Look, most companies are asking the wrong questions about AI. They're treating it like a search engine, when in reality, it's rewriting the rules of how businesses are built. Welcome to Using AI at Work. I'm your host, Chris Daigle. Each week, we'll be learning how today's business owners, entrepreneurs, and ambitious professionals are getting more done with smart use of tomorrow's tech. Let's get started. In today's episode, I'm joined by Jim Curry, co-founder of Build Group and the former GM at Rackspace, and one of the key people behind launching OpenStack. He spent 25 years at the front of every major tech wave, and now he's here to unpack what this AI wave really means for you. When you listen to today's episode, you'll discover why the real winners in AI won't be startups, but existing workflow companies with data and expertise. How roles like product management and engineering are already blending, and what that means for your org chart. And we'll also discuss the surprising way Jim is using AI behind the scenes to re-architect investment workflows, boost targeting accuracy, and even rethink the future of SDRs. This isn't just theory, Jim's insights come from the trenches. So if you want to see where your business could be in the next 12 months, don't miss this episode. Right now, every business leader is asking the same question. What are we going to do about AI? If this is you, ChiefAIOfficer.com has the answer. We give you a simple path forward where we provide executive and team training so your people know exactly how to safely use generative AI in their day-to-day. We also manage the deployment and implementation to make sure tools actually get adopted and deliver results. And we'll also guide company-wide transformation so AI becomes part of your operating system, not just another shiny object. The companies that act now will increase productivity, cut costs, and grow faster than their competitors. Those that wait will get left behind. So if you want to make AI work in your business, visit ChiefaiOfficer.com and see how we're helping companies of all sizes finally get results from AI. Hey everybody, welcome to the next episode of Using AI at Work Today. I've got fellow Austinite and uh AI enthusiast Jim Curry on the episode. And Jim, I'm going to do a brief introduction, but if I miss anything, I'd love for you to take a second and just kind of um point out some things. And I got a list of great questions here that I want to ask you today. But Jim is the co-founder of Build Group, which is a growth stage investment firm focused on long-term company building. Uh, and he also was a big part of the growth at Rackspace when he was general manager there, from startup to, I guess they're now a $1.5 billion industry leader when it comes to uh cloud space. And also, Jim, you know, we we haven't, we might not get into it, but this this role is interesting about the launching of OpenStack, one of the foundational open source cloud infrastructure projects. Um, but before we get started in our conversation, is there anything else that you'd like to share before we uh we dive in?
SPEAKER_01Yeah, no, that's that's pretty much my background. I've been uh, you know, I started my career uh actually in uh investment banking, which is relatively funny because uh my opinion of of bankers today uh is not is not uh best, uh although they do play an important role in our economy. But I did uh but yes, I spent a lot of time in tech, and I think I've been really blessed to be through some really major uh waves uh over the last 25 years at companies are at the front end of it with uh you know, obviously the SaaS boom and the cloud boom, the open source boom, and now the AI boom. And you know, for me, it's uh this is why I do what I do and why I love technology so much. It's just it's been a fun 25 years.
SPEAKER_00Yeah, we're about to get into the fun part with AI. You know, before uh for the listeners, before we started the episode, uh Jim and I just caught up a little bit. He's as I mentioned, our headquarters is in in Austin. He's been Austin-based for 20, well, more than that, I guess, longer than I have. And certainly being in Austin has given you some optics into the, I don't know, tech for sure. But one of the the comments that I I saw that was pretty interesting to me is this idea that AI is following kind of the same pattern as open source and the early internet. And um, I just maybe if you could address some of those parallels that you're seeing now.
SPEAKER_01Yeah, I think uh I I call these infrastructure waves. That may or not be the right term, but you know, what I really mean by them is they're waves that lift all boats, right? I think you have an opportunity to really apply uh this against a wide range of companies. Some are better positioned than others to take advantage of it. And I think that's one of the things that I've learned and trying to take advantage of in this current wave. So if you think about, you know, if you think about cloud infrastructure as an example, the ability to programmatically spin up and manage infrastructure substantially changed how people thought uh not only about how to uh run applications, but how to build them. You know, it was very different from the client-server world. And that really led to uh, in a lot of ways, a boom in uh application development uh because a whole lot of subordinate had to go into managing that infrastructure layer uh and uh and more had to go into just building the application. And I think that was a you know real important innovation change. Some companies were better prepared for it than others uh and how they adopted it. Um, but ultimately everyone did have to adopt it in order to compete. Open source is no different. You know, open source is one of those things, you know. I had a very similar moment I like to talk about in around 2000, late 2010, 2011, when I got an email from Fidelity uh after we launched OpenStack, invited me up and they sent me this document where they laid out their architecture, which was all closed source, and then they laid out a theoretical open source layout next to it, right? So this was replacing Oracle with, you know, name your database, you know, replacing Microsoft uh and going all analytics uh and laying all that stuff out. There were critical gaps because at that time there weren't enough open source projects, but it really kind of said, oh, this is interesting. All of a sudden, open source, which is something that developers at startups might have played around with, is now in the enterprise in a big way. And I think that obviously open source, everyone knows what it did in terms of cost, but in terms of innovation, I mean, think about where we are. I always use the database world as an example. You could find any database type you need for any type of workload. They all exist now, right? Massive innovation change. Uh, also, it introduced the concept of cross-collaboration, right? Working together uh across organizations to build uh software. And one of the things I always find I like to tell people is about year two or three of the OpenStack project, when we all of a sudden had thousands of people come to our conference. I always open up our conference showing contributions from different companies and groups. But the thing that always I was so surprised by was I showed was the number one contributor geographically was the United States. The number two was China. And within that, the number one contributor were U.S. affiliated government agencies, and the second was Chinese affiliated government agencies. Yeah. And you're just like, where in the world does this happen? Like it's just it was an amazing thing. So a lot of innovation uh driving there. Now, AI, in a lot of ways, to me, is the evolution of the data science boom. Uh, it's not quite right to say that, but uh, we initially started Build Group to invest in data-enabled SaaS companies. So those companies who understood that the workflows they were building generated unique and valuable data insights to their users or to the machines that depended on them, and were really focused on driving insights out of that. Um, and you know, a lot of folks didn't really get that. Data science wasn't something that necessarily got broadly adopted by everybody early on, but that's where we focused. That happened to work really well with AI. And early on in the AI days, we're able to build some really interesting companies. But now, since then, AI has gotten so simple to adopt and use, uh, that now everyone's able to take advantage of it. And it's across the board. You can use it in your product, you can use it in your operations, right? And your operations, we all know what's going on with the cursors and the back offices of the world, but the front end, what you can do on the front end is amazing. Um, and I have examples we can talk about that we've done within our portfolio, but is also lifting all boats in terms of innovation uh and in terms of uh what's being done. But you know, the the thing I find also really interesting about it is in some ways, uh it's turning people who've never been developers before into developers. Yes. And most importantly, it's turning them all into data scientists. There's no barrier between your ability to interact with data real time and get insights like there was before, where you had to have a data scientist to get the data properly conditioned, get an artifact built for you, send it to you. Then you had the mod, you know, make changes to it, send it back to you don't have to do that as much anymore. And it's really interesting what that's going to empower.
SPEAKER_00You know, you mentioned a couple of tools that a lot of the listeners may be familiar with, but there was a point of view that I found in the research that I think that some of people may find um, I don't know, they might plan out, disagree with, and your position that the real winners aren't AI native startups, but they're the existing workflow SaaS companies. Do you mind kind of sharing your perspective on that?
SPEAKER_01Yeah, of course there's going to be native AI winners. Uh, you know, big, large uh LLM providers, or, you know, we already know that that's the case. And there will be agentic AI companies that win. Um, the way I think about it though is in the end, AI is valuable when you have two things. The first is data, right? And uh ideally that data is proprietary. If the data is not proprietary, anyone can build around it. And the second is expertise, um, really good vertical or horizontal expertise or market expertise to know not only how to train the model, but how to make it so that it actually does what the end user knows they want to do, right? You have to have some knowledge around that work. So, example I'll give you is we had a company called Case Text, which was in the legal AI space. Initially, they were not uh an AI company, they were a legal research workflow company, uh, but they built an early AI team, and most of that AI team were actually lawyers um who really understood how lawyers worked, uh understood what you wanted the output to be, understood how to make it trustworthy, that type of stuff. So I think that's where they had the advantage. The problem for agentic AI startups is they're trying to compete by having the better trained model to a great extent. And there is th there are ones that will win in that space. They'll probably be the ones that raise the most money so that they can get the most customers quickly so they can get the most data to train the models. So it still all comes back to the data. And you know, I spend a lot of time looking at a lot of these new tools that are coming out there, and especially on the go-to-market side, you know, if you want to look for, you know, uh top of funnel outreach tools with enrich and enrich your data, there's dozens, if not hundreds. If you want to look for SDR replacement tools, there's dozens, if not hundreds. Most of those are gonna fail. Um, and I think that that's gonna be the challenge for VCs in this market is I think a lot of these, the hype that we've seen and going to these markets, there's only gonna be a couple winners. I I don't think the you know, tech has shown that yes, you can build highly specialized offerings and that will evolve over time, but in terms of like big category winners, there'll be one or two in each each major category.
SPEAKER_00So uh do you wanna do you mind showing your cards as far as the uh the SDR agentic platforms or tools that that you think might uh might have a stake at being one of those winners?
SPEAKER_01Oh, that's a good question. So so far, what I'll tell you is we've written our own. So we've been playing around with it. So, you know, I know what you do, Chris, you help people understand how to adopt AI. And, you know, for you have to think about where we are in this adoption cycle. So, first of all, when the way I always think about this might be a bad analogy. So let's go back in history a little bit and think about the Yahoo to Google transformation. Um, right? So if you think about Yahoo back in the day, and I think you and I both could are old enough to remember using Yahoo, it really wasn't search, it was sort of an organized file structure. Like you went to their homepage, you clicked on sports, you went to MLB, you went to the Astros, and back then you saw the Astros sucked. Um, right. Uh, but you it was really sort of organized for you, and you did the workflow yourself. And then all of a sudden you got Google and Google just put it into a box. And I remember people's initial reaction, like, what do I do with this? And the workflow was behind the scenes. They were doing it for you, right? But it was a mindset change in how to use it. And a lot of what's going on now is with AI is very similar. People are trying to take the models they know to play uh AI against it. So if you look at, you know, take examples, not I can take my family, I can take people I work with. To a great extent, people think think of things like Claude or Chat GPT as a replacement for a search engine, right? You know, and and that's okay. That's how they start, right? Where the first things I do with people is say, hey, why don't we flip it back around and have it ask you questions instead? Let's give it a roll and try to start the interaction, right? But that's how they use it. The same thing's going on within the world of uh uh of startups and how they're used. So the even the question you asked, like SDR replacements. I think over time, like that concept's gonna go away. Like this idea, we've set up this thing that's really based upon capability of individual peoples, right? Uh you want some on top of funnel, specialized SDRs to go to AE's, AEs, you know, then go to customer success. I think all that's gonna kind of go away. And that's why we're trying to adopt it. So that gets back to the original point, which is why do you build your own stuff right now to a certain extent? Well, part of it is so people understand how the stuff gets used. So, what do you need an SDR to do? Well, you got to give it training data, right? Give it all you can about your products. Second thing you need to do, you need to give it all the stuff you can around uh customer conversations. Let it train on customer conversations. Third thing you might want to do is, well, how do I speed up customer conversations so it can learn? Well, maybe you want to build another agent that actually acts as a customer, right? And so part of it is just starting to learn how to use these tools themselves and and work. And so I think that's a uh that to me is really important. But a lot of these concepts are gonna go go away over time, even like in the back end, the concept of product management and engineering, I think are gonna merge to a great extent over time.
SPEAKER_00You know, this is I I agree with you 100%. Most people, we've built a reflex over the past 20 something years of using Google, right? Like, so we've been that's that's how we think that we need to get information. And when people translate that into generative AI, they're they're missing the boat. I had not considered though that entire roles, like you just mentioned with the SDR thing, we're thinking about how do we take Google into Chat GBT is the equivalent. How do we take an SDR into AI? But it's not the same. So that's a fantastic point here. Now, this this concept about really understanding the roles and having that be, I guess, the driver of how you use AI as compared to like not just the roles, but the activities that occur within those roles to drive the translation into a generative AI tool or an agent or something like that.
SPEAKER_02Yeah.
SPEAKER_00This idea about AI's biggest value being invisible, it's not necessarily in the outputs, but more in the operations. Yeah. Can you talk a little bit about this? Uh why most people get started, they use it to I talk to execs all the time, and we have them scale on us on one to five, where are you with AI? And it used to be a lot of ones and twos, but now we're seeing more threes and fours. And when I ask them, okay, wow, that's great, what do you use it for? They say two things writing emails, summarizing documents. And I think that's kind of like just the tip of the spear for a lot of people. But what you're suggesting is is something much deeper, this kind of orchestrating decision making and things like that. Can you share your perspective on that?
SPEAKER_01Yeah, I'll give you an example of one of the things I've done internally uh at Build Group, because I I was working on it right before this call. And I'm trying to do this also to help my team uh learn how to adapt these tools. Um, you know, because what you forget, by the way, is even if you're in your 20s or 30s, you've got the existing paradigm in your head. Um so all of us are trying to kind of learn how to use this tools. But the example I'll give you is um we we, like anybody else, build target lists and we've used traditional tools to build those target lists. But what I try to do is add on to that in terms of how we do our work and and marriage what I do. So the first thing I did was have an agent that actually takes a list of companies and scores them for me. And I give it a set of criteria, uh, things I want it to go and look for. And is it perfect? No, not by any means. It's probably 80% there, 75% there. I've never really kind of gone and back tested it, but uh in general, it gives me data that I can kind of check against and say, okay, if I want to go and look, I can get an idea where it is, but I have it tier it, tier one, tier two, tier three. Tier ones, uh, then get a deeper level of research with some more information. Um that by human or by an agent? By the machine. Um and one of the things I do is we use a CRM called Affinity. Um, I use Affinity to jump dump information into Google Sheets, but then I uh I've written my own Google Sheet scripts to query Chat GPT uh and fill out columns with additional data. And that could be anything. It could be look for information on number of offices, look for information on uh the last time this company hired up an employee that looks like S. So it starts to fill in that information. And this is when you start to get beyond what existing tools can do. And also, by the way, I didn't write a lot of that code. I just sat there and, you know, asked, I've used various things. I've used uh, you know, I've used Claude as a big one paper of mine, but I just tell it what I'm looking to do uh and play around with it uh and get it going. Then uh once I get uh a list of those companies put together, I have it generate a first draft of an email for me that loads into my Gmail outbox um so I can review it and look at it. And so I show people that as an example, that goes on completely behind the scenes. All I do is kick it off by getting her a list of companies. Once I get her a list of companies, uh, it goes and does it. Now I'm gonna change things as we go, but in the but in the end, uh I'm using it as a teaching moment for everybody I work with too, which is like, okay, I didn't know how to do this stuff before. Um, and you know, uh I have traditional investment analysts on my team that are really smart. They've never called an API. Yeah. Yeah, how do you call an API? I go, well, have you actually sat down and asked one of these things? How do you call an API? It'll walk you through it. Uh it will give you the code, right? And tell you how to do this step. So that's an example of it, though, of one of the ways that I'm I'm doing it today, internally at least.
SPEAKER_00That's another example of that behavior change. I don't know. Well, did you ask the models? Nope. Ask the models, right? One of the things we do in a lot of our we do a lot of training remotely, and we'll get comments in the Zoom and stuff like that. And my my support team has been trained to take just copy that query from a you from an attendee, go to perplexity, hit enter, and then provide them with a link back to, and it's not because we're being snarky or trying to like show them, but we're trying to like remind them you don't there's no questions that you that you that need to go unanswered anymore if you can change that reflex to oh let me ask a mob. So perfect example. Um, you know, a lot of the people that listen to this podcast, we've got everything from consultants to CEOs, but we we really target the conversations towards operational executives.
SPEAKER_02Yeah.
SPEAKER_00What are some of the things that you think they should be doing right now? We've already talked about how people are reluctantly shifting into a new way of thinking about this stuff, but what do you think they should be doing about AI right now?
SPEAKER_01So I think the easy one is to just think about how to do it on the back end, right? And you know, when I talk to a lot of private equity firms, for example, the way they look at AI, again, on this whole theme of taking existing models and applying new technology against it, they see it as a cost reduction opportunity. And that's what they're doing. And I do think that's that's true. You need to go and look at that. I think customer support, I think automation of that is good. In a lot of cases, you're probably gonna actually deliver better results to the end customer than if it's human-led. I think that's important. But I think what's really interesting is on the creativity side and really on the go-to-market side and the ability to drive more growth with at low marginal cost. And, you know, we could go down a wabbit hole in this because I think VC is in real trouble over time, uh, because I do think the ability to grow at low marginal cost at lower levels is is we're getting there uh really quickly. But I'll give you a couple examples. So you're asking for some tools on the front end. You know, a lot of our companies use clay or they use ample market or they'll use Lem List to get data. Um, but what we try to do is go a level beyond that, which is a lot of what I just laid out for you is taking the data on leads and getting better enrichment of the data so that you can get hyper-targeted. So when you think about, you know, even with the the list that you get today, even with these a better AI tools, to a great extent, they're still very spammy. They don't necessarily do a great job of narrowing it down to specifically what you're looking for. Um, and that's where right coming in and saying, okay, I've got an initial cut of it from these systems. Now I want to go in and really target, I guess, criteria I care about uh to get to really focused uh outreach, and then use that to craft a custom outreach message that looks different than saying, Hey, Chris, I see you're doing great things over there. I'd love to chat with you, right? We get those emails all day long. Highly personalized. We just went through a project with this with one of our companies, and um it we're early days in this, we're three weeks into it, and they went from uh a response rate of two to three percent to over 50%. Now, that was a combination, and by the way, at much lower volume, right? You got down to a lower number of targets, so it's a lower number of targets, but I think the messaging to them is really good, right? Yes. I'll give you another example that uh I've been really intrigued. So I learned that one of our mentioned product management earlier, and uh a lot of our companies are really kind of standardizing around cursor or claud code. A lot of people use cloud code as well because you can interface it directly in the terminal window. Um, but it was interesting because I learned about one of our product managers, a junior one, was all of a sudden submitting pull requests. And so I did a call with her to look at what she was doing. And sure enough, uh, she was spending a lot of time on issues that were coming up from customers, small issues, right? Think about, you know, change a button collar or, you know, uh this report's missing uh a key. Piece of data that should be there. And what she would do is get in a position where she would actually, you know, obviously she's in a position where she can use the systems to query our code base, figure out what's going on, get recommendations from it on how to potentially fix it, work through it, um, and then uh try to get it fixed. And initially, they weren't even letting her really try to run tests on her own, unit tests on her own. The engineering team's now letting her do it because they're like, well, we don't want to deal with these issues anyway. Um, we want to work on bigger stuff. And I'm like, well, that's really fascinating. You know, this uh this product manager's not an engineer, never read a line of code, knows the product really well. So this idea that you can marry someone that understands product requirements, what they're trying to do with customers, with the ability to real-time build code for them. I mean, it's powerful. Um, that's what I mean by the innovation side of things. I'll be really curious to see where that goes. Uh, that's a trend that I'm really interested in seeing on on how the product and product and engineering are going to merge over time.
SPEAKER_00So, this is a uh a good segue into my next question, which was about the I'm always curious about the future of org charts, right? And this seems like a blending or almost the creation of a brand new job title or role. So, and then we've talked about agents and how they're doing better quality work and faster than what a human would do uh in some cases, like the enrichment and uh list selection and and that sort of thing. How do you kind of see this new version of an org chart and an AI-driven company?
SPEAKER_01That is such a good question. Um, you know, because we've talked about that concept a lot, uh, and what ends up happening. But I think we're gonna go through a period of time where roles are gonna merge. And uh a lot of it, I don't know how to think about it, uh, uh other than it's gonna merge along the lines of, you know, managing the customer, an automated customer journey or a partially automated customer journey, you know, is kind of the way I think about it. Um, and I on the back end, I do think you're gonna have uh an elimination ultimately over time of this concept of product and engineering. I think those two things are gonna come together, but I also think support's gonna become less of a back office thing too. This it will become more of a kind of a customer journey thing. So I, you know, it's a really good question. Um, you know, one of the things that I try to help people do is think about agents as people. Um and you know, write agents with row roles and give them very space. You could certainly build an agent that does everything, right? But I also think there's value in say, let's build a hyper-specialized agent and get used to managing chains of ages to do different things. Uh, and because by the way, every platform has different strengths too, right? Every model has different strengths. So you might want to be doing, you know, different things within that. So, you know, I, you know, if you look at the ages I have, I have a red team and a blue team agent that I use for investment committee work that I essentially like put together the negative case, put together the positive case, right? And I I actually get them to interact with each other so I can see if I can get information out of it. Um, so to me, this concept of kind of again replicating human roles in these systems is gonna be really important. But I think also um the human activity is gonna be orchestrating those to a great extent. But it it's hard to know what that's gonna look like.
SPEAKER_00Let me ask you a question about this this customer journey concept, or or really any place where a human is gonna be touching an AI output that is intended to I'm not gonna necessarily say trick them, but where they don't, they don't, they may not, the user may not know that they're interacting with AI.
SPEAKER_02Yeah.
SPEAKER_00Do you think how how far out? First off, do you think we're there now? And if we're not, do you think like what does that timeline look like based on your experience?
SPEAKER_01Well, I think people are not first of all, I don't know if a tech is quite there yet uh for a lot of use cases. Uh and I think that, you know, the trust factor is not there yet. You know, I'm not trying to, you know, you know, we we all can remember, you know, distrust of the internet or distrust of whatever, right? I think a lot of people are having a hard time trusting agents of what they can do. So it might almost be user acceptance challenges more than capabilities that are slowing it down. But, you know, like you know, we were just talking about SDR. So just think about this concept of I come to a website, right? We already have, we already have PLG motions where people come in and sign up and never talk to a human, right? So that will continue uh and we won't change all that. But there are situations when folks need information or want to have an exchange of information. Uh, and that goes from light exchange of information, like, hey, I just need someone to ask some questions of a demo, all the way up to heavy enterprise stuff. It'll take longer for enterprise to get there, but the stuff that exists just to be up above PLG, there's no reason you shouldn't be able to automate that. And the way I think about it today would be you know, Chris comes to a website for one of our companies, um, immediately is greeted by, you know, a familiar, we already talked to chatbots now, but you're created by an intelligent agent. That intelligent agent is acting as an SDR, collecting information, sharing information with you. Um, you may say, hey, I'm interested in a demo, uh, in which case it could spin up a demo immediately for you. And how already has information on what you're interested in, can walk you through that. Or you may say, hey, I'm not really interested right now, I'm gonna leave, but it's always there to come back to it. It says, Yep, come back whenever you're ready, or whatever time of day you want, and I'm gonna be here ready to go with the same context you left with, right? And it's gonna stick with you all the way through. At the same time, on the back end, humans are gonna get in this data and knowing when to jump in and intervene. So the concept I've been thinking about on this would be even if you had schools of SDRs, digital SDRs, um, I still think you're gonna have a human on the back end that's there to kind of watch when you need to escalate or get involved. Uh, and I I don't know that will ever go away. I think we'll always have some level of that going on behind the scenes. Um, but that may become more what humans do. Humans may be sort of like an escalation point uh for things. And maybe that's what these junior jobs become. Maybe a junior SDR uh is now gonna be someone that actually manages digital SDRs. Um interesting. The process learns how to be an AE or learns how to manage a fleet of digital AEs. I don't know. It'll be really interesting to it's hard to think this stuff is it's like anything else. You just kind of get out and do it. Um start being around with it and see what works and what doesn't. The hard part for me right now, with even with startups with young companies, is when you have people and processes in place, it's really hard to get people to change. It's just it. And uh that's one of those things that you know you have to kind of it's hard to make. I think everyone's looking for like these big revolutionary moves. Revolutionary moves are gonna exist with these brand new startups that don't have any legacy, right? Evolution is where it's gonna happen in companies, even in startups that are, you know, five years old. And that evolution is gonna be more at the individual level than it's gonna be at sort of corporate level, right? It's the same with like open source or developer tools. It's not like all of a sudden corporations mandated these things. You had individual developers who just started adopting them and using them. And I think the same thing is going on with AI and the enterprise as well. The example I gave about the product manager, just what she decided to stop doing, right? That to me is that's what I'm trying to pay attention to when I look at things, like, okay, how are things going? What can we learn from it? How do we build upon this? Is it valuable, not valuable? Are we still learning and just need to kind of let's see where it goes? Those are things I'm trying to track right now.
SPEAKER_00So, you know, that's a question that I I get asked a lot, but don't have an answer for, and it's generally boils down to what's going to happen to the humans. So this uh idea that you just shared about a junior level role essentially being an output evaluator from agents before or an escalation environment is interesting. Just generally speaking, and I know it may not be popular to say that AI is going to be replacing jobs, but kind of where where do you see the human fitting in? And it doesn't even have to be five years out in the future, but with what you're already doing, how are you seeing the human create ongoing viability in the enterprise?
SPEAKER_01Yeah, that's a good question. I don't have a strong opinion on this one way or the other. I'm a tech optimist, and so uh my belief is humans will rise up to another level in terms of the contribution they make. Yeah. Um, you know, you could go back and think about, you know, early days of software development. I mean, even before it was software development, it was it was really writing software on a chip, right? Uh, and then, you know, you got abstracted away from that and you continue to get abstracted away, right? Developers have not only uh it continued to exist, they've thrived. Um, so I don't know how to think about what's gonna go on in this space. Um, I think that the speed of it's the challenge, right? I think if you had 20 years to see things change, I think we'd figure it out and it wouldn't, but it's happening really quickly. And also people are trying to make dramatic moves uh because they're feeling the need to like, okay, I want to go cut 30% of my staff in this area or 60% of my staff in this area, um, and are trying to make those moves. But I don't know. I think the way I would tell everybody and the way I would tell my, you know, told my 16-year-old uh son, like, you gotta learn how to use this stuff. Yeah. And not because I think that's what you're gonna do for your career. Like, I don't expect you're gonna be uh, you know, involved in AI engineering, but you're gonna have to figure out how to build whatever it is you do around it. Everything's gonna be a part of this. But in the end, AI is there to replicate what humans do and say and uh how they work together. Um, it can't replace everything we do. I mean, otherwise, what's happening with the humans, right? Machines are gonna, yeah, because in the end, it it always involves human activity of some sort, human going to a store or human, you know, you know, doing what consumption, right? So it's hard to uh envision what that's gonna look like. Um I also believe there's gonna have to be some societal reorganization at some point, most likely, um, which is a harder discussion because you start to get into politics. Uh, but I don't know. You know, it's one of those things that I I just think it's we gotta wait and see how it evolves. Um, I'm neutral on whether I think it's gonna be, you know, how I think it's gonna impact jobs in the long term. In the short term, it could, it's gonna have an impact. I don't know how we're gonna adapt to it and what that's gonna look like over time, but um, I certainly don't think anyone it's in anyone's best interest to allow people to not have hope or jobs. So I think we're gonna have to figure out how to solve that pretty quickly.
SPEAKER_00Okay. The next thing I want in your opinion is I know you're advising and investing in a lot of companies. AI is certainly one of the top considerations, but you also mentioned the speed. The way that we tell executives that AI is not an event that occurs, it's a a series of steps that compound with you know uh efficiency and you know more bandwidth and all that kind of stuff. How are you advising companies, peers, friends when it comes to, hey, we're we're feeling the pressure? We know that like we've seen some anecdotal uh case studies popping up in industry magazines about how this company's doing something that we'd like to do, but we like we're just getting started. How do you advise this kind of balance between speed, but also necessity?
SPEAKER_01Yeah, I think this is where you gotta this is where you probably do need to think about the low-hanging kind of fruit on the not just the cost side to the competition side. So the example I will give you is if you are a brand new startup right now, application startup, not an AI startup, but you're developing a workflow tool, um, you can probably get 90% of your code written by uh by machines, right? If you have a legacy code base, yeah, you might get 40%, right? That becomes a real challenge really quickly, right? Because if you're worried about getting eaten from the bottom, so if you look at the companies we invest in, if they're 5 million in revenue right now, they've got an established code base, you're at real risky and eaten from the bottom because people can just go faster on what they can do, right? It's also cheaper, right? So I tell people you got to make that move pretty quickly. And a lot of times that involves is re-architecting your environment and doing that so that AI can start doing much more of the writing of the code, right? But it's not even just that. You've already got people and processes in place that you're gonna have to have an honest discussion about how do we move them out and change it because you have to make those changes, right? So I think that is an easy one uh example. So that that's that to me is one where you start. Um, but I also want people to get optimistic, right? So yes, you had to go and compete. You had to deal with the fact that everyone's gonna use this to leverage on the back end and you're gonna have to figure it out and pick pick those areas that are gonna be really relevant to you in your competitive space for us, for our software companies, that is developing your software fast, right? And being able to continue to roll that out. The front end is where the optimism comes in, which is where I think you can use this as a tool to bend the curve of growth and grow faster. And I think you can take companies that historically have not grown very well and help them grow faster and low marginal cost, right? And so I always kind of do it as like both things, which is like, okay, you got to do this stuff because you need to compete, but I also think you can do some really amazing stuff around the front end of your business. Um, you know, I was even meet with a friend of mine who has a home uh repair uh business that he started, and he was just asking me about a lot of these tools about for generating leads and how do I use it to better manage my business. So I think people have to look at that and adopt it, but also remember it's very rarely the company that adopts it, it's the individual. Yeah. And when you see winners that are doing interesting things, give them the success to expand it. The same way when you let you did that with developers with open source and with their tools, do the same thing on the business side. If you see people doing interesting stuff, give them the flexibility to go ahead and do that. Don't give them a bunch of mandates. Don't go tell them they have to use AI. Find the people that are natively wanting to go and use it and look at how they're doing it, and then encourage that through the uh through the whole organization. There every company has that right now. That's what I've started to discover. Even startups which are growing fast, they don't have everyone that's all in on this, but there's always some in there that are doing interesting stuff, and those are the ones that I'm really focusing on.
SPEAKER_00That's another topic that whenever we're working with a company, we like to introduce this idea of you need to find those evangelists internally, organize them, and build something around them being able to assist with the change management, not just, hey, the boss sent a memo, but this guy sits next to me, he's using it. Do you uh have you seen any effective frameworks? Do you have any uh suggestions for anybody who's we got a couple people in our our company already who I know they're using AI, whether it's yes, you know, advertised or not? Yeah. How do we leverage those existing enthusiasts to kind of help manage change management, but also or assistant change management, but also disseminate that education throughout the perhaps reticent or resistant users?
SPEAKER_01Yeah, I uh I'll take go back to my rack space days for some lessons on this and how we culturally uh dealt with not only adapting the technology changes, but really kind of uh ingraining our culture of people. And we were a services-based company, and one of the things we did early on was make sure that we celebrated in front of the entire company on a regular basis, heroic efforts on behalf of customers. Nice and tell those stories, and we would give the whole story on it. Um, when we were moving over to cloud, right, which by the way was a huge cultural move from us from going to selling someone a minimum of $2,500 a month server at a minimum with a three-year contract to say, okay, there's no contracts. And by the way, they could start with a server that if you're running full all-out scored bucks a month, right? Uh, how do you get salespeople interested in that, right? So a big part of that was regularly finding success stories and elevating that up in terms of the stories we told. So what I've been trying to do within my organization, which is small, and then with our companies, is do the same thing. Y'all bring people together to talk about how they're using AI and ask them to tell the stories about what they're doing within their jobs. Don't have to come and, you know, have an agenda around it, but celebrate that. Um, and if people don't people want to bring examples that they're doing in their personal life, bring that too. But talk about how people are using it to deliver a better outcome and celebrate that and make the time for it, right? If you make the time for it and you show that that's important, people will do those kinds of things and bring them in a room together. You do this, Chris. You uh, you know, bringing people together to share these ideas is really important, right? And, you know, I do a lot of in a uh reading of people that talk about different prompts to use, how to kind of get the best out systems. I actually do best when I see people doing stuff live, right? When I can see real case studies. Uh and so that's what I like to do is get people here, should just say, show me how you do that. Like show me how you got that result. What exactly did you do? Did you look at alternatives for that? Can I copy that from you? Uh, all that kind of stuff.
SPEAKER_00You know, we're actually gonna be that's one of the things we're working on right now, uh, is developing a framework for companies to establish that kind of AI internal council. And I don't mean necessarily like the strategic side, but the tactical side, right? And what you just shared is gonna be something that we're gonna introduce. Like, who's already using it? Show us how you're using it so that other people can go, oh, that's not just for the MBAs or the the C-suite, it's for me at all levels of the company.
SPEAKER_01I think that's a I'll g I'll give you I'll give you an example I just did too. So um I redid our entire website, Lovable. Yes. And why did I do that? Well, because we were making changes, they take a long time, and I want to be in control of it. And I didn't want to sit down and do HTML code or deal with WordPress and like that. I wanted to be able to manage it. And also, like, it's not cheap, right? Paying someone to kind of maintain all that. And I just sat down with Lovable and sat there and basically rebuilt our website and and I think I made it better. And I shared it with the team, and the team's like, How did you do this? And I go, am I honestly, you can read my entire problem testry and see everything I did with it, but yeah, that's how I did it. And they asked me, Why did you do this? And I said, Well, first of all, I love learning new tools, uh, and this one's easy to learn. And I said, But the second is think about now. You want to make a website change today? I can do it. And once you learn how to do it, you can do it, right? And you don't have to wait for someone else. And so again, this is an example. I'm sitting with my team and talking through and showing them, right, how to do this stuff. But again, it's not even necessarily that I care that they use lovable. I just care that they sit down and say, huh, let me think about this looking for other tools we could potentially use to do this kind of stuff, right? Or, you know, in my case, I talked to my team about you should never ever build a slide again. You should tell AI how you that you wanted to build the slides. And if you do anything else, you're wasting time.
SPEAKER_00Yeah, we we have um we use the hell out of gamma. I think I saw my bill yesterday, it was twenty dollars. Oh, it's easy.
SPEAKER_01Well, I I you know it's funny, uh, you know, I got you probably like me, but I'll sometimes not be able to sleep at night or I'll wake up early. And I think I spent maybe three hours redoing our website. Yeah, maybe. Uh I it made it easy because they have a Figma plug in so I can pull over our original files to kind of get started from. But I mean, it it's just it's amazing what you can get done. And uh by the way, that stuff, I think everyone in tech does certificate's a learner, uh, or I like to think they are. I just the it's it is uh a smorgash board of tools to try. Um, and they're all easy to try. Um, and that's the whole thing. I just said people don't overthink it, just grab something and start doing it, that you can show the impact in what you do every single day.
SPEAKER_00You know, for the listeners, if you're you know, you don't know where to start, because I get it, the tool overwhelm is something that we hear about a lot. This advice that Jim just shared, just get a tool and play with it. Like take 30 minutes, take three hours, whatever it is. And I think that you'll be blown away with how much you, as maybe a new user, are able to get done with these things because it's it's natural language interaction, right? Like your prompts and lovable, were they engineered prompts?
SPEAKER_01No, uh, I I I intentionally don't try to learn uh prompt, you know, I didn't go research like write prompts for lovable. I just started writing my own. I started learning as I went along. Uh, but what was really uh, and I I will at some point go and actually look at third party psych. I was funny when I read my it's like I'm talking to humans. I use words like please and thank you, which is real funny because it's just natural, you know. Uh it and what that tells me is it it feels a lot like I'm talking to another human. Um and I find it really interesting. And when I started trying to write more stilted, I didn't like it. I actually wanted to have a conversation with it that was much more like I was talking to a human. Um but that's the thing is like pick something, think about any task you have to do. Um, you know, look, if you're trying to plan a tr uh a trip, like, well, why don't you just say there there are great tools out there for for you if you don't want to use it for research, if you just want to build an itinerary, like there's all kinds of things you can use to do it. I'm trying to do one right now that's hopefully a better version of bands in town. Uh as you can see, I like a lot of music in Austin. So, you know, Chris, by the way, if you want to go see live music, I'm your guy. Um, you know, a lot of people ask me all the time, like, you know, who should I go see? And, you know, first of all, I asked, well, if you integrated Spotify with, you know, Bands in Town so you can track all that stuff. Like, well, no. Uh, and by the way, I'm not sure I've got a great Spotify list, but here's what I like. I'm gonna have to find a way to better curate this stuff with my friends because they get asked these questions all the time. But that's my point. Find something that's and it's not like it was 20 years ago we had to go write code. You just have to find the right tool to try it out on try it out with us.
SPEAKER_00Yeah. Jim, I'm looking forward to us catching up in Austin. Um, this has been a fantastic conversation. And I I got actually I got a million more questions. But for people who are vibing with the way that you're looking at things, which I am, where do we find out more? Like, are you putting out your thoughts anywhere?
SPEAKER_01Yeah, I have uh I I I post quite a bit on LinkedIn. Um and uh it anyone's welcome to email me anytime, gym at Voggroup.com. Um, I run our IT system so you can I get the catch alls too. So if you get it wrong, it'll come to me. But yeah, you know, I reach out. Um, I love having this topic. Um, you know, what I told I'll tell your listeners what I I tell the students at UT when I talk to them, which is I would love to hear about things you're using AI for. Yeah. If you can introduce me to a new tool, if you can introduce me to a new prompting uh uh scenario and how you use a specific tool, I would love to have that conversation. And in return, I'll share something with you uh that I'm doing. And like I said, I do it every day. I try something new every day on this. Love it.
SPEAKER_00Well, Jim, we're gonna include all of those links in the show notes. And for everybody that was listening to this, um, again, if you have questions, it's Jim at buildgroup.com. Uh, based on this conversation, I think that if you enjoy our podcast, you would enjoy having uh Jim get back to you with some of your questions. Or if you have anything to share with him, that's a fantastic idea. Again, Jim, when we get in person, I think we'll have a much longer conversation. Um, I really appreciate your time on this. And like, I've got more questions than I started with now. This is great. But again, looking forward to catching up with you in person. Everybody, thank you so much uh for listening to the episode. We'll be back next week with a fresh one. Thanks, Chris. Thanks, buddy. Thanks for tuning in to Using AI at work. Don't forget to subscribe for more conversations about how to use AI at work. And a special thank you to our sponsor, Chief AI Officer, for empowering businesses with AI education and training. Visit their website for a free AI readiness assessment and AI strategy guide to help you get started using AI at work. That's www.chiefaiofficer.com. Well, thanks to our producer, Evan Destalnier, for making this episode possible. Follow us on Twitter at the handle using AI at work and visit www.usingai at work.com for free resources to help you harness AI in your role.