A Practical AI Guide for Business Leaders with Brad Groux

Speaker A: Get ready to discover some of the most actionable DevOps techniques and tooling, including performance and reliability with some of the world's smartest engineers. Hey, I'm, um, Joe Colantonio, host of the DevOps Toolchain podcast, and my goal is to help you create DevOps toolchain awesomeness. Did you know that 95% of AI rollouts fail not because AI, uh, doesn't work, but because companies use it the wrong way. And it's not because the technology is weak. It's failing because most companies roll it out without context, strategy or empathy, but how work actually gets done. In this episode, Brad, founder of Digital Meld, shares a grounded real world perspective on how organizations can use AI and automation to eliminate busy work, increase efficiency, and unlock higher value work without replacing people. Listen all the way to the end because this episode gives you a practical framework to move forward with confidence and quick heads up before we dive in. This episode was recorded back in September, but after a four month pause while we retooled the show and rethought sponsorship, we're officially back. The format's sharper, the focus is tighter, and starting now, episodes drop every Thursday. The conversation you're about to hear is still incredibly relevant and honestly, more important now than ever. You don't want to miss it. Check it out. Hey, Brad, welcome to the Guild.

Speaker B: Hey, Joe, thanks for having me.

Speaker A: Really excited to have you. I know you have a lot of experience around DevOps, AI, uh, development, all the things. So just before we get into it, like, how did you get into AI? You know, it is hot right now. But do you have a background with AI that got you even more interested in it?

Speaker B: Yes, my background's traditional it, like, uh, hopefully some of your listeners, they're as old as I am. I got involved with it because of Y2K, of all things. I was going to be an engineer, and then I was in college for engineering in the late 90s. And then Y2 came along and like I was told, hey, I can make 25 bucks an hour opening computers for Y2K. I'm like, okay, cool. And I kind of never looked back. And then so I was in traditional system administration and stuff for through the, you know, early 2010s. And I got more in the consulting side. I worked, spent five years at Microsoft as a PFE or Premier Field Engineer. That's just a, uh, enterprise consultant for those who are not really familiar with that. And then from there I went to B2B Technology Consulting. So major large projects and stuff. And then worked a, uh, no code startup called Uncork for several years as a solutions architect. So I'm very familiar with the space, but I saw the rise of AI coming, so I decided to start my own company in the consulting space. I'm like, why am I consulting for these other organizations? Whenever I can cater solutions, I think better for our customers. And just saw the rise of AI coming, to be honest with you. And I'd gotten into automation before AI. I think that was my entry level into that by, for folks may be familiar with like Power Automate and the Power platform. That was my first foray into that in 2016 or so. You know, my goal is to be more efficient. You know, Elon Musk, for all his flaws, he says the best process is no process. And his goal is to, you know, be SpaceX, not Boeing. Right. You may iterate really quickly, you may blow some things up, but you're going to learn when you blow those things up. And I think AI is the perfect place for that right now.

Speaker A: Love it. Love it. I actually got married December 31, 1999. We postponed our honeymoon, uh, because of Y2K. We didn't want to travel because they said we were going to now have cash and planes were going to fall out of the sky. So that's kind of funny. Brings me down.

Speaker B: Yeah.

Speaker A: And it came around with the whimper. Right.

Speaker B: Like, I remember that as well. Yeah.

Speaker A: Yes. So, you know, with AI, obviously there's a lot of hype and you start a company around, there is any common, I don't know, misconceptions around what people may have about AI adoption that you meet a lot of.

Speaker B: Yeah. So, you know, the MIT study made a lot of, uh, that came out that said 95% of businesses, their LLM rollouts aren't really working, I think in July, and then it was reported in the last couple of weeks. That's kind of sprung up recently. And so I think people, uh, are thinking that AI is this turnkey solution and it's this end all, be all. And at the end of the day, I think if you live in a space, you realize it's just a tool. And the tool is only as good as in the hands of the craftsman who's holding it. Right. And AI specifically, it's all about context and the person ability of it. Like, so what's good for me, for my AI, is not good for an engineer or it's not good for a doctor, it's not good for a lawyer. And so I think you have these guys in Silicon Valley trying to build these solutions. That's the right thing for everybody. And I just don't think that's the future of AI. Uh, I think the future of AI is catered solutions, black boxes living on someone's desktop, they low code or no code, a platform that works for them on their little ecosystem and that's all that really matters to them. I think that's really the future. And I like to say the beauty of AI is it helps you punch above your weight class. And the same thing with low code, no code solutions as well. So you can have people. Microsoft, uh, coined the term Citizen Developer about 10 years ago when the power platform came out. And I think that's the future is you're going to have people that know their industry knowledge and their industry specific knowledge is what makes them special and they're going to use the tools to help boost themselves within that industry, not so much. They're going to have this overarching AI that does everything for everyone.

Speaker A: I totally agree. I've been vibe coding for months now and I can't imagine putting this in the hands of someone that doesn't know like health care or software testing and saying go at it, uh, because you still need to have contacts and you need to have that industry expertise to actually get it to work correctly in my experience. Is that kind of similar to what you're saying and seeing?

Speaker B: Yeah, for sure. Like the best way to utilize it within that specific business vertical. Right. Is you're going to have to empathize. Like as technologists one of the things we used to really struggle with is empathy. Right. A lot of us are very pragmatic and it either works or it doesn't work. Ones and zeros is the word we look in. And the world is made of varying shades of gray, I think. And like you need to empathize with that end user. I tell basically McKinsey and company that, you know, the big consulting firm, they're really preached. You go to somebody say what problem are you trying to solve? And then you're going to use whatever is the best tool for that. I think a lot of time people are just saying AI and most of the time they're probably just talking about automation. It's not really AI. Agentic AI is just a new term for business process automation that's been around for decades. Right. You're just injecting AI in little certain places. And I think the big thing is I do want people to know that they can do this like it, it is if you take the time. The beauty of AI is it's the first technology that's ever been able to teach you how to better use that technology and so have it learn more about you and about your business and your context and what's important to you and then say, okay, now that you know all of this about me and my business, how can you help me, you know, AI? And then it'll uh, give you like a roadmap that again it's only going to be at 80% there. The last 20 or 30% is always going to be your secret sauce is what I'd like to say.

Speaker A: Absolutely. So random is that do you think AI is thinking I think how the

Speaker B: models are currently trained? No, but I think with the new like I think the Google World 3 or the once we start. So I, I love the vision learning side of things. So training things to read lidar radar or thermal or just RGB cameras and like understanding is someone wearing their hard hat or not. I think how we're training stuff now is we're literally just throwing data at it and it's literally just going down and just brute forcing AI. Right. We're going to create these virtual worlds where you create and trains AI specific to their learning with what they see and like, just like a baby does. I think that's the future of orange. We've scraped pretty much all the data we can scrape. Right. There's just not a whole lot more data we could scrape from an LLM perspective. And so I think creating these smaller models that are very specific to a very use specific use case, like methane detection is something we work with. Like you don't need to train that on literally like how to summarize Harry Potter. You just train it on how to notice methane and how to recognize methane.

Speaker A: That's absolutely. So how then can leaders spot the difference between hype and the tools that actually deliver value? Do you recommend a proof of concept? Like there's so many options out there. Like how do you even know how to direct them? Like they may come in going no, we want ChatGPT and you might be like you said, you just want automation or like how do you actually dive into what's actually going to work for them?

Speaker B: Well I think again one of the reasons I got out of the corporate IT space is because I can't take the top down approach. Uh, with technology I don't think that makes sense. Like you come in and they try to brute force sharepoint or SAP or some other big behemoth that is a great product if you take the time to cater it to your actual needs. But at the end of the day it's a, it's a one size fits all type product. I think you talk to the boots on the ground and say what problem are you trying to solve? What are you doing every day that you wish you could automate? What are you doing every day that you could see AI? And so I think you have that conversation and I think what we see with, especially with enterprises is you have these silos and that's how it's been for security and compliance. We can thank Enron, NCI, WorldCom and all those guys for that. Sarbanes, Oxley, but I think that's gonna have to come down. What we'd like to do is start with the center of excellence. Whether it's AI center of Excellence or data center of excellence. And so you're bringing people in from operations and sales and accounting and legal and HR to come together and have regular conversations said we're going to define how we do this as a company and it's not coming from a dictation from the top down. Again, we like to talk to what we call the boots on the ground. You know, think of the NCOs in the military. They're the ones that really get the work done while you know, they're moving chess pieces on a board at a higher level they're kind of separated from what those day to day problems actually are.

Speaker A: When you do implement say Genai into a uh, company, do you train it on like a local model or do you use like a Gemini or a chatgpt? I know a lot of companies are afraid of security, uh, do you just train it just on their domain and their document so that you're just working with what matters in their context.

Speaker B: So we like to tell our partners, we call partners because we can't succeed without them and their input. Like we meet them wherever they are if they're a company that, if they're, if they have Microsoft 365 it probably makes sense to just enable copilot as the GIF for their organization. If you live in the Microsoft ecosystem, is it the best? No, but it's the lowest barrier of entry. It's pretty good. It gets you about 90% there and then you help train that internally within our organization and point it to this specific stuff. Obviously security and compliance is really big. You don't want people not in hr, ah, to have access to HR records and be able to query that those sorts of things. And then we have folks where we do black box things where like we use Gemma or Llama, one of the open source models and it just operates in their own little ecosystem. We have it done on people's desktops. Um, we work with field, field workers a lot in oil and gas and out in the middle of West Texas. And they're like, we just need this one specific model and do this one specific thing. It's like, great, we'll just get a Mac studio or something and run it right there local. I think that's the catered solutions is I think where the real power of AI is. These generative models are great, but they're generative models, they hallucinate, they don't train, they don't tell you when they're wrong. They just make stuff up on the fly, you know, like try to have an argument with your gen AI. It's not going to happen. It's going to say, you're amazing, I love you. You know, like. And so I think that's the key thing is just teaching people there's a difference between the types of AI. And again, we always start with what problem you're trying to solve, you know, and then you can you introduce like, what's your turnaround time, what's your cost? And that helps us drive. And do we go for like a Gemma or a Llama or do we go for like a paid, you know, AI model? It's really good. Uh, this example is invoices. Everybody has AP Automation. There's no reason for us to go custom make and train an AP Automation model whenever. Azure, AWS and Google already have those. And for 40 or 50 bucks a month you can scan in 5,000 invoices. You know, there's no reason for us to do that. So just it really says what problem you're trying to solve is where we start.

Speaker A: Now I know you make kind of a joke of it, but it almost lulls you into thinking that it's really thinking in that, oh, that's a great idea, Joe. Uh, Joe, you're really thinking this through. It never really gives you enough pushback. So how do you tell a partner that, hey, AI can do so much without them thinking, oh, it's going to do everything. Obviously it's going to agree with you. Like, do they have unrealistic expectations like that? Like, oh, I do have a great idea. All my ideas are great because it's validating me.

Speaker B: Yeah, I think that's where the evangelism part of it comes into. So I, I go to speaking engagements. I, I do podcasts like yours. I have a podcast like mine called Start Small Day Big. Getting people to know that again, the beauty of AI is it's the first technology that's ever come along that helps you learn how to better use AI. So you can say, hey, I have six months, I need to upskill for this. And I'm in this business sector, I have two spare hours a week. Can you give me a roadmap for that? And it'll give you like a six month roadmap for. That's pretty good. And you have the ability to do that. And as a business leader, if you're trying to implement AI, use AI to help you learn more about AI. But also check it like, you know, compare and contrast against Gemini and Copilot, uh, or chatgpt. I have tabs for pretty much all of them open and I compare and contrast too. And then look for those use cases that are very specific and then hone down on those use cases. Don't use these generative AI models. You know, do a very specific use case because the context is where the real, real secret sauce is that.

Speaker A: Speaking of context, are there any use cases that you think it could be applied to where it's not talked about as much? Everyone thinks code helps you with code. Is there anything else you see like, geez, you could actually use it for this. Um, and it's actually a good fit for this, this particular use case.

Speaker B: Yeah. The funny thing is, is we work with blue collar workers a lot. I'm from Houston, Texas. We're an oil and gas city, we're an energy city. Uh, most of our biggest clients are construction and engineering. You know, think of like a project manager on a road project, like a billion dollar road project. Think of all the invoices, all the time sheets, all the stuff that's tied to that, to the, and like these, these project managers are billable at like 300 an hour and they're going from site to site and they're like, what can you automate for their day? They're going out the site, taking pictures and they're coming back and running a report in their truck and doing all of that. Let's automate that. I think the biggest, I think blue collar is one of the, the greatest. I was just talking as a lunch meeting and there's um, a company that creates rubber seals and things was like, they have these 100 year old machines that they've added computers to, you know, for their CNC machines and Mill machines and stuff like that to get these insights to better understand like what their churn rate is, what their turnover rate is, what their efficiency rates are. And I think people overlook the blue collar work. And we have a saying down here in Texas called redneck rich. And so you see these guys driving the F350s and they have half a million dollar boats and a million dollar house and like these guys just, they wear coveralls for a living, you know. And I think people overlook that. And the big thing is like Silicon Valley is thinking about all those guys in Silicon Valley. They're like thinking about these B2B we're selling the Deloitte and Microsoft, the KPMG and all these big companies. Whenever there's so many mom and pop shops that are a hundred million businesses with 80 employees that you would never hear about in that space. Whether it's oil and gas or chemicals or we have a company that does frac sand. So like you would never think that this company would be worth a hundred million dollars, but it is.

Speaker A: So it's, it's interesting that AI has gotten to that level that those type of companies are know uh, about it then I would think.

Speaker B: Right, Yeah. I think the big thing is again we try to get and be evangelists. We could talk to chambers of commerce. I talked at Texas A and M and A University last week. I think that's a big thing too is creating a community. I jokingly say that we're creating a cult of AI having conversations like this one because you're going to have 15 or 20 person companies within the next five or 10 years that are billion dollar companies and it's going to be those people that have that industry knowledge. I know how to use these tools. But someone who's uh, a toolsmith and knows how to use tools, I don't know how to woodwork and how to work with that specific wood or that marble or whatever. Just because I know how to use that tool doesn't mean there's the ones that the AI experts, they just don't know it. I'm the tool expert that shows them how to unlock their full potential with AI. And I think that's some humbling from a technology perspective. Like we don't have all the right answers. We can point them in the right path and help train them and upskill them. But at the end of the day the getting the most out of that AI is going to come from that business use case. And we're technologists, we don't have that or Understand that business use case and we need to like let that go.

Speaker A: What's the main thing that you do then when you go into an organization with your partners, do you create like uh, a custom AI, uh application based on their context of their needs or is it a little bit of everything? Maybe you create an AI automated flow once you understand the business process or is it like what are the type of things you cover?

Speaker B: Yeah, so I started the business to make software because you could sell software a heck a lot easier and you can sell scale people. Right. But I quickly understood that especially these mid market companies think people, you know, 80 to a thousand employees that make hundreds of millions of dollars a year, they don't have dedicated IT teams. Maybe they have an MSP that they work with and then they bring in SMEs as they need it. And so I quickly found out that, and even OpenAI has, it's like uh, we, we have to be business analysts as much as we do technologists. I need to understand what problem they're trying to solve and then we utilize the tools and our knowledge of these tools as quickly as possible to figure out if you're a DevOps engineer. Right. You know that not every full web stack is created equal. You're going to use whatever web stack or hopefully you're using whatever web stack makes the most sense for whatever that job. Like you may not need, you know, Microsoft SQL or Oracle on the back end. You could get away with the MySQL or Supabase or a PostgreSQL or something like that. It's like you have to have those same sign of decisions as you're like architecting out the solution for the individual. If it's a company of 20 people, you don't want to put in Oracle for example, like how the heck are they going to support Oracle? Plus can they even afford Oracle? Let's be honest. And I think it's really in the future I think these white glove or catered services are gonna be really good. And we tell them we meet you wherever you are. And so there's really nothing, we've done stuff really small like literally you know, automating the inbox for a uh, receptionist who's, who's trying to let go of the uh, or do the scheduling for the executive conference room for a 200 person company. But we impressed her, we impressed everybody because she talks to everybody. So like no job is too small. And I think that like we'll sweep the floors if you need us to

Speaker A: is what we Say what's persuasive? What A.I. uh, does, is the time saved? Is it the insights revealed? Uh, like what is the mean? Like they m get it once they see it.

Speaker B: This is the real disconnect I think with most leadership. Because leadership, uh, especially in large businesses, if you work at a enterprise, you understand this. They think in quarters because that's what their bonus structure is. Right. Meanwhile, the boots on the ground, they're thinking in years and they're thinking like the US vs China AI debate. China's thinking in generations, right. And we're thinking in political terms two, four and six years. So I think that's the real disconnect there. And so our uh, what we try to teach, again, the leadership is like first, like don't think what you know the company needs. Go talk to your boots on the ground and get that understanding and trying to flip that on its head and understand that you're going to make your business far more efficient. If you talk to the people who are going to be utilizing these tools and trying to fit a square peg in the round hole, I think that's our biggest challenge. And again, but having conversations like this has really helped with that because again, we're just building that cult of AI and hopefully people learn more and more as it goes. The uh, beauty of AI too is it's something that everybody's interested. My 72 year old mother watches podcasts and stuff on AI and goes and watches videos. Like in what world is that technology been like that? I go give talks at AI events and there's there ranges from 16 years old to 86 years old. It really is amazing.

Speaker A: That's awesome. So when it comes to AI, especially when you talk to leaders at these companies, how do you ensure that they're using AI responsibly? Like, uh, is there any transparency, privacy or like ethics concerns? You have to kind of guide them like, huh, hey, you may not want to use it for this or anything like that.

Speaker B: Yeah. The number one thing we take when we talk to users is like, we're not teaching you how to use AI and automation to eliminate people. We talk about it to punch above your weight class to amplify those people. An example, we have someone who was doing accounts payable for a $1 billion road construction project here in Houston. And so she needed to make sure she's a woman with an mba. She's brilliant. She was doing the invoicing and she was doing the timesheets and literally it's just manual data entry. And she was Doing it because it had to be right because it's a billion dollar government based project and can't be wrong. There can't be any discrepancies. But this woman, like a woman with her master's degree or an MBA was doing this work and like this is mostly data entry. Let's automate 90% of that job and then she's the human in the loop. I see human in the loop, at least for the next 10, 15 years, especially in business for any automated process. Because you only have one chance to make a good impression is the way I say it. Like, and if you make the wrong mistake In a business B2B space, you're really going to have egg in your face. But anyways, we cleared up the time for that woman, that 90% and now she's chasing more work. She's still that human in the loop, making sure that all the numbers are right and it's really, really important that she does that. But now she's generated millions of dollars, extra dollars because she has her mba, she's chasing more bids and more work and they're doing more work. And so they're with the same headcount, they're able to, you know, elevate their business without ever really having to scale up.

Speaker A: So have you been seeing that though? Like you implement AI, uh, and every company is like, they thought they want to play something or they know they're gonna, they get to have them work on higher value, uh, activities now like what are you really seeing across the board? Because a lot of people say it replaces them. Like, oh, it doesn't, it's, it's different than the older technology. It actually there is no, you just change a title of what you're doing. I don't know if that makes sense.

Speaker B: Again, I think the leadership who think that's going to happen, it's, it's hubris. They think, oh, we're just going to take this solution and off the shelf and we're going to train it for a couple weeks and it's going to replace, you know, Deborah and accounting, that's not the real world we live in. It's just not. Yes, there will be some, anything that's manual data entry type stuff, that stuff is easily replaceable but hopefully what we have, we literally have this conversation. We come and do lunch and learns and we start the process. Like now that you freed up time for this person who was doing data entry, upskill them. Because the single hardest thing as a business owner is hiring good people. That's the single hardest thing. It's the single biggest time sink. You know, if you want to get a junior executive that's like a 80 or $90,000 just to the recruiter, like that's not the six months or a year it takes to upskill them to where they want to be. So our goal is to again amplify people to free that time and again. The leadership focuses on ROI a lot and we like to take it in time savings. Okay, you saved this much time here, you freed that person up to work like on more meaningful projects. Let's think of every business, especially small and medium business, have a huge backlog of projects. Like you're just freeing up all that time to work on these backlogs that you've been planning these honeydews.

Speaker A: Right.

Speaker B: That you wanted to do for years and years.

Speaker A: 100% agree. I'm a very, very small, small company. I have like a really good operations, uh, manager. Sometimes she's doing these piddly things where if I used AI to replace that, she'd uh, be able to free up to do sales things that actually bring in more money. So I wouldn't want to replace, I want to replace those low level activities and that would definitely free a lot of things up, I think.

Speaker B: Yeah. And one of the areas, again, I'm one of those people, I love taking things apart, knowing how they work. I call myself a jack of all trades and master of none. So I get those people, they don't want to let go of that little thing because they like seeing how the sausage is made. But you can still see how a sausage is made. And you're just doing the quality inspection at the end. You don't have to watch it go all the way down the assembly line. You're seeing it, you know, you're seeing the product go in and the finished product coming out. And I think letting go of that power, that's going to be the big struggle. And that's where we, you know, we elevate with lunch and learns and training sessions and those sorts of things to lead people. Like, yes, you think you're doing really fun and cool things now, but just think of all the things you could do if you were more efficient. And the biggest thing is like the more efficient your business is, we hope people scale their business without having to increase their headcount all that much. And that just frees up more capital to dream on even bigger things and even bigger things. And that's the beauty of our, of AI too is because it shows The ROI is pretty easy to prove when it comes to time and things, but laying the foundation, that takes some time. So like, you can't just build a skyscraper overnight. Right. You're not going to just turnkey solution with your AI overnight. You're going to have to do that little bit of extra training. Much like if you've ever used ChatGPT, the first time you use it, like, this is pretty cool. It's answering my questions. But the more context you give it, you upload PDF documents or give it reference links or give it context clues, it's going to give you better responses. It's no different than in the business space either. It's just as a larger scale.

Speaker A: All right, so that's, that's a good point you bring up. I guess it comes to mindset, there are a lot of people that do coding and testing that, like, not the minutiae, but like getting into the guts and like creating a function or a page object when it could just do it for you and probably do it better using code, uh, assistance. Do you see any resistance? Or like, how do you change the mindset to say, look, it's not replacing you, it's going to generate more code and therefore it's going to need you to be able to look at it and make sure it's doing it correctly. You may not just be writing it line by line.

Speaker B: Yeah, I think for the best coders out there, think of yourself as now a team lead. You're leading a team of AI agents and you're correcting that AI agent. Like you're elevating yourself to a code leader now. And look at it from that perspective. Don't look at like, this is replacing my job or it's doing most of the work and where's all the fun? Like, just step yourself up one. Place that in the chain. Right. Uh, like in the hierarchy. And look at that. Like, that's like. I think it's going to take the younger generation coming in now. They're AI first, right. They're using AI in school and they're all these things. Us, you know, as the older generation, we're having to rethink how we do things. But think AI and automation first. If there's something you can use AI to automate, why wouldn't you not do that and then focus on tackling the bigger products? Like an example for us, we do vision learning with LiDAR, radar, thermal, and we do multifusion models. I have a PhD on my payroll and she's amazing. And if I Have her working on the nuts and bolts and doing the stuff she can do, she can develop, that's great. But the real magic is in the algorithm and those things. I would much rather her focus on the cool stuff and that's the secret sauce that differentiates you from everyone else. If you're just writing a full stack web app, how many companies in the world can write a full stack web app? Like, come up with a new use case, come up with a, you know, Microsoft Garage projects or the 8020 rule. Take 20% of your time to focus on something you would never have worked on and who knows where you can end up. That's the way I look at it.

Speaker A: So this might be a little contrarian. This is one of the first technology I've seen where it actually gives older people a leg up because the AI is taking what I think is what an intern or someone new would be learning. But you already have the context now where you get the most out of AI. Um, what do you think of that? Is that a weird thought? Is that something you've been seeing?

Speaker B: No, I think you just nailed it. That's the real difference is what I try to tell these project managers or senior executives, like you guys, you have 30 years of business knowledge that these young guys coming in, they have no clue about. And you know, one of my co workers and partners, uh, said is that basically within the next 10 or 20 years we're going to have like what's equivalent to the burning of the library of Alexandra as this older generation retires and all of their business knowledge is in their head, right? And like, how do we unlock that, how do we utilize that? And so my goal is to have these younger, you know, like engineers, for example, we have these senior engineers have been doing over 30 years taking these engineers out of college and they work together and then it's like a brain dump, which you would normally do in a training session now, but now you're using, you're training an AI specific model for how your company does within an organization and standards. And I think that's the huge thing you do have the leg up. The more industry knowledge you have. That's your biggest differentiator. And I tell people that all the time. It's like, if I came up with an idea on how to, you know, revolutionize the chemical industry with an AI process, I would never do as well. As someone who's been in the chemical industry for 30 years who had the same idea and then upskilled themselves on building a POC. It's much easier to upscale yourself with POC than have 30 years of industry knowledge. Right.

Speaker A: You know, that's an issue I haven't really thought a lot about. I think companies should like, do you see that as that's like could be a really uh, destroy a larger company. They may think, oh, I'm gonna replace people. But actually that's gonna be a problem as people start retiring and aging out that they're losing all this context of the actual business and they're over relying now at AI that point where they're actually kind of not as productive, I would think.

Speaker B: Yeah. And like I, I'm sure you probably have some listeners out there who've worked at large enterprises. I've worked at Sheryl, Exxon, Chevron, all the big oil and gas companies here in, in Houston. And even uh, in the mid-2000s, like they had 16 bit applications that were written for the specific use case. And the person who wrote that application died or retired 10 years ago. And it's like this is our most business critical application and we have no idea how it works. That's happening everywhere across. Even in mom and pop shops where you have these, you know, machinist shops that have been working for 30 or 40 years and they have the specific, you know, machine that they use and only one guy knows how to use that. I think that's going to happen. The beauty is you can use a, uh, I get you like now we just record everything, Just record everything, get that context, have a meeting, whiteboard stuff, capture those whiteboard sessions and over time, but you have to have those conversations. That's the thing, you have to be proactive about it. And nobody's really thinking about that there. I guarantee you nearly every organization in this country, there's one person that's doing far more work than everybody else realizes. And if that person goes out sick or is a car wreck or like leaves, that just throws a giant wrench in the efficiency of that company. And it's on business leaders to know where that is. And the only way to do that is to talk to your people, don't work in your silos, remove those walls, no stupid questions, have lunch and learns. Like that's why we create that center of excellence. Right. Because you know, you know, I bet you there's some overlap too. You have inefficiencies, you have people in accounting doing a lot of the same sort of work that people in finance are doing or uh, people in legal are doing. And maybe you can come in and work together to streamline that process and say this is our way of doing it across the entire org.

Speaker A: Absolutely. I actually work for GE Healthcare. We had some technology that was built on like a database from 30, uh, literally 30, 35 years ago called like MIPS that no one knew except this one guy that was about to retire. That's exactly what happened. So I could see that being almost amplified now once AI starts taking place. For sure.

Speaker B: Yeah, definitely.

Speaker A: Awesome. So what are some other things you think people need to know about AI if they're trying to become an AI first type of powered business?

Speaker B: Well, first of all, I think you need to understand the geopolitical aspects of it. And I think that's not so much fun for everybody to know. But you need to realize our reliance on TSMC in the South China Sea and then the geopolitical aspect of it, the power consumption. Like if we unlock AGI tomorrow or artificial general intelligence or ASI artificial superintelligence model, we couldn't power it. We don't have enough power like our power generation demands. We don't. It takes eight to ten years to stand up a nuclear reactor. It takes eight to ten years to run transmission lines, the high transition lines. It's going to take everything we've got until we unlock fusion power. Basically. I think that tale of AGI I think is very oversold. Uh, AGI is coming, but we're pro, I would say at least a decade away. I would personally, I would say based on those with chip constraints, like because as a small company I go and I try to get on Microsoft Azure, AWS or Google. I was like, hey, I need cloud based GPUs. I need 10 of them to train this model for this industrial use case they go, that's great, stand in line like everybody else. But meanwhile, if you're a Google or a Microsoft, I mean any large company, any Fortune 500 company can say, oh, I need that. And they're just put to the front of the line because they're spending millions and millions of dollars. So that's where I think that catered solution is like, okay, fine, we'll just go around that. We'll go buy $10,000 Mac Studio and you can run Deep Seq locally in the middle of, you know, West Texas sand. You know, like I think you need to be. There's not. Don't get set in your ways. The best tool is going to change every single day. So use whatever's the best tool for the problem you're trying to solve.

Speaker A: So I know, I hear that a lot about energy consumption. What does that mean? Is it the compute power that's going to be needed for as more and more companies utilize it or is there something else?

Speaker B: What's going on? No, it's definitely the compute power. And so like they're building gigawatt data centers, right. And we have like five or six under construction right now in Texas that are gigawatt data center. You know what a gigawatt is? That's one, that's the use of one nuclear power plant. And we stopped building nuclear power plants in the 60s because of you know, Three Mile island and those sorts of things. Microsoft has paid 100/4 million dollars to stand back up one of the nuclear reactors, a three mile island because there's not enough energy. It's going to take everything we can throw at it. And China is using everything they can throw at it. That's the difference is our adversary in the space for lack of a better term is they're using everything about it. They're full fledged ahead. And I think it's going to take as this we're giving more power to the more powerful people. Right. The Sam um, Altman's, the world Elon Musk. And so I think it's really important for the smaller people to have a voice there. And the only way to know that is to know about these things and to say no, this is the right way to do things and we're going to do this ourselves. Otherwise the people who you give the keys to the kingdom they're going to run and play and they're just going to leave you behind. And I think but the beauty is technology is a great equalizer. You have the ability to do all the same things they're doing now. You just need to be more, a little bit more nimble with it.

Speaker A: Yeah, I feel like we already lost that battle almost. I don't know, maybe I'm pessimistic rather than optimistic with that.

Speaker B: I would say look at the transition from like GPT 3 to 5, right. Like GPT 5 was supposed to revolutionize the world and people complained so much they had to roll back the force grow. That's the thing is we've hit that threshold. They scrape the Internet like they've scraped. There's not so much data, more data that can do it now AI is producing more data that we can then train the models on too but there's no reason you can't do that internally as well and they don't have access to. As an example when you look at the Cameras from an industrial use case to see if someone's wearing a hard hat or a vest. It's PPE detection, you know, or someone's in a no, no go zone or kill zone. Google and Amazon and Microsoft, they don't have access to that company's recordings of that. So that's your secret sauce. Use your data that is specific to you to differentiate you from the others in that space, whether it's methane detection or it's your, the pro. Whenever you have a manufacturing process, we have a company, uh, that's a brewer. It's a $50 million brewery and they have their manufacturing process and they what's their yields when they use this or that? And Silicon Valley doesn't know about that because it lives within their ecosystem. And so I think it's kind of overblown that they're going to take away with everything because the context is key. That's the real, real thing. And again, OpenAI knows this. That's why they've spent $10 billion to stand up their consulting arm, much like Microsoft, Google and Amazon has, is because they realize they can't teach people enough, fast enough, how to consume these solutions. And so they've spent $10 billion to stand up this consulting space. And there's the arms race of talent and all those sorts of things. But the beauty is you don't have to have a PhD to understand how to do it. Like a vision learning model or just a basic model. It's all out there. All the information's out there. On Google, you can take Harvard classes for free in your underwear. Like, if you have the will and the ability, you can do this. I know you can, because I did it. And I'm. I wouldn't call myself the sharpest tool in the shed.

Speaker A: It's a good point about almost a company has an IP that they may not be aware of. Is that something you help with that? Uh, you say, hey, you know, you have this unique IP that you probably could leverage to give you a leg up. And having someone on the outside looking in probably would help with that, I would think.

Speaker B: Yeah. And so what we built is a health and safety platform again that we called Rubicon, and we announced it about a month ago. And so where you can monitor things like fire detection, PPE detection, you can read things from, um, you know, vessels, you know, looking at, from a inspection perspective. And as over time they can train that model and then license that model out to other people in that same space. Right. There's no reason, if you have this ip, why would you not want to license that out? And yeah, that company could. They may be your competitor and stuff like that, but your opera. We look at things like what we do is, isn't different, it's how we do things. And I think more companies need to think about it that way. The world is big enough for all of us to have our success, I think is really what it comes down to. But if you're not utilizing your data to at least transform your internal business operations, you can also use that data to help improve other businesses as well. And that's where that licensing comes in.

Speaker A: Nice. So I know some people hate this question because you never know. No, we're getting close to 2026. Is there any trends you see that people need to be, uh, more aware of as we go into 2026 and beyond? I know you seem to be on the forefront of visual almost. Is that something that's being overlooked or they nail the trends people think you think should be on right now before it takes off?

Speaker B: I would just say Iot in general. Again, manufacturers have sensors everywhere, right? And they're probably capturing data. Again, we work with road construction as an example and one of the things we worked on with the Texas A and M Traffic, uh, Institute or working with on them is like, how far back do you need to know where the. Whenever you have road construction, do you do the taper? You're closing a lane. Like how far do you put that taper, capture that data? Like, okay, last time we started the taper on the 60 mile stretch road three miles back and it was 40% more efficient. Like they're capturing all that data but they're not utilizing it. And so if you have sensors somewhere in your operation or uh, your business process, and a sensor is literally anything capturing data, that's where the secret sauce is. And again, especially if you're in manufacturing or any process that's based and every point in a process, whether it's a business process or like a manual process for like a manufacturing is to say that's a data point. And that's the thing too is once you get your, your SOPs, your standard operating procedures and your processes documented, start capturing that data because you'll find where those inefficiencies are. And this includes like just project automation or just internal, you know, bite collar type stuff as well. Like any process to start capturing that data. Because what separates Microsoft, Google and Amazon from everybody else is their data. All their products and services are just a means of capturing more data that's all it is. That's where their secret sauce is, and that's where you could do that as well. And that's the thing. If, uh, a product or, uh, service is free, you're the product. Right. So like Instagram and Facebook and TikTok and those things just know. Think of it from that perspective. Change your perspective and think of, like your business that way as well.

Speaker A: You know, I've been trying to call IoT since 2015 as a new trend, but you're right, with AI, it makes it a lot more possible. Is that often overlooked? And you think that people actually have this data hanging around and they're just like completely missing out, um, on the low hanging fruit.

Speaker B: Yeah. How many businesses out there have SCADA systems? Right. And the SCADA system's the bane of their existence because there's an outage here or an outage there and it stops their assembly or stops their chemical process or something like capture that data and you'll be able to understand what those things are. And that's where we're seeing even things from, like manual, uh, detection of, you know, anything you need to put a set of eyes on, you can. Like pipelines is another example. They fly airplanes. My brother's a pilot for Southwest Airlines and he flew a thousand hours flying pipelines every day. Just to say, is this pipeline, is there a leak or something like that. Now you can just set drones every 10 miles, you know, and every day it goes. Checks the pipeline. Checks the pipeline. And so just think from that use case perspective, if there's anything you need to. A gauge you need to look at, that's at the top of a vessel, you know, 300ft up. Don't send a person up there, just fly a drone up there. And you can see that gauge.

Speaker A: Absolutely. All right, let's start wrapping things up. I guess. First thing, top of mind is what do you cover on your podcast? Small Think Big.

Speaker B: Yeah. So start Small Think Big is literally trying to just teach people that they can do this. And the goal is, is how we start with our customers. We generally start with the low hanging fruit. Hey, let's get the ball rolling. Do we work well together? This kind of gives you that roadmap. The first 10 or 12 episodes are talking about the use cases and how you can get started, how you can have that internal conversation within your company. And then since about episode 12 one now we bring on experts to kind of echo what we've done. We've had people from Hollywood talking about how, uh, AI is changing you know, animation and movie making and stuff. We've had people from construction, engineering, we've had people from education. How that's changing. It's literally it. That's. The beauty is like that you can glean insights from all of these and understand that thing. And just the goal of the podcast is to teach you, you can do this. Again, if I can do this, you can do this. I guarantee you. And the beauty is you have these technologies and these tools to help you get there, too.

Speaker A: Okay, Brad, before we go, is there one piece of actual advice you can give to someone right now to help them with their AI journey? And what's the best way to find or contact or learn more about digital melt?

Speaker B: The first thing I would say is get involved, um, and maybe even locally. Go to meetup.com. if you live in a large city, you may have, like an AI meetup. Just start talking to people. We have one in Houston called the AI, uh, called Houston AI Club. Go talk to people. Go find people. Again, the context is what's really key. If you work in an industry, go to industry events and say, hey, how are you guys utilizing AI? And if you're not, why don't you work together to kind of come up with use cases for your industry? That's, um, my biggest thing is it's. It's funny, as we get more digital and we get more separated, I think what's going to differentiate people in the future is that human touch, that coming together as people and understanding the best way to use things. And so that's how I'd encourage everybody. And that's why I started my podcast is hopefully have those conversations. Conversations and kick that start, Kickstart that. Um, you can find me anywhere. Uh, I'm Brad Grub. B R A D G R O U X. I'm the only one in the world. So if you Google that, you'll find some embarrassing stuff from like tech forums and stuff in the early 2000s to sports forums and game forums and things. I'm, um, on LinkedIn, happy to have a conversation. Just say, hey, this is how I would do things if I was in your shoes. You know, I always say, put good things out there, good things will come back. I'm not a salesperson. I will never try to sell you anything. If you want to work with me, great. If not, I hope you still succeed. Know the world's big enough for all of us. Start small. Think big Podcast is on all the major podcast platforms, on video and Spotify and YouTube and just linktree.com digitalmeld and you can find us anywhere.

Speaker A: You can find links to all this awesomeness down below. And for links of everything of value we covered in this DevOps Toolchain show, head on over to test guild.com p202 so that's it for this episode of the DevOps Toolchain Show. I'm um, Joe. My mission is to help you succeed creating end to end full stack DevOps toolchain awesomeness. As always, test everything and keep the good. Cheers. Hey, thank you for tuning in. It's incredible to connect with close to 400,000 followers across all our platforms and over 40,000 email subscribers who are at the forefront of automation, testing and DevOps. If you haven't yet, join our vibrant community at, uh, Test Guild and where you become part of our elite circle driving innovation in software testing and automation. And if you're a tool provider or have a service looking to empower our guild with solutions that elevate skills and tackle real world challenges, we're excited to collaborate. Visit test guild.info to explore how we can create transformative experiences together. Let's push the boundaries of what we can achieve. With Lutes and Liars the bards began their song A tune of knowledge, A um, melody of code through the air it spread like wildfire through the land Guiding tester showing the secrets to behold.