trends, or solutions that manufacturers need to know in today's rapidly evolving times. Now, sit back, enjoy the discussion and let us know what you think.

Speaker 1 - Michael Thiessen

Hello, my friends. Welcome to the All Axis Podcast. I am your host, Michael Thiessen. And um at the beginning here, I want to kind of set the stage because in last several episodes, we really have been focusing a little bit on, not even a little bit, quite heavily, on the entire AI thought process and where we are and what we're doing, what should we do, what are the concerns. Uh and and because it is such a hot topic that everybody is asking questions about AI. I mean, we can't help ourselves anymore in our daily lives, um, not even thinking about AI, if that's Copilot or Grok or ChatGPT or any of these other AI that are open AI for all of us. I mean, we can't help ourselves. Every single place we go, there's some influence that AI is being utilized. And of course, uh, from us in the manufacturing society, especially uh in the automotive mold and dye, that's also something that you guys are asking about quite a bit. So, of course, it's only fitting that we want to have an expert on today that can really talk about the the entire thought process of AI adoption in at your company or uh in the manufacturing society or just in general, and then kind of tell you what's working or what's what is applicable, what is not quite working, what are your benefits out of this entire AI? We're gonna be talking about all of all this AI hype, and uh we talk about the adoption and and all of this, and then also we're gonna we're talking about where is it actually getting stuck? What are some of the problems that we're running into, especially in manufacturing? So I'm gonna be talking to a gentleman who is the founder and CEO of Fountain City. This company is a technology studio that deals with AI agents and that builds autonomous AI uh agents. So we're gonna be talking about him uh and about his company, Fountain City, how what the education around AI looks like and the adoption, this word AI governance, we're gonna dive a little bit into and all of that kind of stuff. Sebastian Chedal, you have been doing this thing for I guess about 25 years now, working with multiple different brands. I would say you have quite a bit of knowledge when it comes to this. So listen, I gotta step back in time with you because Fountain City, you've started this company when was this back in 1998?

Speaker 2 - Sebastian Chedal

1998, yeah, in Amsterdam, the Netherlands. Yeah, so quite a while ago, and it has evolved a lot. Our biggest we've ever been is 22 people, but most of our history has been well, most of our recent history has been smaller. Right now we're actually even smaller than historically because we have a lot of agents that we're bringing in right now. But yeah, so we've uh 98 is when we first formed in Amsterdam and Netherlands, and then in 2008, I moved over to Oregon, Portland, Oregon, and so now I have to assume that back in 1998 you weren't talking about agents and AI, or were you talking already or looking at that already?

Speaker 1 - Michael Thiessen

Because for most of us, AI is pretty new.

Speaker 2 - Sebastian Chedal

So actually in 2001 or two, but somewhere around there, I did a project for Heineken, which was essentially a early version of an LLM, or an early version of you know AI systems, because what I built for Heineken at that point was a system that took natural language text, you could just text a phone number. And Heineken go likes to go on these tours to show off their um you know their technical astuteness as because they they love to buy or help support lots of small breweries around the planet.

Speaker 1 - Michael Thiessen

Right.

Speaker 2 - Sebastian Chedal

And so when they go visit a small brewery, they might have a beer, right? Because they sell beer. Of course. And as they're in the bar talking, so the the idea was that they were bringing this projector with them on a screen, and so you could just text this phone number with something about soccer, because you know, beer and soccer goes well. They go hand in hand, yes. They go hand in hand, and you might say something like, Oh, I want to see Zidane score in the World Cup in Paris in whatever year that was, 2000 or something like that. And then so I built the whole system that took that text message you send, translated it into something the computer can understand, you know, nouns and verbs and things like that, broke it apart, synonyms, and then matched it to a whole database of videos and then shows like small clips based on what you're sending. So that was kind of like an early version of that. And I was already, I mean, I've been interested in cognition and how to make computers think and evolutionary systems forever. So, like the world is now catching up with all of that, with the large language models that have become so popular these days.

Speaker 1 - Michael Thiessen

Well, with all this story, I almost envisioned you that you're like Einstein, white hair. Okay, and uh because you definitely are a genius, because back in 2003, at least we here at Tabis, we didn't think at all about AI or computer learning, even though we are working in the computer field with our cat cam system, it never crossed our mind. So you were an early adopter, really, when it comes to this AI, and throughout your uh what isn't it, close to 30 years now almost that you you've been doing this? You've ran probably across a lot of the industries. I mean, uh if that's manufacturing industry or some of the other industries that are out there. So I've got to ask the this question with in terms of the AI adoption, especially in today's world in 2026, what is your overall thought process or the read of the of these industries of where the uh the AI adoption stands? Is it being utilized or not, or where does it where does it sit right now?

Speaker 2 - Sebastian Chedal

So I'm gonna just talk about the last two years in particular because things are very different right now in 2026 than they were in 2024. So 2024 into 25, there was a lot of hype deflation, I would call it. But even with the hype deflation, there was already a big disparate difference between people's exposure and experience with quote-unquote AI systems. I say AI systems in general because it's kind of a big bucket within that. You still have today, and you know, today meaning we're almost mid-26, but you know, Q226, that there are managers, executives, sometimes in very big enterprise companies who are barely just still using chat GPT in a browser and are just starting to get familiar with it.

Speaker 1 - Michael Thiessen

Right.

Speaker 2 - Sebastian Chedal

And yet there's there's been a lot of advancement the last two years, but a lot of it has catalyzed and become this new uh new emergent uh abilities and capabilities have really appeared this year in particular, the first few months that has really pushed things across. So adoption looks very lumpy right now, even within companies. I was consulting actually for um manufacturing company. Um they do software, but it's related to manufacturing. I was consulting with them three, well, at the time of this recording, it was three weeks ago, and I finished that up. And in that company, I I cannot name their name because of NDAs. Of course. But in that in that company, it was their product engineering team was very advanced with AI, like to the point where I was impressed with all the things they were doing. But then their non-product engineering teams are still in the stone ages and struggling to get access to things. Meanwhile, they've got multiple licenses for different, potentially overlapping solutions. Their marketing team is begging the IT team, which is now the product engineering team, so there's like this confusion of roles going on to install for them these MCP servers, but then they're not using the official ones through their enterprise licensing that are that make the MCP controlled. Instead, they're just getting on the side, on the hush, like local MCP servers installed on their laptops to connect to their marketing system. You know, so it's like the governance and everything on that is just like all over the place. And then the education is really lopsided as well. So it's you just have the screaming ahead product engineering team, which they're doing feature releases that used to take six months, they're doing them every six weeks now, with you know, with completely, you know, no all bugs solved, product release tested, validated, you know, six-week cycle instead of six months. So it's a huge acceleration in productivity on that side. And meanwhile, the rest of the company is just scrambling on the ground in the desert and not knowing, you know, how to and just missing basics.

Speaker 1 - Michael Thiessen

But but that must be incredibly difficult for you. I mean, when you walk into these types of companies, well, you have opposite worlds where you have the early AI adopter and really the geniuses in these in this company uh adopting it quite well, and then you have potentially even management that are supposed to be the decision makers that are that are kind of pulling back the reins a little bit and go, oh, well, not so fast. Uh, I don't know what this thing does. And they're kind of stopping it. That must be incredibly fast. How do you deal with this when you walk into those type of companies to convince both sides that this is the best thing for their company?

Speaker 2 - Sebastian Chedal

Well, my role specifically in the collaboration was the technical aspects of AI governance. But there was the services we were providing were also non-technical as well. But that was my focus was on that. So, in terms of the governance layer in that, you know, so my focus is on, well, first helping management understand even what kinds of levels of AI is going on, because a lot of what they're focusing on is the things that they should have been focusing on in 24, 25, but not the new stuff you can do with AI in 26, specifically. You know, now you can have actual agents that do entire jobs independently, and they weren't even thinking or realizing that they have people in their company who are building those. They were thinking more about I want to make sure that a human doesn't take private information about our company and goes into a browser window in ChatGPT or Gemini, whatever, and pastes in that information. It's like, that's great, you're thinking about that, but they're now you've got your whole data that's accessible to these agents, and how are you even controlling or managing and having oversight over those? So it's um it's a combination of policy, right, which are rules that need to be, in my opinion, as simple as possible and not super restrictive, otherwise, people will go around them. And then it's also the systems in place that enforce the policy so that people need to remember less rules. You know, if you if you have a system that just naturally filters out private information when you use an AI, or maybe the AR you're using is vetted to be 100% secure and you don't need to worry about private information going in it, that's way easier than trying to get people to never make a human mistake of posting something in that they shouldn't have. You know, you're about it's bound to happen, even if it's by error, uh eventually, if it's just a rule instead of a system that's enforcing it. So that was a lot of my focus, as well as just helping management to get a better sense of what's even going on in their company, so kind of uh you know, outside perspective.

Speaker 1 - Michael Thiessen

But if if management is not on board or not a believer into this, uh I wouldn't even say believer, that are not immersed in it in this entire AI world because they just learned how to talk to Chat GPT or asking simple questions. How do you or giving the guidance of how to do certain things within a company?

Speaker 2 - Sebastian Chedal

Yeah, I think education really applies across the board. It doesn't just apply to the people in on the floor who are doing the work, but it also applies to management. And I think there's it depends a bit on the person that I'm talking to. Some people are very eager and excited and want to learn and they want to figure out all this new AI stuff, and others are you know resistant, skeptical, don't understand the value. And so depending on where someone is, I think the conversation is different in that regard. Sometimes projects that take place within companies are from that place of skepticism as a whole, like the entire, not just maybe the upper management, but the whole company might have skepticism. And then the kinds of projects and initiatives that take place are things that are designed to prove validity in the value that LLM systems, you know, AI systems can provide. And I LLM, but it could be machine learning, you know, whatever the flavor of AI is for that particular situation. Um, you know, so then we get into roadmap planning and prior prioritization of projects and things like that. Sometimes organizations have a different issue, which is that they have a mandate, you know, maybe they they don't want to fall behind on the AI thing, so they come up with a mandate like everyone, 20% speed increase, or everyone should have AI installed on their computer, and we need to figure out those licenses, but there's nothing beyond that, you know, and those are really either not show in a lot of value, and then they're like, why are we not seeing a lot of uptick? Because there needs to also be a North Star, like a driving direction that is turning that is applying these systems to things, you know.

Speaker 1 - Michael Thiessen

So which industry sees that North Star? Or which which are accelerating more in your personal opinion when it comes to the AI uh application and its use?

Speaker 2 - Sebastian Chedal

Yes, so I guess just to give a bigger picture of the kinds of clients that I work with or have worked with, so I would say historically about 60 or so percent of our clientele has been in manufacturing. But you know, the other 40% has been pretty broad across different industries. In 2024, the vast majority actually of our work was in manufacturing. But in 25 and now into 26, we've seen a lot of proposals. Either proposals haven't been going further with manufacturing, or we do a discovery process and then it just kind of stagnates. And so manufacturing industry, at least from our small sample set of what we've been seeing, the manufacturing customers we've been talking to most are still not moving as fast as other industries. Other industries are moving a lot faster right now, especially this year, though, really. It's the this year is when things have really picked up. Where we're seeing the strongest demand is actually from digital agencies or other software development groups. Uh, anything that, or any company that has a product that is software or that uses software or builds software are seeing huge upticks of AI adoption. And and I think that has a lot to do with the breakthroughs that have been happening since December in agentic development. But right now there's also a flurry of speed around everything that is autonomous agents and agent development, uh, or just embedding agents into applications and services. So, you know, we have a few projects that are very distinctly different from each other. The only thing that's similar is that we're using AI to build applications that have AI in them. And then the AI that's in it, we're doing work to, you know, whether it's security, hallucination reduction, you know, error rate reduction, but also cost optimization, observability, you know, stuff like that, as well as the actual feature development of those applications. So those range from data intelligence platforms, connecting ERP systems together into on-demand reporting that you can then just talk to with conversationally uh all the way over to like voice agents that we're building that listen into executive calls and then offer very very well thought, considered questions at just the right moment, which is not that easy for an AI to do AIs. You know, if you want to get good response, usually you have an AI that just goes blah, blah, blah, blah, blah. You know, but here we have an enterprise environment with management layer, and what you you know, for the system to be good has to be.

Speaker 1 - Michael Thiessen

These companies now, these industries that you just now mentioned that that uh you've done you've dealt the most right now with, what would you say is their biggest gain? What is what are some of the tangible items that they're walking away with? Is it all about just the speed or the the the quality of the decisions that they're giving you, or maybe even cost? What what is it? What is their best their gain that sets them apart?

Speaker 2 - Sebastian Chedal

So there are three, yeah, and you mentioned the three. So there are three main parts to any technological innovation revolution for it to really be a tidal wave of takeover. Yeah. If it's better, faster, and cheaper than it takes over, it's a it's a huge revolution. The better part is the part, in my opinion, that's holding back the pace right now. That's the harder part. That's where you know, if I look at where we spend time, it's kind of like making a sword out of metal. So you you can get a sword really quickly, but to really get a good sword, you have to polish it and polish it and polish it, right? Like all the time comes in the sharpening of the blade. And it's like that a lot with these AI systems. You can make something that posts terrible content on your website at high speed for cheaper than you would ever pay a person tomorrow. Like that's not hard, but getting it to post articles or to do jobs, let's say it's jobs within your you know, manufacturing process, whether that's inventory management or sales facilitation, whatever it might be, getting it to do it as good or better than a person, that's where the craft comes in to achieve that. And so where I believe where the impacts have been really strong right now is in those areas where the AI, not only the model, but the harness, which is what the AI goes into, the control that kind of it's like the body of the brain. Isn't another way to think of the harness, right? Okay, so the harnesses and the brains have been hitting these catalyzing moments in a few different fields this year, which have really created that catalyzation. So everything that's related to sales, content, production, uh thought leadership, you know, everything in that area has really crossed over the hill. And then any kind of job that is very deterministic or workflow-y, but still requires a judgment call. Because that's what the LLM essentially does, is it helps you to automate a qualitative decision that normally before in code you couldn't because the code wanted it to be either one or two, right? Like by or zero and one, I guess binary kind of choices. Right, right. Um, so all of that has really opened up, plus also the actual coding itself, whether that's coding those things I talked about or actually writing and developing software code, which then opens up the world for SaaS kind of applications. So we're doing a lot also of custom apps, like one application. Actually, that that data intelligence one I was talking about, that one is saving the company $60,000 a year. And in and it only took about, I think, that first release. And there's going to be more features added, but it was a fraction of the of the 60k. So they're they're paying a one-time cost to build of a fraction of their annual, and then the monthly amount is like way less than you know their old SaaS platform, and it's connected. So you can create these custom systems now that solve your particular needs. And ERP pain, and especially manufacturing, I mean, I think like one out of two manufacturing companies I talk to always bring up how they're unhappy with their ERP for insert the reason, and it's never the same ERP. So apparently they all they all make people unhappy in different ways. But there's now real solutions to connect systems up in that way. We had a project, it didn't go all the way through last year, but which, sorry, you know, it's not a released project, but we we got far along with one last year where it was about connecting the ERP over to a system for doing custom reporting and then also connecting all of the IoT systems, the Internet of Thing, capable um manufacturing, essentially robots that do different jobs on their floor to that centralized system so they could get more advanced on-demand reporting from all of their equipment and from the ERP than the ERP currently allows them to do. And and I think their ERP was trying to remember the number, I think they were charging them six thousand dollars or five thousand dollars per custom report that they wanted them to create. And I remember when I saw that number, I was just like, that's that's crazy. Yeah, because once you build this system custom, making a new report is just it's peanuts, it's really fast. Yeah.

Speaker 1 - Michael Thiessen

So with those kind of gains, I mean sixty thousand dollars that they're saving, that that's not small change for for a lot of the companies. But you mentioned earlier that back in 2024, you were heavily involved in them in the manufacturing sector. I mean, it's it's only fair to ask because this is a a podcast that is really focusing on the manufacturing side. Yeah. Um, what is Changed in your mind, or what do you think has happened that manufacturing has become a little bit more stagnant? I mean, uh I would I would think that they're welcome this, uh welcoming this because of the labor shortage and all those kind of things that they're running into, that they would apply AI on every single corner of their business.

Speaker 2 - Sebastian Chedal

Yeah, I would say there's multiple factors that are taking place. Back in 24 as well, we had the pressure of an election season, which created uncertainty. We had, I don't know if it was a recession, but there was definitely economic stagnation. And then going into 25, several people that I know in the manufacturing world, or you know, or clients I was talking to, or clients that I had, or even just conferences when I would go and network with people in manufacturing, there was a theme that was similar there too, which was that a lot of the tariffs that were being placed added and removed were creating a lot of uncertainty in the manufacturing sector. And uh, you know, and lots of people were pro-tariff, but they even the people who were pro-tariff said that the thing that was difficult for them was not knowing if the tariffs is gonna stick, if they're gonna how long they're gonna last, or they just appear out of nowhere. And so that was creating a lot of chain disruption. And then, you know, and so for them also problems with potential in some cases, problems within their in financial investors not wanting to invest yet because they're not, you know, you can't calculate your projected return on something if you don't know what the tariff's gonna be in uh X number of months. I don't hear tariffs as much anymore right now, like that has definitely quieted down, but that was a theme during part of 25. Another thing in 25 that I saw a lot was education around two categories. One, management would hear AI is amazing, you put it in, your world's gonna change instantly. And so then we would they would ask for this project. I remember this one client in particular, they had a very long list of all of the security credentials and certifications that they wanted the new system to be built in and enterprise and all these things, and they just kept adding more and more things on the stack of what they wanted the system not only to do, but to be accredited in. And you know, by the time we came back and told them the price tag, because like it's now we're talking like a serious deployment, they're like, they were just shocked. They were like, I thought this was just gonna be really quick and easy project. And cheap. And cheap. Yeah, and I thought it was gonna, you know, in that exact and that company in particular, they uh that manufacturing company was in a situation where they told me they have more demand than they can handle and they have production capacity to double, but their only bottleneck was around the whole sales enablement, um, you know, all that middle part of like actually being able to get your customers in, find out their specific, because they have to do custom quotes. Uh, so they have sales engineers that build out custom solutions for the manufacturing. Uh their customers are other manufacturing companies, so they they're building out solutions for them. So that was their bottleneck. And so we were looking at a project to do essentially AI-assisted sales engineering process with augment um helping to do the the quote for the engineering quotes that are at least 80% accurate really quickly, and then all the sales follow-up and enablement to make sure that so that there's the sales engineers are spending as little time as possible on the stuff that they don't want to be doing anyway, which is a lot of the follow-ups and chasing and and also the the initial buildup of the quotes, you know. Anyway, so this whole project for them was projected to potentially double, you know, help them to double their revenue from 40 to 80 million a year. And so the ROI on this platform was going to be hit within basically a month or two, but it was still sticker shock for them. They were just like, oh no, this is way too expensive. And then we had another proposal, also very substantial. They wanted to do a lot of intelligence between, you know, their, they have uh engineers that go on to these webinars that do these show and tells, but then they have lots of questions, and then it's really time consuming for them to answer all those questions. So they have like this, and then internally they have very limited people who have all the knowledge in their head, and so they were looking at a solution to kind of take that knowledge, make essentially a kind of a virtual engineer that also to protect against retirement, because a lot of people in an in in that company, but I think a lot of manufacturing companies are um eight, there's an aging problem. So anyway, so that project was around that, and again, it was gonna need a big uh, you know, uh like five people working on this project for a certain number of months because it was a lot of information to pull in. And then we were almost there to initiate it, and then CEO jumped in. Bless his heart. He was he said, Hey, I've I met this kid who just graduated out of school. He seems pretty sharp in AI. Let's not do this project, we're just gonna hire him and see what he can do. And I remember that moment where, like, me and the other four people, which was also some of the you know, people on the client side too. So even on the client side, the everyone just was silent looking at each other, so confused. Because we were just like, wait, what? This is like a really involved multi-system project that that you want to do with really ambitious goals. And is it I don't know. So, you know, there's the and and so hard to read into that exactly what the motivation is. You know, is it that he thought that this person could actually replicate what you know is taking a bunch of engineers to be doing who have all this experience? Is is it just a buyer education challenge in that case, right? Like, why does it need to be this sophisticated? Or maybe, you know, they're just not feeling ready and they want to just see what they can get done working from the bottom up. But there was a lot of that, those kind of two themes. Either, you know, I'm gonna just do it myself, how hard could it be? Or I thought this was magic, why does it cost money? So those were the kind of the two two big themes in manufacturing last year that I was seeing a lot.

Speaker 1 - Michael Thiessen

Well, let's look at one of them because I think this is huge, because when you walk into the room, you I think you can kind of feel or sense which people have the knowledge and which one don't. And then I think you would automatically go into like a teaching mode and education mode. How how do you handle that? Because obviously you're there to promote your your company and you have solutions of how to do it, and all of a sudden you are almost like the kindergarten teacher and trying to trying to teach them of how this AI is going to be implemented and what are the steps are and what what the end result and the gains are gonna be. How do you deal with that?

Speaker 2 - Sebastian Chedal

Well, I mean, a lot of it depends practically on who it is I'm engaging with and talking with. You know, if I'm if I'm directly engaged by the CEO of a company, it's a very different kind of conversation than if the procurement department or the marketing department or the VP of technology and so forth. So I'm not always tapping right into the top layer of a company, especially when it's very big. Um, but that said, I I mean one thing that I I mean this has always been true of just how I work personally, at least, is I'm pretty good, or I focus a lot on translating things that are technical to non-technical and then back and forth, or you know, I I also can cross over into other domains like marketing lingo or design lingo, or and I'm mil multidisciplinary. That's kind of my background. But I would say that you know, different people care about different things. So if I'm talking to the CFO, they want to know about how much this is gonna cost and what are the risks and what's the asset value and what are the possible scenarios? In what scenario is the asset value worth zero and we wasted all our money, and what's our ROI, and blah, blah, blah, blah, blah. You know, and then also for them, they want to know about the alpha gain, like if they're if the growth rate of the company is 4% and this is projected to bump them up to 8%. That's what they want to know. And then how do we then measure that over time? Blah, blah, blah. So I mean, you're having a different kind of conversation there than if you are with a leader who is, you know, oh, in some companies, you know, there's there could be resistance within certain teams or people. People are worried about, you know, job retention or uh, I mean, well, job retention is a big one, but then there's also people who are just anti-AI for other reasons. So that's a very and so then that person might be wondering about maybe the leader is very pro-AI and is trying to figure out how do I get the team to be on board, how do I resolve these these points of tension? And so there I'm working more on a level of how do you turn this into a vision away from a mandate? Let's how do you frame projects, like reframing a project so that it's there to support your team, to augment them. So, because a lot of what we're doing really is to help people to be doing less of the things they don't want to be doing and more of the things they do want to be doing. It's the it's like the microcosm of the business as a whole. If the business as a whole is trying to lower its bottom line and increase its top line, people are trying to reduce their bottom line in air quotes, which is their time that is mundane and wasted on stuff, and increase their top of the line, which is time that they're doing on things that is more fulfilling, more thoughtful, more interesting.

Speaker 1 - Michael Thiessen

Have you ever run into a leader uh in one of your meetings that you have the sense of that he wants, I'm gonna reduce 50% of my labor cost if I have this AI thing there?

Speaker 2 - Sebastian Chedal

So it yes. I mean, there are companies whose goal is to lower the bottom line. I think the easiest place you can make cost benefits from these LLM systems, the very first thing that's easy is cost reduction. Cost reduction could look like someone is spending out of their 40 hours a week, 10 hours a week doing this road basic activity. So you automate that away. That person then has 10 hours free to do something else. You're not necessarily eliminating that person's role, you're just opening up their time to be doing other things. But I mean, the you know, the truth is there are people who are looking to replace certain roles with agents that are doing that job instead. I'm not gonna say that that's not happening. I think the type of work that people need to be and are also moving into are just going to be different kinds of roles that are using AI to augment their capabilities. Rather than so, I think it's almost like um the vision I have in my head is like a slippery balloon or something like that. If you just try and eliminate the human being entirely, what ends up happening is the human just kind of slurps into a different kind of position, you know? So people are absolutely um so the second phase that we're gonna see, going back to that, you know, right now it's how do companies reduce their bottom of the line. But if company A has five times more output than company B, the question isn't how do I reduce my bottom of the line, it's how do I match that level of productivity. And you need human beings to be managing and monitoring that. Right now, when people are thinking about Chat GPT in 24, 25, they're thinking about the analogy would be you're on the floor building, let's say, a car, and up until now you've been using the welding equipment yourself. And so you roll in this little robot arm that hopefully is going to pass you exactly the right tool when you need it. That's kind of the AI in a browser perspective. But where we're moving right now in 26 with these autonomous agents and the roles that people need to take in this new world, it's more like moving people behind the glass on that car manufacturing floor, and it's only robots building them the cars themselves. So people are still very the kinds of jobs and functions that people need to be training for and doing in this new world are just very different than the old ones. You're not manually making belts anymore by hand, custom made. You're building the machines and maintaining and managing the machines that build the belts. So there is job displacement, is is gonna be a natural function of this.

Speaker 1 - Michael Thiessen

Right. And and I think quite a few people that we're of course we talked to too, they're apprehensive with this entire AI because they're saying, well, if if AI AI will take over my job, and then I'm out of a job and all those kind of things. So there's this natural scare of, okay, I'm gonna give in to this AI, I'm gonna give AI a chance because eventually it's gonna it's gonna take over my my job. I mean, I I have a personal situation where this person uh used to work in in the automotive field, he left, and automatically uh this person uh or the manager then came to his co-work and then pretty much asked, Hey, do who do we have to backfill this individual? And then the answer was, Well, you have nobody. Oh, and then of course, the natural progression was, oh, let's figure out if we can do this in AI. So there's of of course that's care. But what should leaders when they deal with you, and especially there's probably more than you out there, because this AI field is such a booming industry right now of learning and educating and all that kind of stuff. And there there probably is going to be a difference between a real application or just a a presentation or a demo, and how can a leader or a company figure out the differences between a company A that just gives you a really beautiful demo, but really the application may not be correct, or company A that really says, um, you know what, we can really help you out from this. And what should they be asking those particular individuals?

Speaker 2 - Sebastian Chedal

So I think the tangible outcomes is really important. And thinking from the end result backwards. It's kind of like if you build a house, you think about the end product, and then you work backwards on how you're going to achieve that. Roadmap planning within that is really important. How do we so I when I work with companies at this early stage, especially ones that have lots of ideas, I like to map out all, you know, first you're brainstorming on all the different opportunity areas that exist, either from their own ideas or I ideas that I bring in. And then from that, you're prioritizing between impact and effort. And if you want to try and get some quick wins, then you're trying to hit the quadrant where it's lowest impact. Sorry, highest impact, lowest effort first and focus on. And then I think you can, I mean, a lot of these gains you can really see in most of the time in three months, maybe six at most. If it's a big, wide enterprise and you're trying to do a big large large rollout, obviously it's gonna be more than that time. But that's not usually where companies are starting. They're starting with you know a department or an easier system across the whole company, things like that. And I think there's just different sub-areas to be thinking about that the person you're bringing in has should be thinking about. So that's you know, maintainability, how is the system going to self-improve and learn? Uh, how what is the interface people are going to be using? Are you thinking about security? Are you thinking about context and memory and and knowledge? And how is the data? Because you need to have good data for these systems and you need to have good process. And if you don't have those things in place, then the project needs to include the either the cleanup or the creation or the organization of data and process in order to make the system successful.

Speaker 1 - Michael Thiessen

Very quickly, what does the future look like in AI, especially in your mind? What does that picture look like? And hopefully it's not gonna be like Terminator where AIs and the machines take over.

Speaker 2 - Sebastian Chedal

So the future, and it depends how far we go into the future, but I'm I'm gonna go a little, maybe like a short midterm. So, you know, not like this year, but in the next few years, what we're gonna be seeing. I I like to think of how people are interacting with these systems in a vertical and a horizontal direction. So the vertical is the building of these AI agentic systems to enable then people to work across the horizontal. So, what I mean by that is more and more of the things that used to take time to quote unquote do to go between either problem solution or idea result are just all compressing and shrinking to zero as the platform, in quote, which is this AI infrastructure, AI-enabled infrastructure, is handling that part of the equation, which is also because we're going to be seeing huge um acceleration in software development in general. So software tools. I don't I don't think we've even seen the beginning yet of how we're about to write more code than we've ever than we've written in the last 20 years in the next year. I can guarantee you that. So it's it's going really fast right now. What that might look like is, and let me actually take a step back to talk about the kinds of areas that people are going to continue to be important in, because that'll connect the dots here. So the first is directing, and by that I mean, so of course it's directing at the top level, but even within your function, your discipline, it's the direction of that. So it's you know, your expertise, why are we going left or right to begin with, uh, navigating the ship, you know, within your engineering department or within your marketing department, whatever it might be. Then it's the the next area is the interaction layer. And by interaction, I mean people to people. You know, people want to buy from people, people want to listen to podcasts done by people, Michael. So I think you're good. I still have a job. So, you know, anything that's people to people is going to stay at least for a while until the new kids grow up and they think it's it's stupid that you're not just talking to robots, they prefer it. I don't know. But right now we're we have a long runway of that still. People want to talk with people. And then the third category is ownership. And by, you know, so if the AI system isn't doing the right job, but also if you're talking to a person and they they're worried about the product or something like that, it's not that you can't point your finger at the AI system, you still point it at a person, and then that person takes care of that problem or they own it. You can't grow a business with just one person, you need to have delegated responsibility. So ownership is the third kind of uh area. And then the the building of the platform or these solutions is you could argue it's both building and supporting, you know, the creating of it, and then the maintaining and supporting of it. There's a support role also in the human element of it, too. So, you know, what that might look like in practice, you might have a conversation with someone about specing out, I don't know, I'll just invent something like the custom heating cooling system for their factory, because that's what your business does. You build those things for other facilities. You'll talk with them, you'll learn everything you can, you recorded that, you know, the conversation is just naturally recorded. You get back to your office space and the the black box, which is all of the AI platform, the you know, software platform that's behind you, supporting you, is already coming up with possible solutions for you to look at, uh options to consider, and then asking you questions about things that maybe you or the client need to weigh in on to help with that decision. So you're having a conversation at that point, reviewing so you're more reviewing, orchestrating rather than doing a lot of that manual work.

Speaker 1 - Michael Thiessen

Right.

Speaker 2 - Sebastian Chedal

And then and then you get back, you know, the results that you then just review, check off. Maybe you have something to change, maybe you don't. If you had something to change, then the AI system learns from that so that the next time it doesn't keep doing the same mistake again. So it's learning actively. And so it's kind of removing a lot of the friction, I would say, between idea and manifestation. It's actually really, I think it's really exciting, and I think there's a lot of this resistance to it, but it's also incredibly empowering for anyone who wants to be faster to manifest whatever it might be, whether it's sales or products or creative ideas.

Speaker 1 - Michael Thiessen

Right. Well, it's a good thing that Arnold isn't coming anytime soon, so we're we're safe on all of this. But let me close out with this and ask the the one of the final things that to you, because obviously we're dealing with time here. Obviously, most of our clientele, I have a feeling they're a little bit behind the eight ball. They're not quite understanding where they are, maybe they're behind AI, they're really not understanding it. What is your advice to them? What should they be doing right now that haven't been involved in asking questions in the AI field, and maybe even they want to communicate to you and contact you and get and shooting your brain and seeing what's happening uh in this AI field and how they can get involved. What is your advice there?

Speaker 2 - Sebastian Chedal

Things are moving and changing very quickly. I think having so you know, staying up to speed right now is very challenging. I mean, I find it challenging and I'm super deep in it. And by challenging, I mean there's just way you have to be picky because there's so much happening on a weekly basis. I would say, especially for management, I think you need to check a few boxes. First, you need to have a little bit of exposure to it to kind of get it so that it's not just pocus pocus to you. Next, you need to understand, I would say, the levels of capability or the potential benefits and gains that are happening right now in different areas. That might include also an understanding of the classification of different sophistication levels. Everything, and there's and it's a little challenging right now because everyone is misusing terms. Like everyone calls everything AI, and everyone calls everything agentic now when it's not really that agentic, you know. So having an understanding of these levels of capability so that you can become an educated buyer, right? I think is a really important area to be focusing on. The next area to be thinking, so I would also do an um we do these AI readiness assessments. I think I have a blog post on that that you can just you. I mean, you don't need to hire me for that. You'd be I'm happy to be hired for that, but you can also just review the blog post and check yourself on where you see you your uh level of sophistication is across the different domains. Domains, usually that's also different between depending on your company size, it could be different in different departments and so forth. And there the approach is really looking at the weakest link in that chain and then focusing there because it's going to hold you back. You know, some companies, it's they have no AI governance policy at all, even though they're moving ahead in other departments, and that's scary. You know, and then another company might have really good policies in place for the AI governance. I mean, this usually doesn't happen, but it could, but their AI adoption is just snail pace in so they need help more in that. So yeah, I would say uh AI change management, focusing on that change management in general, the principles of change management, governance, technical capability, understanding, and then complement it with a bit of personal experience. And then thinking practically about how do we turn this from experiment into realized value and where where is that value best realized? Because sometimes it's a clear ROI of time savings, other times it's something that's more that's a little less indirect, which is knowledge retention. There's other areas you could also focus on that are important, which is getting new talent in the door. You know, new talent wants to come in. I mean, new talent will bring with it AI capability because kids are all you know up with the new the new coolness, right? But you need to also create an uh an attractive environment to attract those kind of people as well.

Speaker 1 - Michael Thiessen

So it's a little bit of a if they want to communicate with you or contact you because you seem to have all of your eggs in a row and you know exactly what the AI world does, because otherwise we wouldn't be talking to you. What can they reach out to you?

Speaker 2 - Sebastian Chedal

Different ways. I'm pretty easy to reach. So there's the company website, fountaincity.tech, and contact form. I will definitely see that. You can also find me on LinkedIn. I do have a YouTube channel that right now it's seem I guess it's like once a month I'm posting a video, but sometimes it's once a week. And there's also a newsletter on Fountain City if you want to be more of a voyeur than an active reach out. But uh, but yeah, I'm very I will be happy to talk to anybody about anything. I'm that brings me a lot of joy. It's just hear what people's questions are and see if I can help them.

Speaker 1 - Michael Thiessen

Well, Sebastian, I really appreciate you to coming on here and and giving us your thoughts and also the trends and where we're going and all this. So thank you first and foremost for uh joining us here. I hope that we run into each other with the entire AI surrounding us. Hopefully, it's not quite gonna happen where these robots are all running around and we're standing behind the glass, even though just recently I saw that yeah, one of the airlines is trying to use uh humanoids to push uh the luggage stuff and do all that kind of stuff. So it it's it's gonna be an interesting future. I agree with you on that, but and it's gonna happen relatively fast. So thank you for your insight. Thank you for your advice. And any final thoughts from you since uh we still have you on here?

Speaker 2 - Sebastian Chedal

Uh no, thank you so much for having me. This was a great conversation. I I really appreciate it.

Speaker 1 - Michael Thiessen

Awesome. Well, my friends, you heard it here first. AI is here. Grab it while you are still in the infancy stage so you can learn along with it. Start with the small things first, then uh get invested in it, learn what it really means with AI governance, what that all means, and get connected with the right people because you could spend a lot of money going in the wrong wrong direction. And this is why companies like Sebastian and Fountain City, and there's probably many other that can help you out doing this. You should be looking into this because you know what I'm hearing a lot of is I don't find the right people, I don't have enough experienced people, all this labor buzz. So definitely look into this AI. We don't want to replace individuals, but we also don't want to close our doors. Thank you so much for tuning in. We hope you found that our discussions are giving you some thoughts and some ideas, some insightful items to think about of what you can do in your manufacturing sector. Feel free to share our episode within your network, send it via emails, via text or whatever. We'd love to hear your feedback. So send us your feedback. We're here to help you, we're here to help uh give you more ideas uh that you can find in your own four walls, and that's the reason why we're on the air. So until next time, keep shaping the future, and we'll see you soon. Thank you. Bye-bye.

Speaker - Voiceover

Thank you for joining this episode of All Axis. Please leave a comment or review with your feedback or what you'd like to hear in future episodes. To learn more about manufacturing technology solutions and Tebis's capabilities, visit our website at Tebis. That's T E B I S dot com.

Head Shot

Michael Thiessen

Host