Ep 67- Navigating the AI Revolution in Business with Devlin Liles

Transcript

Transcripts are generated by machine learning, so typos may be present.

BTB (00:00):

Welcome to the Building Texas Business Podcast. Interviews with thought leaders and organizational visionaries from across industry. Join us as we talk about the latest trends, challenges, and growth opportunities to take your business to the next level. The Building Texas Business Podcast is brought to you by BoyarMiller, providing counsel beyond expectations. Find out how we can make a meaningful difference to your business at boyarmiller.com and by your podcast team where having your own podcast is as easy as being a guest on ours. Discover more at yourpodcast.team. Now. Here’s your host, Chris Hanslik.

Chris (00:42):

In this episode, you will meet Devlin Lyles president of Improving Devlin is a leading expert in the application and use of AI for businesses. Devlin shares several helpful ideas relating to AI for businesses and believes that a business’s readiness for AI is mostly psychological. Devlin. I want to thank you for taking time to join us today. Why don’t we start by just telling us a little bit about yourself and, and kinda your background and your role with improving. Sure.

Devlin (01:12):

So Devil Miles, that seems like an odd thing to say. <laugh>. Uh, so I’m a technologist by kind of trade and training. So I started writing software when I was very young. I was eight when I started programming. That’s pretty young. My dad got me into it <laugh>. I started my first software company when I was 16 in high school building. Used car websites and that kind of thing right@thekindof.com bubble expansion. Okay. And so decided I was gonna not do that as a career. I was gonna go become a professional soccer player. That didn’t work. So I kind of fell back into it as a hobby and kind of continued on that most

Chris (01:45):

Programmers think of a professional soccer player as a dream. Right.

Devlin (01:48):

<laugh>. Yeah, absolutely. And so I ended up kind of falling back into my hobby as a career and then came up through kind of corporate IT at Tyson Foods and then got into IT consulting and I’ve been doing that for the last 15 years. So that’s a bit about me.

Chris (02:03):

Okay. And, and let’s talk a little bit about improving where you serve as president. Tell us a little bit about what improving does and your role there. And then, you know, one of the things I really wanna focus on as you know, is thing that’s on most people’s minds over the last oh, 12 to 18 months is ai. So let’s just kind of couch it in that context.

Devlin (02:23):

Sure. So my role with improving has kind of evolved over the years. So I actually started as a consultant delivering to our clients. And I came in kind of two and a half, three years in. And so we have an equity share model. So I grew an equity share at improving and then took over as president, uh, here in Houston in 2017. My global role for improving as chief consulting officer. So I own client delivery, thought leadership, go to market and employee growth kind of that space. And so AI has been a big part of that conversation. Now the interesting thing is I get to live in a time machine somewhat in this space of AI’s been a big part of that story for us for five to seven years. The world, uh, with chat GPT kind of making it a part of the zeitgeist is really catching up. And so it’s cool to have these conversations and really talk about it because a lot of our excitement and like, oh, it’s gonna be utopia from 20 17 20 18, when the, there were some big strides being made forward and now get to kind of relive with everybody else.

Chris (03:25):

Interesting. Yeah. So I guess you’re living it for the second time.

Devlin (03:28):

Yeah. And it’s, the thing is that going through it the second time, you get somewhat of the hindsight in real time, which is interesting. Yeah. ’cause we ended up helping a lot of customers apply some of these technologies. And technology always has this kind of pull to off the shelf Right. Systems. We used to pay tens of millions of dollars to build custom. Right. Think about CRMA client contact management system. Right. Right. Almost everybody has today. Yeah. In the 1990s that was a multi tens of millions of dollar project for only the biggest companies to really have a unified customer relationship management system. And today I can go put in a credit card and sign up for HubSpot or Salesforce or Dynamics right off the shelf. There’s this pull to off the shelf that happens in technology which leads to the middle market and small businesses being able to take advantage of what used to be incredibly expensive technology. And that’s actually what we’re seeing in the AI space is it’s driving from, I no longer need a hundred million to approach this problem. I can actually apply this for 20 bucks a month.

Chris (04:31):

Yeah, that’s a great observation. And and yeah, it’s so true that it becomes, I guess better efficient and and more economical. Right? Yeah. Each time I guess as technology is with us and develops longer. That’s a great kind of segue. I want to just kind of start with what are some of the key factors, uh, a business should consider when evaluating their readiness for adopting AI into the business?

Devlin (04:57):

Interesting. For adopting AI into the business readiness is mostly psychological because there are pieces in the business today that you can do better. We break this into kind of three parts when we talk to business leaders about this. One is how do you do your job much more effectively? Right? What’s the super human version of Chris? Right? There’s AI tools to make that happen. Like I’m a very well augmented human. I have tools that analyze my notes and make sure that I don’t forget things. I’ve got tools that keep reminders and stuff on my personal network. Now they’re not spamming my friends with like text messages to buy things, but it’s going, Hey, you haven’t talked to Bob in two months. Here’s what you talked about the last time. So I can reach out to Bob like, Hey man, we haven’t caught up. How’s your wife doing?

Devlin (05:48):

How’s your son doing? Like those kind of things. That’s the superhuman version of me. ’cause I want to stay connected with my friends, right. Just bad at it. And so it covers that gap for me. So that’s the first part is like that personal productivity side, which is mainly just a resistance to change that you’d see in any technology adoption. It’s psychological, organizationally, people have tied their identity to the work they do. And so changing that means a like an existential crisis sometimes, right? Sure. Think about a bank teller when the ATM came out right now, we still employ a lot of bank tellers, but their jobs drastically changed. It’s that moment where we’re not gonna get rid of a bunch of humans and have robots doing those jobs. What we’re gonna do is change the job of the human to guiding the controlling and managing the robots.

Chris (06:37):

I think that’s a important point to kind of reemphasize for the listeners. ’cause I think so much that’s out there. You you have, you see these news headlines and articles. I think people think robots are gonna take over the world. Mm-Hmm. <affirmative>. And I think the point you just made that that’s not the case, but the role the human will play will adapt and change. And while that sounds scary in a vacuum Mm-Hmm. <affirmative>, if you actually take a moment and look back, that’s what’s happened throughout kind of our evolution, especially in the, in the, in the industrial world and in the, in the business world in the United States. Right. Jobs have evolved and changed over time. And I’ve heard you say this before, so this is nothing different. I want you to dig in a little deeper on that to help the listeners understand and maybe some historical points to uh, compare to so they, you know, it makes a little more tangible.

Devlin (07:28):

Absolutely. So think about the way we did accounting before the PC was invented, right? So before the Apple two we’re talking in the 1970s right before computing devices were in everybody’s office on everybody’s desk, right. The way we did accounting was we managed the book and you wrote entries in and you had somebody checking the math and you had the, you know, 10 keys sitting there with the stream of numbers coming out of it. Right. And your accounting department was massively larger than it is today. To be able to accomplish that, it had to be right. Which was a big overhead for a business to bear, right? And you had these big accounting firms who would help with economies of scale or whatever, but like that was really the ball game, right? Right. And it took a long time to like close out the books and do tax audits and those kind of things.

Devlin (08:19):

Now fast forward to the introduction of broad computing power that sped up that process. We still have accountants, we still have bookkeepers, but in most businesses you can close the books on a month in 10 days, 30 days if you’re, you know, got a lot of moving parts. It’s not, hey we just closed January and June. Right? Right. Accrued accounting became far more prevalent. We had less financial fraud overall. Right. The, the stories about it happen more often, but like we have less by volume. Right? And we’re actually getting more insights out of that because it’s no longer just tracking all the pieces but going, Hey, did you notice last month you had increased expenditures in this area without the increased revenue tied to them? And so we get business insights on top of what we used to get was just transactions. We not only have lower accounting costs, but we then have better outcomes from it.

Devlin (09:14):

AI’s gonna do something similar from a business perspective. It’s going to allow us to get, it’s going to allow us to get better outcomes or lower our costs and give us pricing power in the market. Because all technology is labor compression, right? What a welder by hand used to take hours to do on the original factory floors and you know, structural integrity of the original cars that we were rolling across an assembly line, right? Think 1930s, 1940s. We now have robotic welders who can do in 15 or 30 seconds with far more precision with less human injury. Right now the quality checking the x-ray and all that is still reviewed by a human to make sure that weld is solid. And even that we’re automating some of, but like that evolution allowed us to produce stronger, faster, cheaper, safer cars. I think we’re in that space where AI is largely going to be applied to the problems that are on the edges of humans do a lot of it, but we’re not very good at, ’cause like our bookkeepers, right? There’s that

Chris (10:21):

Whole notion of human error. Yeah. Not that there won’t be computer error as well, right? Oh

Devlin (10:26):

Yeah. And

Chris (10:27):

So you kind of, that’s where the check and balance comes

Devlin (10:29):

In and the idea of technology is just gonna solve everything hopefully as a civilization. We’ve moved past, right? The 1970s to today. I use the 1970s ’cause that was kind of the broad evolution of, of available computing, right? To today, every new technology has created new problems. A a joke with our team that yesterday’s solutions caused today’s problems. And that’s a good thing because one, we always have problem to solve and two, we don’t have yesterday’s problems. So AI being introduced is going to create things like we now need to manage bias, the computer error, right? That’s not something we do today very well when we talk about humans, right? Right. Like how do you manage bias at scale in a thousand person company is like, alright, HR in an army of training but with a computer you can actually try to start tilting at some of these things. Now does that mean we’re gonna do it? Well we’re gonna do it better than we do it today, right? Probably we’re gonna do it wrong and have to create tomorrow’s problems.

Chris (11:29):

Yeah, I love that perspective. So what, what are some of the obstacles or pitfalls that you’ve seen that businesses encounter when they’re trying to implement technology and maybe even speci obviously specific to AI technology?

Devlin (11:43):

So there are two, one of them is perfectly valid and it’s gonna be some learning that we have to overcome. And I’m gonna start with that one. The belief that I have to spend a ton of time and money to correct my data, right? Because traditionally over the last 20 years you’ve had data engineering and data warehousing and data lakes and like you had to clean it and curate it and do all this work. That belief is a little antiquated, right? You can bring in raw data and then actually use a lot of these automated systems, AI systems to clean it up with you so that the labor of that is way less scary. Now that’s the pitfall most people fall into is, oh I gotta get my data cleaned up before I get any value. Hmm. Okay. And so that ends up raising the price tag of going after these technologies and ultimately keeps companies from getting some of that benefit.

Devlin (12:30):

’cause they don’t wanna pay that cost and then makes sense. The second pitfall is building your own. And what I mean by building your own is every business has unique challenges and they have their particular flavors, right? It’s why where SAP works for one, but you know, Acumatica would be better for somebody else as an ERP system, but you don’t have to reinvent the wheel. And we keep doing that, right? I was just talking to a friend of mine, Houston based company yesterday, 500 million in revenue and we’re like talking about one of their AI initiatives. It wasted $6 million, didn’t get anything out of it. Wow. And we were talking about it, I’m like, you can do that with almost off the shelf tools. Everything you guys were trying to accomplish in about four months for about half a million. And the difference is that they tried to reinvent all the wheels. We don’t need to do that. Just like you’re not gonna build your own email system, right? You don’t need to build your own baseline architecture for a large language model. Use one of the foundational ones that’s off the shelf and you don’t waste a lot of that time and effort

Chris (13:33):

And that gets you. Yeah. Good way to get started. Yeah. It may evolve from there.

Devlin (13:37):

May evolve from there. You may hit a problem where you do need to build your own, I kind of, the rule of thumb I use is if your IT budget doesn’t start with a B, you’re probably not building your own machine learning models. <laugh>.

Chris (13:48):

<laugh>. So that raises a, I mean, I guess a good question and that would be, how can companies distinguish between an AI solution that actually what is gonna offer value? Mm-Hmm. <affirmative> real value versus just a company following the hype, right? And being misguided by the solution. Maybe they, they choose

Devlin (14:09):

Fall in love with solving the problem, not the tools. So if, let’s take my company, right? We spend a lot of time trying to solve one big problem. That big problem was knowledge. We grow the VIA acquisition, we’ve done 14 acquisitions in 14 years and we always create knowledge silos. And so when we bring in somebody, our current team doesn’t know their stories for like selling their skill sets, what they’re good at, those kind of things. And they don’t know all of our stories. And so we had this big knowledge silo gap problem, right? Right. Now ultimately what that means is when a customer goes, Hey do you do X regardless of what X is, they’re gonna say no because they don’t know the stories. Now how do I overcome this? I could do training, right? But then I gotta do that training every time we acquire a company and we’re doing like we’re aiming for two to four acquisitions a year, which means that’s not a sustainable thing because of the the labor cost, right? It’s like okay, well maybe I allow the silos to continue and just accept that’s part and parcel to the business. It’s possible,

Chris (15:19):

Possible but you’re gonna miss out a ton of opportunity.

Devlin (15:21):

Exactly. Or we take all their stories, they’re case studies, they’re customer testimonials. We loaded them into what we call echo, which is a AI enabled chat bot and it literally reads SharePoint, right? It did. It’s not like, it’s not parsing data. There’s no big data engineering effort. It’s loading word documents, PDFs, all this off SharePoint and they just chat with it. And I go, Hey, have we done a deal with a major energy company? And it goes, yes, here are the three that are most relevant to you. And then it embeds the PDF and goes, and here’s where you find more details so that the sales team on a sales call can have echo up on another window like, Hey, have we ever done that? And it goes, yes, in this office, here’s the people to reach out to again. And that level of knowledge access would’ve cost us thousands of hours of training, right? And so it’s that type of thing. Focus on the problem. Where do you have pain and where are you wasting hours? You don’t actually care as a business owner unless you’re selling AI as a product, right? You don’t actually care if it’s an AI solution, an automation solution, or just really clever software. You just want the problem solved. And by not falling in love with the tool but falling in love with solving the problem, you focus on the right thing. Because the value add, the ROI is all about the problem, not about the tool.

Chris (16:43):

Look, that makes sense. It’s easy to remember for sure. And I mean I think you’re right. I think most business owners agree, I just need this problem solved effectively and efficiently. Yeah.

Devlin (16:52):

By the way, you find these problems by going, what would it take for me to five x my business today? The things that immediately pop to mind, you’re like, oh well this would break and this would break and this would break and this would break. That’s your list, right? For me it’s like, well I need five times the as many account managers and my accounting staff’s gotta grow and I’d need better hiring. Like that’s my list. Do I need five times as many account managers or do I need to help automate a lot of the account management and administra to make them more effective? How do I upskill and get my recruiters leveraging AI sorting and those kind of things to pull more people into the pipeline, right? Yeah. Like that’s my list by simply going, what would it take to get bigger by like a big number If five x isn’t scary enough, tech has zero on there,

Chris (17:38):

<laugh>, that definitely makes it scary. So let’s talk about, uh, you know, there’s a lot that’s been written and something we’re doing kind of here, you know, here ourselves and that’s, you know, with AI out there, what are best practices that businesses should be considering around policies for using, evaluating, adapting, you know, AI technology in the business? You know, there’s a lot that, you know, I think it’s probably best practice that there, there should one, yes, you should have a policy, but anything you can kind of, you know, guide the listeners on, on, on those issues around a, a competent and well thought out AI policy.

Devlin (18:20):

So it, it’s got a few pieces. Number one, data privacy needs to be forefront in that conversation. Primarily to protect your business and to protect your uh, competitive advantage, right? So if your AI usage or acceptable usage policy doesn’t include something about how data privacy should be evaluated, that’s a big gap. Now your opinions about data privacy are gonna be your company’s opinions, but those tools that are cheap and freely available today are largely cheap and freely available so that they can use your data to train a better tool. Is that okay with you? Right. Some people are like, yeah, it doesn’t matter. And some people are like, no, I absolutely can never allow this data outta my control. At which point you gotta choose different tools, right? So data privacy’s number one,

Chris (19:10):

Dude, to that point you may be aware of this and uh, I recently wrote a little um, blurb on it, but you have the New York Times lawsuit Yeah. Saying that all, you know, trained

Devlin (19:21):

On copyrighted material. Yeah.

Chris (19:23):

Trained on copyrighted material. So that’s kind of, to me somewhat of akin to data security and privacy. Oh absolutely. And and that’s a whole, it’s a whole other issue about copywriting and licensing around information. So maybe we can talk about that in a minute. Mm-Hmm <affirmative>. But let’s keep on the data or AI kind of policies. And so you said most important thing data privacy. What’s next?

Devlin (19:42):

Second is vendor resiliency. Now this is going to sound a little tough to like the indie developers who are trying to launch their product, but last year in the US there were 6,000 plus tools launched on the AI hype wave. Now the punchline to that story is over 4,000 of have already failed. Already had to either pivot or gone outta business vendor resiliency. If you’re gonna start pulling these into your business, evaluate the vendor, are they gonna survive long enough to be valuable to you? Or do you now have a broken tool that’s no longer being accessible that you’ve woven into your business that is gonna drive you towards some of the bigger vendors, the ones that have been around for a while. And as it kind of should if you’re weaving it into your ops now for experimentation use, use the little players like that makes sense to me.

Devlin (20:33):

But when you’re talking about a broad policy vendor, resiliency is gonna be a big thing. The other side of vendor resiliency is how are they going to indemnify you from the inevitable lawsuits in this space? Right? Right. Microsoft, Google, Amazon have all said if you’re using our tools inside the license agreement, there’s indemnity. Right? That’s a pretty big shield, right? Yeah. Microsoft actually said that they would, if you’re using their AI services, they would protect you and defend and pay a settlement if one ends up happening for copyright infringement. So like the Times article thing won’t hit the consumers of those AI tools. Microsoft has stood in front of it and said we’re good. That’s a big shield. It’s big. Now if you’re on a small to mid-market software player, can you put up a shield? Right? Right. To your customers as a customer I need to start caring about this.

Devlin (21:25):

And then lastly in that policy, some centralized knowledge repository, some centralized store. Because what we found is everybody’s play, everybody’s trying, experimenting using these tools. They’re wiring in their favorite one. I do this almost on a daily basis. I kick out unapproved tools from meetings that somebody like wired up, like a meeting transcriber, listener bot, and I kick them outta meetings and send a note to whoever did it. I’m like, just to be clear, not approved <laugh>. Right? Here’s the approved one, don’t use that one. And everybody’s just so expense control in some kind of central review. It doesn’t have to be heavy handed. Ours is literally just a let us know when you’re experimenting so we can check in on the experiment ’cause it might be something we wanna share. Yeah. Right. But some kind of central, you gotta have centrals, right? Yeah. Because a lot of these are SaaS based. A lot of ’em are out kind of in the ethos of like knowledge tools, like note taking tools that I use. There would be no way for improving to know that its IP is in that tool if I didn’t tell them. Yeah. And so you’ve gotta, you’ve gotta have kind of a reporting and honor system for the employees to tell you where your data and vendors live.

Chris (22:33):

So one of the things that I know that improving and the leadership at improving, which includes you, you’ve done a great job of you know, building a culture in a company that embraces technology, embraces innovation. What, what can you share about that experience and that journey at improving to maybe help others understand, you know, how they may be able to do the same thing?

Devlin (23:00):

Absolutely. So I, I have the oddity of looking at this kind of, if I look back down the mountain, it seems like it’s a long way but all I can see is looking up the mountain and it’s still seems insurmountable, right? So I guess first would be the journey doesn’t end. Don’t let the size of the mountain scare you. Just take a step. Yeah. Right? For us, we have a lot of like growth and planning kind of baked into our employee management model and we call it path, that’s our employee growth systems. Okay? And part of that is maintaining your marketable job skills. Literally what we call hard skills, right? The marketability of a person to maintain because there’s this kind of natural degradation if I stop learning, I become less and less valuable ’cause the market moves ahead of me, right? And so recognizing that truth and going, okay, what are you doing this quarter to grow with technologies?

Devlin (23:56):

Then we go, okay, what new tech are you learning or playing with or experimenting with this quarter? What we have found is as long as there’s a vehicle for them to share that back to the company and make an impact, people are highly engaged. If it is just playing over here and then they have to come back over here and do the same thing they’ve been doing for 15 years, less engagement. And so creating the vehicle in which their experiments can have long lasting impact on the business created a lot of engagement. And then the other side of it is, we recognized a while ago that if you are not growing, you are dying as a business. Amen. And that’s true for all of our people. It’s what we call the plateau of slow death. Like you’ve just decided to coast that will have an accelerating decline in your value to the business.

Devlin (24:45):

How do we help people stay on a plateau of slow growth where they’re still incrementally investing? Sure. Now for us that’s five hours a week because we’re a technology company, it moves quick, right? That might not need to five hours a week for somebody in manufacturing, distribution, et cetera. But probably an hour a week just reading like there’s the Wall Street Journal podcast, right? There’s this podcast that’s phenomenal for staying abreast of what’s happening. Like consume an hour a week of new information for you and your team and you’d be amazed at what doing that week after week will do to the business. Like it just accelerates. And it sounds very simple. It was one of the first steps we took.

Chris (25:27):

You know that the dedication to being intentional about the learning and self-improvement on a weekly basis I think is amazing That I mean any business, right?

Devlin (25:37):

I believe so. I’m amazed how many business owners and friends I have that work in businesses and they’re so busy that they’re too busy to survive.

Chris (25:48):

I’ve said it here in this firm before and yet to repeat it and we’re all can be victim of it and guilty of it. But busy can’t be an excuse. Yeah. I’m too busy to do X when, when, when X is strategic work on how to improve the the company. Mm-Hmm. <affirmative> or yourself busy can’t be an excuse. ’cause if it is, then nothing will ever get done. ’cause you always feel too busy. Yeah. Right.

Devlin (26:13):

And so I pay for a lot of tools. I’m a well augmented human right? <laugh>, one of those tools is summaries of like business articles and books and all that. And so while I was sitting here waiting for this conversation, I was reading one of those and it’s that overarching approach of like how am I getting value out of those moments? Like when a meeting wraps up early, do you sigh in relief and like walk out and waste 10 minutes? Maybe that’s good recovery and you need that for emotional balance. Okay. But is it intentional? Did you go, Hey, you know what, I need emotional balance and chose that. Or did you go, I got 10 minutes, I’m gonna read that book summary or I’m gonna read an article or I’m gonna check out what’s on HP j’s innovation stuff like those questions,

Chris (26:59):

Right?

Devlin (27:00):

Just making the consumption of data an option mentally for all of us. This is why I say like a lot of our barriers are psychological. ’cause the technology’s actually not scary once you start exploring it. Sure. It’s only scary when it’s like Skynet and Terminator from the movies. Right? And so then it’s scary and that makes sense. But

Chris (27:19):

What, it’s interesting because it just occurred to me, let’s bring this full circle from the beginning of the conversation, right? What you’re talking about and recommending people be intentional about that self-learning that discipline around self-learning and improvement is really gonna be essential as new technologies come online. Because we, you said earlier, right, technology’s gonna force the worker to adapt and the only way you can adapt is by continuing to learn. So to be successful alongside technology like ai, it’s gonna be essential.

Devlin (27:50):

This is actually, I’m a future optimist. And what I mean by that is I think that technology elevates humanity, right? Very similar to capitalism, elevating humanity. It has made life better. It’s increased longevity, it’s done a lot of things. Now that’s not to say technology’s perfect and we live in a utopia like, but it is, technology elevates us, but it makes us do the harder version of life, right? Technology allows us to play life on hard mode, right? So like social media, I can doom scroll forever, which means I have to own the choice, right? Before that technology enabled me to stay connected with all my friends. I didn’t have to make that choice, right? Right. AI, by taking a lot of the complexity, a lot of the time consuming tasks off my plate means that all that’s left are the difficult tasks. It’s the hard mode tasks.

Devlin (28:43):

And getting really good at the hard mode tasks is the value creation in the future. It’s, hey, I gotta go write this software. The writing of the software, the actual typing is gonna get much easier. Just like accounting, just like bookkeeping, just like going through and like automatic scanning of discovery documents in the legal space. Sure. Used to be very time consuming now is being accelerated by AI and automation. So now then the hard part is understanding what software I need to write and why, understanding what those transactions mean to the business and why understanding what in that discovery is pertinent, important and relevant to the story I’m telling. Right? Like all the hard tasks get left, the difficult tasks. Sure. Because those are the ones AI is really bad at

Chris (29:29):

<laugh>. Right? Thankfully for now. So before we wrap this up, I definitely wanna ask you your thoughts on regulation and you know, what you think congress should or shouldn’t do around putting some regulations in the AI space.

Devlin (29:48):

So AI regulation is coming, like that’s going to be the case. Uh, any sufficiently developed technology ends up getting regulated at some point, right? Should do Transparency to empower a educated consumer is phenomenal. Right? Like stating if you’ve baked an ethical bias or a political or religious bias into a model so that the people who are using it can choose. Right? That makes sense to you realize

Chris (30:14):

That the the output is

Devlin (30:16):

Tilted.

Chris (30:17):

Yeah. Has some bias

Devlin (30:18):

In some way. Right? That’s great to know as a consumer, right? And luckily that’s where a lot of the early regulations in this space are tilting. The shouldn’t do side of it is dangerously close to that, which is then publish how you built the model to prove that statement. Which is a lot like saying give everybody your proprietary trade secrets, right? There’s a reason that OpenAI stopped publishing a lot of their, and here’s exactly how we built it. And that’s because a whole bunch of other companies took that research that they had poured tens of billions of dollars into and created additional models that were almost identical in performance. Right? Right. Now they’re different and they were developed by different teams and all that. But like there’s a reason it went from we have one major version of this to we now have 15 publicly available commercial models, right? That gets dangerous. Sure. When you start regulating people to destroying their business. And so that’s the line I’m hoping we walk the stifled innovation that happens on that second one we’re seeing in the EU when they passed the, and here’s all the restrictions of ai. You have to publish your training set and your methodology and all this stuff. It’s like awesome. And there was a mass exodus of AI companies from that area. Like yeah. They’re like, nope, we are not going to participate if you require us to kill ourselves. Right? And so no

Chris (31:44):

One’s gonna invest time and money in something that they can’t then have a return on.

Devlin (31:49):

I mean if you look at the open AI side of it, this is tens of billions of dollars in decades of research and development and work to make this happen. Imagine if you then had a law that said, and you have to enable your competitor who doesn’t have that cost to then rapidly get to the same point for a 10th. Right? And so there’s a balance between you wanna democratize some of it, you’ve gotta balance the investment side of it. And if you go too far, which I believe personal belief that the EU did, it just causes a significant drop in investment. Yeah.

Chris (32:24):

So, you know, kind of with that in mind, where do you kind of foresee the evolution of AI over the next five to 10 years?

Devlin (32:35):

We have largely looked at AI as the Jetsons robot or Terminator where it’s this one thing that is omni powerful, omni capable, right? Omnipresent. I don’t believe that’s where we’re going the best minds in this space of which I get to talk to. I’m not one of, I beg

Chris (32:56):

To Differnce but go ahead.

Devlin (32:58):

They would tell you that it will be a cloud of things. Like imagine that you’re surrounded by Chris’s swarm of empowering bots. You’ve got a bot that helps you manage your schedule. You’ve got a bot that helps you take notes from a meeting without having to like jot them down and all of these save you 10, 15, 20 minutes an hour and a half a day. That means somehow Chris is doing 50 hours of work in this eight hour day because you’ve got this super human capability that’s empowered by all of these things. That’s where we’re headed. I just saw, I was playing around with a, a toolkit that there’s been a lot of hype over the last few weeks is the video generator pika, it’s like mid journey or dolly or stable diffusion for images, but does videos. Okay. Like cinemagraphic grade quality. The problem is you have to also get really good at understanding camera movements and placement and blocking and all these things that directors have known for decades.

Devlin (33:59):

And so it’s not built for this average consumer. It’s built for making folks with that knowledge massively more successful. Right. Being able to go and here’s a rough of my movie idea, right? Here’s a short of my movie idea for a thousand dollars, not 70, right? Right. Like that will accelerate the creative space in movie making, but it’s not gonna get rid of a need for that knowledge base. Same thing’s true with like geophysics and well planning and the energy space. How do we conceptualize all of this and make a human significantly more powerful? So this team that includes a drilling engineer, geophysicist and all this can plan wells and make financial analysis and all that in days not years. Right. That acceleration is where we’re gonna see it and we’re gonna see it through these kind of micro enhancements. Okay. I carry several of them with me.

Devlin (34:54):

I’ve got a note taking system that maps all of the connected topics that I’ve been researching and digging into and it’s wicked fun and crazy. But I built a chat system on it that runs on my laptop. And so I can ask questions in my notes. I’m like, Hey, in my last Vistage meeting there was a speaker who talked about this, what were the key takeaways? And he goes, here’s the notes, here are the key takeaways. It’s that, that’s crazy kind of empowerment. ’cause human memory is fallible. And so how many of us have wished, like, I wish I had a better memory, it doesn’t have to live in my head.

Chris (35:27):

Yeah. Kinda like what, there was something five minutes ago I said I needed to do and now I can’t remember what it is. Yeah. I mean how often does that happen?

Devlin (35:33):

I carry around Todoist and Todoist integrates with it. It, and so at the end of the day right before I typically leave the office, I get a reminder set from the automation I hooked up to it. Now it looks at my calendar and goes, where’s the right point to remind Devlin to do those things before the end of the day. And so like folks literally like, I don’t know how you do this. I’m like, I don’t, I’m very well augmented <laugh> that.

Chris (35:58):

Yeah, you’ve said that more than once. I know you mean it very well augmented. So I was gonna ask you what some of your favorite AI tools are. I think you’ve shared them just now, but maybe just a, a quick summary of maybe three or four of your favorite tools for the listeners who are trying to frantically take notes.

Devlin (36:15):

<laugh>. So I, for network management, so my personal network management, uh, I use Clay Earth, you literally go to Clay Earth is the URL. Uh, I think it’s phenomenal and I use that to manage my network. It does not spam or reach out to, it just helps me reach out and stay connected. The kind of in my business version of that one is Dynamics. We use sales co-pilot for Dynamics. Einstein in Salesforce does the same thing. So in the business we use a different one because different needs, right? Sure. For note taking, I use obsidian. You can use Evernote or OneNote in this same thing and it’ll do a lot of the same AI enablement through plugins and those kind of things.

Chris (37:00):

And then you mentioned one about just the minute the reminder.

Devlin (37:03):

Mm-Hmm. <affirmative>. So I use Todoist and Power Automate. I’ve combined those two tools. So if you’re in the Microsoft stack, right? You use Office 365 or Microsoft 365, you have access to this one already. I didn’t know it. And so you can go to make power.com, it’s a Microsoft tool. You’ll log in with your Microsoft thing and you can describe what you wanted to do. I did this yesterday. I was presenting to a group of CEOs on and this topic and I was like, take the notes, my handwritten notes that I emailed a picture of myself, take the notes I emailed a picture of to myself, parse them, put the text in my notebook, scan it for action items and put those action items in to Doist. Literally that’s all I described. And it goes okay. And it’s got this massive library of these tiny little tasks and it pulls ’em all together and goes, here’s the automation that will do that.

Devlin (37:52):

And it writes the rough draft, the prototype of the automation for you, and you just click, alright, create. And it goes, this is the permissions I’m gonna need. Are you good with that? Yep. Go. And it’s there and it’s running. I had to write no code. I had to wire nothing together. It just did it. Hmm. And so we’re using this for like back office automation all the time. Like, hey, take this output of our financial system, slice it, dice it in this way, and it writes the pivot table creation and all that in Excel. Like that’s might be half an hour to 45 minutes that I just saved our business partner in accounting. Yeah. And so it’s a lot of these tiny little bots.

Chris (38:30):

Wow. So when you think about AI and how it could be disruptive to industry, what are maybe one of the top two industries you think it’s gonna be the most dis disruptive to?

Devlin (38:45):

So oddly, I think logistics, supply chain and manufacturing are probably those two. One, they’ve typically been underinvested in technology and so there’s a lot of low hanging fruit. But two, it gives pricing power. Like imagine that I can compress the labor to accomplish a task. I can now outprice my competitors who aren’t doing that. And in those two spaces where they’re very commoditized, prices came, sure. If you can be 3% cheaper while maintaining your margins, that’s the ball game and you can just put people out of business. So I think those two are gonna have massive kind of immediate six to 18 month impact. If you look slightly beyond that, the construction space is huge in this AP great, you know, Houston story here has a, a robot called Dusty that they helped to develop. Okay. It takes the construction documents for a high rise and it prints the lay down onto the concrete.

Devlin (39:45):

It uses basically a Roomba guided by ai. It parses the construction documents and in color coded paint prints the lay and it reduces the labor of manual labor, construction labor of building out that building because they don’t have to snap chalk lines and measure everything and everything else. They just follow the color coded thing. Wow. Which also means they need lower skilled labor, which is a labor savings. Sure. Right. And so these things are changing the game and changing the pricing power on a lot of these fixed bid contracts. And so you, you see some interesting spaces where traditionally non-technology based business has a lot of low hanging fruit, like FinTech and financial services has been heavily invested in technology, less low hanging fruit there. Sure. Makes sense. So the disruptive stuff I think is gonna be in those three over the next few years. Okay.

Chris (40:37):

Devlin, this has been such an interesting and, and fun conversations. Thank you for, for doing that. I want to just turn, uh, just uh, to a little bit of the fun side of things when I have a guest in and what was your first job? I guess you told today you were programming, but was that, were you getting paid

Devlin (40:51):

To do it? No. So my first job, there was a pool near our house and I love, like there was a cherry seven up, like you got the bottle cap thing and you could earn points and order stuff like that moment in time. And I, my parents, like, I didn’t have enough allowance to like as much cherry seven up as I wanted. Right. And so I talked to the owner of the pool that we were a member of near our house into letting me like do the chlorine in the cleanup and scrub the pool for cash when I was 12. Like this was definitely not legal <laugh>. And it’s like I’m moving buckets of chlorine and doing all this stuff while my friends are playing at the pool because I was earning $5 a day that I could spend on Cherry seven up <laugh>

Chris (41:35):

On Brew from an early age. Right. I love it.

Devlin (41:38):

So hopefully I don’t get anybody in trouble. I’m not giving you names of Fools <laugh>. Okay. Don’t

Chris (41:41):

Do that. So what do you prefer, TexMex or barbecue?

Devlin (41:44):

Oh, barbecue. Hands down. Yeah. I have a massive pit smoker in my backyard. Like Oh,

Chris (41:50):

For real.

Devlin (41:50):

Okay. Yeah. So we throw a barbecue in Dallas every year for 4th of July, feed like 400 people. We throw one here at our office for Labor Day Memorial Day. Which one’s at the end of the summer? Labor Day. Labor Day for Labor Day feed, like 250 folks. Like Oh, serious. I’m bigger than re serious.

Chris (42:06):

All right. I love it. And what do you like to do for fun when you’re not out speaking on ai?

Devlin (42:10):

So I play a lot of golf with my wife. Nice. And she kicks my butt or I like video games and stuff like that. And so my brother and I play a lot of video games.

Chris (42:19):

Very good. Well, like I said, Devlin, you know, I love the conversations we’ve had in the past. What you shared today was so enlightening and I, I, I know will be valuable to those listening. And like I said, they probably like me took a lot of notes that they’ll try to implement into their daily life. So thanks again for being

Devlin (42:35):

Here. Oh, thank you. This was great.

Chris (42:39):

And there we have it. Another great episode. Don’t forget to check out the show notes at boyarmiller.com/podcast and you can find out more about all the ways our firm can help you at boyarmiller.com. That’s it for this episode. Have a great week and we’ll talk to you next time.