Latest / The Flipped Story / Flipping AI to Reclaim Human Leverage in Business with Drew Donaldson
Transcript
- David: I'm David Gadarian and welcome to the Flipped Story. Today, we're looking at AI not as a tool to strip away human agency, but as an intentional framework to scale impact. Joining us for this flipped AI adventure is Drew Donaldson, CEO of Automata Intelligentsia. Drew, welcome to the show. Drew Donaldson, Automata Intel: Nice to be here, David. I appreciate you having me. David: Absolutely. Absolutely. So I think as a just an easy way to level set, why don't you tell us about, you know, maybe two to five minute short version of what is automata intelligentsia and then we'll kinda work through it all. Drew Donaldson, Automata Intel: Sure. So we focus on custom engineering for more of like proprietary platforms that businesses will use, whether that's internal or interacting with their clients. we also do a good amount of GTM engineering, so connecting their existing products together so they work better together, as well as doing a lot of work in a Gentec AI and building platforms that leverage teams of agents to do work. that would otherwise take humans many, many hours to do in a relatively short amount of time, or error-prone tasks. So, you know, things that, you know, if you if you have a particular process that has to go through a a cycle of reviews and those reviews show more errors. And then in the effort to correct those errors, more errors pop up. we build platforms that are a bit more accurate than kind of human-led processes for a lot of that. data driven work. so that's that's a bit about what we do. I was involved in I started my kind of journey in AI in twenty sixteen, which is a bit earlier than I think most people that have been playing with AI. I started back when machine learning was the rage and that was the the David: Mm-hmm. Drew Donaldson, Automata Intel: cutting edge, the cutting edge of of what you could do with computers. and I've really stayed up up with it from then from then on. and so I I used to run a marketing agency and we implemented AI just about as soon as it was ready for prime time. And and now in this business, I'm a very heavy AI user. I I probably talk to AI more than I talk to humans most days. and David: Yeah. Drew Donaldson, Automata Intel: you know, I I know some people would be like, You're an addict. No, I'm I am perfectly happy to put the AI down when I go downstairs for dinner. I do not need to be in twenty four seven contact. Because i I treat it very much like a coworker at this point. David: Yeah. All right, let me ask you that question. Let's just stay right there. So talk versus type. So you're you're you're you're in AI all the time. You're are you are you physically talking or are you doing more type or what what's your kind of preferred interface? Drew Donaldson, Automata Intel: I tend to type. I tend to type because I am more verbose when I speak and I end up confusing the AI because AI needs clear instructions and I'll bounce around to a lot of different things. And so I will use it for brainstorming. Like I love using voice in the car when I'm thinking of an idea and I want to get the idea out and I wanna get immediate feedback on it and I wanna I want research on the idea and I and I I just want that iterative approach while I'm, you know, driving to the doctor's office or something. so I will David: Right. Drew Donaldson, Automata Intel: use voice there. and then I use voice when I'm with my family and we all have a question that like, you know, my wife and I will be in an argument about, you know, whether Henry Cavill was in a movie. And I'll be like, you know, I can't say it because my phone will pick it up and start talking about Henry Cavill, but you know, I'll say the magic words and ask you the question so that everybody can kind of hear. So that's the only times I really use voice. Otherwise I am very type centric. It's probably not the best. I know I could probably be more efficient, but I just it's it's easier for me to get my thoughts clean cleanly to the AI if I type. David: I totally get that. I we we've had a couple folks on the show that one one guy was was pure voice, but I had another guest who was AI centered and we talked a a fair amount of his his voice journey and how he's kinda been doing some interesting discovery and and workflow kind of development, particularly around what you just said. I think of for a lot of folks it's about ideating, not necessarily doing the work instruction. The work instruction seems to we're not quite there yet. Anyway, neat. Drew Donaldson, Automata Intel: Yeah. Simple tasks. I I I I could see that being the case. It's just so little of the time I'm talking with AI, it's simple tasks. Because like David: Right. Drew Donaldson, Automata Intel: so often I I think people do a lot of over engineering on like, how do I you know, I just wanna delete all these emails from this sender from my inbox. I'm gonna have AI do it. And it's like they spend like six hours doing something they could have done with a search in their inbox and like select all delete. And David: Right. Drew Donaldson, Automata Intel: And it's like, so it's it's like the simple tasks, half the time I'm like, well, is the AI going to be able to do this easily, or am I going to talk to this for 30 minutes when I could get it done in ten minutes? And so most of my conversations tend to be pretty long-running, complex, like multivariate things that I kinda need to type it out to keep track of where I even am. and then the other challenge I've had with voice is Voice doesn't always save what you said. It'll only save the responses. and that sucks because you're like David: Right. Drew Donaldson, Automata Intel: you read the response, you're like, I don't know what I said to get this response. Like, what was the idea getting? Yeah. It's David: Right. How did we get here? I know where we are, but why? Drew Donaldson, Automata Intel: it's kinda like waking up in a car like at the other side of the country and being like, I have no idea how I got here. Like, does you have a map? Is there like, you David: Right. Drew Donaldson, Automata Intel: know, a GPS log? It's it's frustrating. So I I tend not to I I don't use voice probably as much as I should. But I do like it. I do I David: Yeah. Drew Donaldson, Automata Intel: it's it's a lot better than it used to be. It used to walk all over you and if you talk for too long it would cut you off and it's better now, but yeah it's tough. David: Yeah, it's getting there. You can see also a lot of these providers are starting to feature and set our voice a lot more. So clearly we're the guinea pigs and it will continue to to to scale up. But yeah, I tend anyway, it's neat. Drew Donaldson, Automata Intel: So I I use Hermes Agent as my main AI agent on my machine. f as like my assistant MVP developer kind of magical toolbox of tricks, right? David: Right. Drew Donaldson, Automata Intel: and so when my computer's off, I don't have access to that because of Hermes' local to my machine. And David: Yeah. Yep. Drew Donaldson, Automata Intel: so my I was dealing with a hard drive replacement issue. and I try so I I David: God, your brain. Drew Donaldson, Automata Intel: switched on Grok and I turned on their new video. I don't know if it's new, I just never used it before, but their video mode, which lets Grok see through the camera. And I had had no expectations of what it was going to be able to do. And all I wanted to do is ask it if this hard drive chip that I was trying to put in the computer, if the chip was too big or if I was just putting it in wrong. Because I had never it was a it I've only had this computer like a year. I never had to replace the hard drive in it. And I'm not a big like hardware guy. I'm a software guy. So like I don't know what I'm doing. So I turned on the camera and I asked it and it was like yeah that's too big. peel off the thermal shielding or the heat shielding or whatever and it'll fit. And I was like can you can you actually see that? And he's like yeah it's off by like a millimeter but like the heat shielding on that's about a millimeter thick. If you peel it off It was like being on the phone with my uncle asking him about my like radiator. Like it it was just so it was it was so concise. And so I I did I stuck there with a pocket knife and peeled the heat shielding off. And I was like, I guarantee this is gonna be just like dead. Plugged it in and it fit beautifully, and then it worked beautifully, and I was just like, I can't believe how good this is. Like it's it's funny, it's like you think of someone that like works with AI all day long would never be like impressed. I'm impressed every single day. Like there's never David: Yeah. Drew Donaldson, Automata Intel: a day doing this that I doesn't go like, Well, by golly, you're good at this. Like David: Right. And I I think that's part of that's an interesting insight of just the some of the the magical awe and the wonderment of how are these things happening. So let me let me let me flip that for you. So you do a lot of this. This is what you do all day long. When new new clients come to you, business owners, when when they're coming to you generally And generally you don't have to be specific by any mean. What what what what's their headspace like? Is it like very intentional or is it very much a little bit like reactionary to like we know we have an issue, we is it about cost, is it about optimization? Where where are they kind of coming at Drew from like right from the beginning? Are they confident? Are they shitting in their pants? I don't know. What's what's happening? Drew Donaldson, Automata Intel: Not I think there's like there there's a sense of like real optimism because they realize that for the first time these problems that they've been dealing with or this like dream that they've like held on to and like held out hope for is like could actually happen now. And so there's there people tend to be very excited. I don't get a lot of people who are coming in dragging their heels being like you know, you you gotta fix this. Like I'll never survive another week if I don't have like it's it's a lot of it David: Yeah. It's too new. Drew Donaldson, Automata Intel: is hey, we lost our receptionist last year because she retired and we've been having all these other people and we we really just want like someone to like kind of manage the inflows and outflows. And I don't know how to do it, but I think AI can do can AI do this? And then the question is, well, you know, it's it's like the famous it's just heard someone say it's the consultant's favorite phrase, it depends, right? It's like, well, it depends. Like, well, well, what are you actually trying to do? Are you trying to have it answer the phone? Because yes, AI can do that. Can AI have broad, long ranging conversations on the phone? Kinda not really. Like, you know, can it confirm appointments? It absolutely can confirm appointments. It'll do a great job confirming appointments. So, like Yeah. David: statements can get confirmed. Drew Donaldson, Automata Intel: So, like that that's the thing is like they're They they're excited and they're optimistic, but they're also kind of like that it's it almost it almost comes off as like they don't want to ask for too much. Like they don't want to be like a space cadet out there and like pitching me some idea that I'm like, that's not gonna work in a million years. So they tend to be very conservative and like they're just like, Well, if we could just fix like this thing and then I'll do more digging and be like, Well, what else? Like, well, tell me why, like, that's so important. You just said like, This is a problem, but you have these other problems. Like, why wouldn't you just solve all these problems? And they're like, Well, I just you know, I I didn't think AI could could do that. Or I didn't think I could actually have a tool that, you know, could could solve that problem. It's like, yeah, like you gave me the the easy pro like the this is a nobody problem. This I could fix this in 15 minutes. Like this is the big David: Right. Drew Donaldson, Automata Intel: problem. Why don't we fix the big problem and then this fall throw in for free? Like, you know, it's David: Right. Drew Donaldson, Automata Intel: so they it's this very Like I almost r I it it sometimes feels like I'm a bit of a magician. Like I I remember going to see magicians as a kid and always being like amazed at how they did stuff and like being so like nervous and excited and like wanting to ask them to like do a trick. Like if I'd see one like out in public, like and like but not really want to like impose, but like also I really wanted to see the trick. And so David: Right. Drew Donaldson, Automata Intel: it's that kind of vibe I get from a lot of people. I don't get a lot of people I was worried I would get more people that We're like, we want to cut headcount and we we need to develop tools to like slim down. And I think we just from the get go, we did a really good job of like making it very clear on our site that we don't build stuff that hurts people. Like we we build enablement tools, not disablement tools. So if you're going to use my tool to go and fire fifteen people, we're not going to do business together. You can find someone else to build that tool. I wanna build stuff that makes people's quality of life better. Like if you have a bunch of really talented people wasting time on meaningless tasks, let's get them back to doing the thing you paid them to do and get rid of meaningless David: Right. Drew Donaldson, Automata Intel: tasks, not just let them go for no reason other than, I'm gonna save a buck. You know. It that's that's I stuck David: Yeah. Drew Donaldson, Automata Intel: with that from the beginning and I think that's scared off a lot a lot of those type of people to where they don't they David: Yeah. Drew Donaldson, Automata Intel: don't end up on the calendar anymore. David: That's an interesting intentional choice. And I I think you you may have timed that really well also. in that Drew Donaldson, Automata Intel: Do have a background in marketing, so I do kind of I'm pretty good at reading tea leaves, you know? Like it's David: Right. But I I I think that's let's talk about the journey here. So from an a purely AI perspective, then this will come back to what we were just talking about a half a second ago. Where do you think we're at where where are we in this continuum of kinda like how early are we? Are we in the middle? Are we second inning Drew Donaldson, Automata Intel: So there's going to come a point, because this happens with all technology, where we hit a plateau, and the plateaus are usually driven by kind of like three competing elements. one is resource availability, the second is some kind of technological hurdle, and the third is util utilitarianism. so in terms of where we are on the scale, I think we're approaching that plateau. AI is continuing every day to become more and more of a utility in people's lives. It's more and more capable. Eventually we will hit a point where there's not much else it can do for us because as humans, we are of limited capacity in terms of what we do and what we need. Right. So at a there will be a natural point where AI does not i is not further developed to serve the needs of the average person or the average worker because all of their needs are already met by AI. And then there may still be progress, but it'll be progress in specific parts, sectors that either got ignored while the general public was being served, or are just they have that use case that's far beyond what the general use case is. and like the way to look at that is like the automobile industry. Like cars haven't really fundamentally changed since like the seven It's like styles have changed, safety features have changed, but it still has four wheels and an engine up until the electric car came out. Now, why did the electric car take so long to come out? Because battery technology wasn't where it needed to be. Distribution of electricity for car charging wasn't where it needed to be. There was a technological hurdle. And so that's the next part that's going to feed into this plateau: is that we will eventually get to the point that the amount of the the the current technology we have is not going to be enough. And that will probably happen around the time where we have to start moving stuff to quantum computing. But even if you look at where quantum computing is now, that's nowhere near where it needs to be for us to even consider making a jump from what the graphics cards that we're using now in big data centers to quantum units. Not to mention the fact that quantum unit you know, you want to talk about energy. A quantum is going to suck up way more energy, which is the there goes the resource issue. David: Right. Drew Donaldson, Automata Intel: There is going to come a point where we are limited in the amount of power we can generate as a country and the amount of power we can dedicate to AI, and that is going to be a constraint. Now, that can that's the easiest one to move because people are already working on ways to generate more power, right? People are already thinking about this. How do we how do we create sustainable energy for these giant energy sucking Goliaths, right? David: Right. Yep. Drew Donaldson, Automata Intel: But that's the slow process. That's not something you don't spin up a nuclear power plant in 30 days. That takes years, right? That's why, like Three Mile Island, which is out by us, when they were gonna turn it off, it was a big deal around here. Because David: Yeah. Drew Donaldson, Automata Intel: not not for the the I mean the jobs were one part of it, right? The bigger issue is once you turn it on, it's years to get it back. You you don't just flip a switch, turn the lights on, you're like, back up in business. No, it's you have to Go through like a two, three year process to get it back online after you turn it off. And wouldn't you know it, about a couple years after they turned it off, Microsoft comes knocking and then bought 500 gigawatts. They, they, Microsoft for their data centers, is using the entirety of a single nuclear power plant in Pennsylvania just for a data center. That's not David: There you go. Drew Donaldson, Automata Intel: tenable long term unless we develop new technology around energy. And so David: Right. Drew Donaldson, Automata Intel: you're going to have these three kind of caps and they're gonna hit at different points, and they may be three separate plateaus. Likely we're gonna hit energy and technology around the same time. but the the utilitarian's kind of the one that's like more distant in the future. But you gotta think like today's AI is so much more capable than what the general public realizes that if If no further development was done today, it would probably David: Right. Drew Donaldson, Automata Intel: take two or three years for the rest of society to realize the unbelievable superpower they have sitting in their pocket. People are still using it to check movie times. Like David: yeah. Drew Donaldson, Automata Intel: you could create your own movie. Like it's it's so far and beyond where people think it is in terms of capability that I think the utilitarian angle is is a decade out probably, if not more. David: So yeah, okay. So this idea that that that your your firm is focused on of of kinda there's some some some choices, some guardrails, right, that you guys have put in that says like we want, you know, human centered, human favorable, human positive. I I don't know if those are your words, but loosely I think that's where your head is at, right? When you when you come into an organization, you know, so you share that with the the the buyer. rank and file, what's their what's the reaction? 'Cause I know this would scare the crap out Drew Donaldson, Automata Intel: So David: of me if I saw you coming, even if you were saying that, I'd be like, goodness, this is this is we know how this playbook runs and sounds Drew Donaldson, Automata Intel: So David: like it and I and I I'll ask a follow up question in a minute on this, but I just wanna kinda Drew Donaldson, Automata Intel: Most of the problems we solve are the kind of things that rank and file have been begging to be solved for years. It's the things that eat up their time. It's when they go into performance review and they're like, Hey, you didn't make as many calls as you were scheduled to make. And it's like, yeah, because I got sixteen hours of reporting to do every week because you have four different David: Right. 'Cause your C R is garbage. Drew Donaldson, Automata Intel: systems I have to access and it it bogs me down. So like it's very rare that I get pushback from someone in these organizations. About the AI topic, what I do get occasionally is people that maybe should have retired a couple of years ago, but are still in their job because they love it. But they're stuck in their ways. They're not willing to bend or modify or change workflows for anybody. And everyone's scared of them and doesn't wanna like they don't wanna, David: Right. Drew Donaldson, Automata Intel: they're not gonna push them to do it. So you end up having this like lonely island effect where it's like, okay. You know, Michelle is gonna do it Michelle's way, and then the rest of us are gonna do it the easy way. And like that's a pain in the neck, but what are you gonna do? Like no one's gonna no one's gonna say anything to Michelle because Michelle's 65 and has no filter and will make you cry in the lunchroom. Like, you know, and and no one's gonna fire her because the boss is afraid of her because she used to babysit for him. So it's like it it's the kind of thing that like you have these, you know, glaciers. In some of these businesses that are immovable. And you just have to accept that we're gonna go around the glacier. We're not gonna, we're not gonna try to change the glacier, we're not gonna try to melt the glacier, we're not gonna try to win the glacier over. We're just gonna say, listen, Glacier's happy doing it this way. She's gonna be gone in three years anyway. Let's not worry about it. Let's do the thing that's gonna help 95% of the rest of people. And then if she sees that they're all now having more time to focus on other stuff and she's still stressed. Then she'll either naturally migrate or she's so good at her job that a tool generally isn't gonna fix a problem for her. It is, it is she was right to begin with. It's not gonna help her. She's better off doing it her way. And like some businesses are okay with that. Other businesses, you know, try to force it down people's throats. And I always recommend when I'm talking to leadership, like, this is we this doesn't have to be scary. This is what they asked for. They asked for a tool to solve this problem. They've asked you that for years. Your Slack is full of messages about when are we getting a tool to do XYZ? Right. There's nothing to be scared about. You're just delivering on a promise you made six years ago and never had a chance to deliver one. So then it's like, you know, if that's the scope, then it's not like me going in and being like, well, tell me exactly how you do your job. And here, I'm gonna just set this recorder right here, or like the office David: Right. Drew Donaldson, Automata Intel: space thing. So what do you really do here? It's not that. It's like, okay, so you hate this process, right? Yeah, great. Tell me every bad thing about this process and they will talk for an hour about all the crazy things that they have to do to make this process work. And then I just take that and I build a tool that solves it. So it's it's never I never have an issue with rank and file that's like a fear because when we come in we're very transparent about the scope and we don't we don't in we intentionally don't do a lot of like, it's okay, we're not here to take anyone's because that's like the the you know, people say like David: Right. It's not your call. Nope. Drew Donaldson, Automata Intel: The the worst thing or the two scariest words or the or the scariest phrase is like, we're from the government, we're here to help. It's like I'm the second one is like, I'm an AI consultant and I'm here to help. It's like, no, no, no. It's it's like, no, I come in, we're fixing this problem. Here's what I need from you, so that you don't have to deal with this problem anymore. And then usually people are very open and transparent. They're not, you know, I know a lot of consultants come into it from a very broad discovery perspective. And You know, there's value in that, but it also freaks people out. It just does. You you have to do a lot more hand holding to make sure that, David: Hair. Drew Donaldson, Automata Intel: you know, a riot doesn't break out on the sales floor. David: I love it. The the world Drew Donaldson, Automata Intel: The the world. David: this space is very it's very interesting right now. So let's maybe shift into that. Let's talk about the the so you come from marketing. It's your background. do you think marketers or engineers make better AI experts or AI practitioners? Drew Donaldson, Automata Intel: Well, so I do come from marketing, but I come from the tech side of marketing. And there's there's these two sides of marketing. You have the creative side, which is the side most people relate to marketing, right? Writing ad campaigns, designing graphics, like all of that. I was never strong in any of that beyond copywriting. I was I was a very good copywriter back in the day. I still am a good copywriter today. I just don't do a lot of copywriting. I do more coding than copywriting. But that was the only like creative thing that I was really, really strong at. and then I had a video background so I was pretty good with video. Not the best, but pretty good. most of my work was on the technical side, which is CRMs, ad systems, analytics, tracking, identity resolution. So it's like it's the part of marketing that no one really knows is part of marketing. It's this like the David: Right. Drew Donaldson, Automata Intel: underbelly of marketing. So moving from that into software development was not a it's not the as crazy a transition as you think because I was already doing this stuff on the back end. For a decade or more. I mean, I've been doing building websites for 20 years, right? So, like way before AI, we were building websites by hand. And like, but that's part of marketing. But people think, well, you build websites, you must have done the design the create. No, I outsourced the creative. Like my wife did the creative. I didn't touch the creative because that's not my skill. My skill's the technology. And so yeah. So David: HTML and what have you, yeah. Drew Donaldson, Automata Intel: the the biggest advantage I have over your traditional engineer. is that I have a much I I think a much broader understanding of human behavior and like how a how a person is going to interact with an application and where they're gonna where things are gonna get frustrating for them. Like I I I contribute to some open source projects that I'm passionate about, right? And the number one thing that drives me crazy is when these engineers put in features and then don't explain how to use them. And it's not in the docs. And they just assume, like, well, anybody that's used the app, I'm sure they'll figure it out. It's like, no, you put you with turn on accelerator mode with no explanation of what that does. Right. And like it David: Right. Drew Donaldson, Automata Intel: could just be a dummy button. Like it could just be a joke that they put in, but there's no explanation. There's no payoff to that joke besides them laughing at how many people click on it. And they're like, you know, it's like, you know, it's so it's it's the kind of thing that like I think because I've worked with people and I've studied behavior and I understand clients and the struggles they have with technology, that I'm very attuned to like how do you build an app that's easy to use? How do you build an app that's so easy to use that you don't need training to use it? And so like when we build something, that's the f at the forefront of my mind is like how can I build something that I can put in front of my mom, who is not the most world's most technical person, and she can figure out how to use it in less than 10 minutes. And if I've done that, then it doesn't matter what high level, you know, crazy consultant I'm working for that's doing crazy complex stuff. If it's that easy, it's gonna be just as easy for them, if not easier for them. And that's the basis of developing really good applications is is having something that's so dead simple to use, that's so user intuitive. And I think that's where that's where I see a lot of engineers struggle with developing apps. Especially when they're on their own, when they don't work with a team of people who's like, that doesn't make any sense, Charles. Check David: Right. Drew Donaldson, Automata Intel: that out right now. Like they just they get into this kind of mindset of like, yeah, this makes sense. And I'll put this button here. And yeah, no, it totally makes sense to move this settings out of the settings and into this other window that no one's ever gonna look at because I look at it because I have a huge monitor. So it's like David: Right. Drew Donaldson, Automata Intel: I think that's the big thing. It's like it's not do marketers or engineers make better products. It's My background in marketing influence the way I look at products. And so instead of focusing on creating the sexiest stack that's going to get me, you know, a million likes by engineers on GitHub, I'm thinking first of how do I build something that everyone can use? And that and that then makes the products better. Because if you if you build the other way around, if you're if you become so obsessed with the technology stack, David: Right. Drew Donaldson, Automata Intel: no one cares. Like It's kinda like the I always quote that scene from Joe Dirt where he's going and buying the fireworks and the guy just has like the apple jacks and the snakes. And he's like, Where's all your like whiz bangs and zippity doodas like all these things? And he's like, I don't like those and he's like, It's not what you like, it's what the consumer likes. It's like that's David: Right. Drew Donaldson, Automata Intel: it. Like that's the whole thing. David: Hm. I d you know, that that's neat perspective. I kinda look at some of this this space in particular and it's it's it's got a lot of similarity similarity to like early social in that there were, you know, at least what I see of many markers is they're you know, they flipped a switch, they went from SEO to AEO immediately. And they are running a playbook of kind of social media expert, but now AEO expert. and they're they're vibe coding their way through something that's a little bit soft. And it it what it what it's doing on one s on one sense it's it's filling a market with a need, which to some degree validates it, but on another it it it's creating a little bit of uncertainty around what is the value and the discernible value and a little bit of false hope. It's interesting to kind of look at the the AI market overall. and I won't ask you to respond to that because I get you don't want to say anything untowards towards anyone. I'll say it. I said it. Drew Donaldson, Automata Intel: Well, no, I I will tell you, ca so the beginning of last year I saw the writing on the wall. I knew AI David: Yeah. Drew Donaldson, Automata Intel: was coming for marketing. I knew it the moment GoDaddy aired the Arrow Super Bowl commercial. I saw that and David: Mm-hmm. Drew Donaldson, Automata Intel: I looked at my numbers and I was like, websites are thirty percent of my business and GoDaddy just made me irrelevant. Right? David: You're in trouble. Yeah. Drew Donaldson, Automata Intel: Because it doesn't matter how good of a website I build if People look at Arrow and say, well, for free and a GoDaddy, and I'm buying a GoDaddy domain, it can just build me a website. It doesn't matter how good my website is because my value now goes to zero. Because they think they are can get the same quality out of GoDaddy as they can me, because not enough time has passed, right? So my thinking was this is not going to be a blip. This is going to be a cataclysmic shift in the marketing dynamic, which to date was US marketers versus overseas marketers. That was a game we could win. This is not a game we can win. And so I knew that in February and I started repositioning. I knew by the end of the year I was likely going to close my agency down and start this agency. And so we completely inverted our model and the stuff we used to provide for free, not as for free, but as part of the package, but it wasn't the sale selling point. That became the selling point. And then we shifted, we we just kind of eliminated the marketing and over the course of that year I saw the interest in these tech services skyrocket at the same time marketing implode and by the beginning of Q4 I was feeling the burn and so were dozens if not hundreds of other marketers because I would almost every day and I forwarded every single one to my wife because I I like when I'm proven right because I told her in February I was like marketers are going to start dropping like flies by the end of the year. In October, almost every day of the week, I would get an email from a marketing agency that I'm connected with in some way, you know, did business with them or was on their podcast or a friend of mine or whatever. Update on our business. We're changing paths. We're going into N8N automation. We're going into, you know, identity resolution. We're going we're transitioning into AI consulting. Every single one. Same story. I knew what they were going through. I because I saw it and I was doing I was doing perfectly fine in marketing. It wasn't until this moment that I realized like, this is this is this is bad for all of us. Like this is gonna go. And thank God I caught it then and not in dis in October of that year, because I wouldn't have had the time to build this kind of in secret and actually be able to launch on January 1st with people. already lined up to buy products. So it was I think a lot of them are coming from us like the the the uncomfortable truth about marketing is that marketing is not nearly as complicated as marketers make it seem to be because that's how marketers get paid. Mar like I hate to I hate to be the magician that reveals the trick. That's how marketing works. We make it sound more complicated than it is most of the time. Big asterisk most of the time. Yeah David: I'm a marketer, it's very complicated. Okay, Drew? Just to just so when I'm not doing this show, I am a hundred percent a marketer. It is the most complicated thing in the world. I'm gonna just politely disagree with you, okay? Drew Donaldson, Automata Intel: You you can just cut all of this out. Just cut all of it. are parts of marketing that are coming David: Alright. My my part is especially complicated. No, but I I I hear what you're saying. So Drew Donaldson, Automata Intel: Yeah. David: just staying on that, when you when you saw that, outside of marketing, just as a as a business owner, you could see something and I I it that moment is not happening exclusively to marketing right now. it's not. There it's Drew Donaldson, Automata Intel: No. David: It's it's it's happened in a lot of places, but just headspace on that moment. You're watching this ad and is it shit or is it or is it like okay, here we go. Like I need to I d I know exactly what next is and kinda just as a as a business owner, what what's what's your headspace at that moment? Drew Donaldson, Automata Intel: So my wife had gone to bed, because we had my daughter was young at the time and sh my daughter is a terrible was a terrible sleeper when she was a baby, so my wife would be up all night with her. and so she had gone to bed early and I was sitting there watching the Super Bowl, just relaxing. And as that commercial came to a close, the it's like the entire room darkened. It was the biggest sense of dread I'd ever felt about my business. Like we, you know, listen, everybody goes through bad months. Everybody has deals that they were really counting on that don't come through, right? That happens. This was the kind of dread that was like You know one of those movies about like an apocalyptic event where the apocalypse happens right at the end of the movie and there's always that scene where people are sitting around trying to be normal right before the apocalypse. Right? That was the feeling. It was like an asteroid is coming, there's nothing I can do about it, and we are all doomed. Like that was it. And it was crippling. It was like after a week after I could not think of anything else but like how I'm, you know ta I'm I'm carrying myself to my own funeral by being in this industry. And so it was it took me probably four weeks to come to the like, all right, so what do we do? What's the part of the business that AI can't do that's still valuable, that's still hard, that's still complicated. And it wasn't the creative piece. The creative piece was too easy. It was too easy for people to think they were good at it. You can type into Claude right now, write me a landing page, and it will write a landing page that to the untrained eye, to the not trained marketer, will probably read pretty good. Right? It won't be very good. It'll be the sum of a bunch of parts. It won't be creative. It won't have any unique tone or style or flair, but it'll be sufficient for most business owners in terms of what they consider a good landing page. And So I just stripped out every piece of creativity I could and focused only on the technical for the next year. And I gotta say, there is something very relieving about no longer having your work judged subjectively. you know, like I used to I we never did branding. And the reason I never did branding was because I hated feedback from brand went from branding clients. It was always like, It looks old. How so? I don't know. It looks old. Can you try again? Sure. I'll tell them not to do it old. And it's like, you know, a modern low, like a mo very modern hip logo, and it looks old. I don't know what to do with that, right? Now there's people that they s they love that challenge. I don't love that challenge. I like building a system that works. And so I would write an ad and send it off and be like, all right, you know, this is what you're gonna read into the camera. I don't like this. Based on a formula, we have perfected, we have a laundry list of clients that run ads just like this. It works. Just read the script. Like it it there's nothing objectionable in it. Well, it just doesn't sound like me. What doesn't sound like you? Well, I would never say this. Okay, write this. And it was just always this sub subjectivity of our work of whether we're not, whether we're doing enough. And if you are too good at your job, then people think it's easy and don't want to pay you as much. And if you're bad at your job, you get fired. So it's a lose lose for marketers most of the time. and you know, I I'm I'm David, in case he didn't introduce me, I'm just David's inner monologue. This is just what goes through a marketer's head every day of the week. David: Yeah. Drew Donaldson, Automata Intel: And it it's awful. It's awful because you can sit back and look at a campaign and be like, crushed it. Look at those numbers. They're gonna be so happy. And you walk in that client call thinking like they're gonna raise me on their shoulders as as the king cometh. And then they go, Yeah, so that didn't really work, did it? You're like, What are you talking? Did like 40 leads in 24 hours? Yeah, we were hoping for 80. You know, and and you just have to suck it up and deal with it. And so losing that subjectivity, that that subjective grading standard of marketing and and just being graded on, hey, when I push the button, does it turn on? it's such a relief. I I cannot tell you how nice it is to just be graded on a pass fail rather than this weird sliding scale of. And it's not even one sliding scale. It's like a six axis sliding scale because the marketing director loves it, CEO hates it, his wife loves it though. So maybe she'll influence him, but the son thinks you're an idiot and he should fire you right away. Like, great, cool. This is awesome. I love marketing. David: I love that. Drew Donaldson, Automata Intel: But like this is so binary. It's like I had a meeting with a client earlier, and their biggest concern was like, Hey, can we add a check button, like a checkbox that does this? And I'm like, Yeah, sure. And they're like, besides that, everything looks great. Because it works. It just they click the button David: Yeah. Drew Donaldson, Automata Intel: and the right thing shows up. And it's it there's n it doesn't have to they don't get in the subjectivity because they don't understand the architecture of the stuff enough to actually have those kind of like deep discussions that would lead to any kind of subjective grading. It's just like I don't know how to do this stuff, and that's why I hired you. Whereas in marketing, a lot of people think they're great marketers and so they have opinions. You don't have that in the tech side. People don't have opinions 'cause they don't know enough to have an opinion, which is great. Yeah. David: yet. Wait till they start bobcoding. Then everybody's an expert. So you you got Drew Donaldson, Automata Intel: Yeah. David: you got some runway. You got you definitely but I hear I that makes sense. I mean I I do I do live that. so yep. Drew Donaldson, Automata Intel: No. David: All right. So let's come back to automata real quick and question just right out of the gate. Business owners are getting ready to do some AI integration. What what do they suck at? Right, what do they get wrong and what would be your one piece of advice? Drew Donaldson, Automata Intel: the number one thing they're bad at is prompting. Which is so stupid. Because David: Okay. Why are they bad at that? What's the what's the hold back for them? Drew Donaldson, Automata Intel: Because a and I'll I'll give everybody a trick to this at the end. They don't people are very funny with AI. 'cause I've watched a lot of people use AI. I watch my wife, who is, you know, retired from work. She's just is a stay at home mom and raises the kids. and sh her experience with AI is probably about what anyone else has. Like you'd think, you know, being married to me that she would be like some kind of like AI ninja, but no, she's she's just like every other person out there that uses AI for this or that, doesn't not really hardcore. And because people don't really understand how it works and they don't understand how it processes information, they don't talk to it like they they underestimate it. And so instead of asking the question in its entirety. And structuring it, like you might get a an email from a, you know, if you're a politician that's in trouble and you get an email from a journalist with like 15 points that they have to ask. AI can process that and you'll actually get a better answer. And then the other piece is that they only go one round with AI. Like working with AI is like a a boxing match, right? Like you're giving up after the first punch has been thrown. Right. And it's not necessarily giving up, but you're just accepting whatever AI says. The thing that makes us successful in our use of AI is that I am constantly pushing back. I'm constantly saying, You're absolutely wrong on that assumption. Like I'm dealing with a Postgres database issue, right? AI agent was convinced Postgres was not installed on my machine. I was like, you are a hundred percent wrong. It is absolutely here is the directory. And it goes, my bad. It's like, yeah, you're bad. Like it's it's right there. It didn't find it. But a lot of people would take that and be like, I guess I guess my computer just can't do that. I guess AI can't do that. Or they'll ask it in a way that will be like too casual. Or or they'll use terminology that the AI won't really understand what they're what they're trying to get at. Like what is the goal of talking to me? And so like the way I always tell people is like you have to treat AI as a very smart but very lazy intern. Right? It's like Bill Gates ideal hire. He's always said like he'd rather hire a lazy person, a brilliant lazy person than a dumb, busy person, right? So like David: Great quote. Yeah. Drew Donaldson, Automata Intel: yeah, it's it's it's it that's AI. AI is like that lazy. Incredibly smart MIT educated intern, right? If you don't push it, he's gonna give you not the best answer. If you push it, if you ask pointed questions, if you make it think, it's going to give you much better results. When people go in and they type, write me a blog article on this or write me a blog article for my business, it's lazy. It's gonna give you the lowest common denominator, right? It's trained to do that. These companies are losing money, they cannot afford. the the the the amount you pay on tokens is not their actual cost. So they have to make these things so efficient. So the way they do that is shortcuts. They say David: Yeah. Drew Donaldson, Automata Intel: what's the lowest common denominator that could answer this query. And if you just put in write a blog post, you're not going to get a good blog post. You're gonna get the lowest common denominator blog post for your business. But if you are hyper specific, if you say, I want to write an article based on this you know CNN report on this happening in this part of the country. I wanna pull in labor statistics from the DOL. I wanna go and look through tax law and pull out, you know, these parts of the tax law that just passed and how they're gonna affect. If you give it that kind of structured work, it will do it and it will outperform your wildest imaginations about what it's capable of doing. But too many people just tiptoe. They ask they ask How do I get my kid to go to bed? And it David: Right. Drew Donaldson, Automata Intel: goes, and it gives you the same advice every other parenting blog is going to give you. But if you say, my child is four, she has ADD, she does not do she doesn't like eating dinner. We have her on these vitamins. this is her daily schedule. If you feed it all that information, all of a sudden you get a whole different set of recommendations that no one else has told you. Or things that at least make you ask the follow-up question, be like, okay, well how does that work? Like, or how, how could I do that with my kid? Or or I don't have that, you know, I don't have that capability where I live. What else could I do that's in the same line of it? And so that's where you really unlock the power of AI. And so it's it I almost feel like there's that there's that have you ever heard the the the Guru on the mountain story. David: No. Drew Donaldson, Automata Intel: Okay. Very quick story, right? So there's these three monks and they're in training, right? And as part of their training, you know, they rise through the monkhood, and one of their rewards is they get to go to the top of the mountain and ask the guru one question. Only one. And they only get one opportunity. This is the one and only time. They've all done a lot of work to get here. And so three of them head up the mountain together. And they're at base camp, and they're only can go up one at a time. So the first one goes up. And he goes up to the top and it's a long walk and gets to the top and he goes to the guru and he's like, Guru, what is the meaning of life? And the guru looks at him and he kind of cocks his head and he reaches into his pocket and he pulls out a walnut and he puts it in his hand and then he sends him down the mountain. And this whole time, this guy is looking at this walnut like, What does this mean? Well, David: Right. Drew Donaldson, Automata Intel: why did he give me a walnut? He didn't even answer my question. Like, do I get a redo? Like, this is insane. So he's so he's so in his head that he just walks right by the other two. And the other two are like, he's been like, he learned something. And so now they're really excited. So the second one goes up, goes up the top. Guru, what is the meaning of life? Guy looks at him again, reaches in his other pocket, pulls out a walnut, puts it in his hand, closes his fist, sends him down. And now this guy, he doesn't have quite as good an attitude about it, right? He's not trying to look for you, David: Okay. Drew Donaldson, Automata Intel: he's like, What a ripoff. I've dedicated my whole life to being a monk. This guy gives me a lousy walnut as the answer to my question. Like he's no guru. So he's starting to lose it. So he gets down, and the the guy, the third guy's like, what what happened? He's like, he gave me a walnut. And the other guy, who has now come back and joined the group, is like, he gave me a walnut. And the they're they're like going, you know, good luck to you, man, but I guarantee you're gonna come down with a walnut. The guy doesn't answer questions. So the third guy goes up. And he's up there a long time. I mean, seasons are changing, right? And these two monks are just sitting up there at base camp. They're running out of food. Like they've seen all sorts of weather and they've just had to endure it because they all have to go down together. And so they're just waiting for this guy. And finally, six months later, this guy comes down from the the the top of the mountain. What took you so long? Why were you up like we were up and down in like an hour? It's been like six months. What did you David: Right. Drew Donaldson, Automata Intel: ask him? The monk says, I asked him why he gave you guys walnuts. David: Okay then. Drew Donaldson, Automata Intel: You have to ask the right question. Too many people David: Love it. Drew Donaldson, Automata Intel: ask AI the wrong question. They ask the easy question. They ask the question that's not gonna get them the real answer because they don't know how to ask the question. So that's the biggest thing is like don't focus on all the shiny stuff. Focus on just the how am I communicating with this? And how do I get better at communicating with it? And how do I push it because it's way more capable than most people will re ever realize. That's why I'm saying I'm I'm I work on this stuff 16 hours a day. I am Every day shocked at what it's capable of. David: Huh. There you go. Drew Donaldson, Automata Intel: So yeah. David: Love it. Alright, True. We're kind of getting to the end here. So I do have a request of you, of most of our guests. we we've had a a lot of really wonderful contributions on what we call the flip story recipe book. So I'd love to to get your recipe, if you don't mind. Drew Donaldson, Automata Intel: I would love to give it to you. this is this ties perfectly into the the kind of the prompting focus. I want to introduce you to the idea of meta conversations with AI. Most people have a transactional kind of relationship with AI where you ask it a question and it responds, or you ask it to do something in the case of an agent and it does the thing and then tells you what it did. A meta conversation with an AI. is where you ask AI how you can better prompt for a specific outcome by instead of focusing on structure and how to pack all the stuff in, say this is what I'm trying to do. How do I tell you to do it? This is such a secret hack of like I don't know why people don't do this. I do it all the time. It is so powerful. You can do it cross-platform. Like that's one thing I've tried is I've done like I've had Grok write me prompts for Claude and vice versa. You can also just do it with Claude or Grok or whoever your favorite is, right? Like you just say, This is what I'm trying to do. I need you to write me a prompt that I'm going to feed an agent that accomplishes this. And I want you to think of all the things that like I didn't think of. And it will go and write you the most perfectly structured, beautifully outlined, hits all of your points and 10 things you didn't think about, that you then just copy, put it into a new session and hit send. And it will do wonders for your output. Absolute wonders. And it's it's the kind of thing that people don't think about because they're thinking too transactionally and not relationally. It's here to help you, but you you it it's not gonna know how to help you unless you tell it. And so by admitting that, like there's plenty of times I tell AI, by the way, I have no idea what I'm doing in this. Like this is kind of foreign territory for me. I don't know if you're even able to help this, but here's my idea. Can this work? Is this feasible? And then if it goes, yeah, it's feasible, I go, Great, write me a prompt to do that. And then it will do it. And all of my work is done for me. Like I just then take that and plop it into wherever it needs to go and it works. phenomenally well. I have built so many prompts that are mission critical prompts for businesses using this methodology. And it there has not been one time where the output of these meta conversations has been anything even close to what I would have come up with on my s own because it's able to think in such broader ways about a problem than a human can. We are very limited in how we think of problems and our our total experiences And our stuff. And it's not to say that I don't go and edit them sometimes. They don't, they're not all right all the time. But in terms David: Right. Drew Donaldson, Automata Intel: of I need to get AI to do stuff for me, and I doesn't ever do what I want it to do, tell it that. Be like, I I'm an idiot, and I just need you to hold my hand. And this is what I'm trying to do in plain English. Can you do something with this? Just write me a prompt. Don't tell it to do it. Just say, write me the prompt. Because it changes how AI is going approach the problem. Because when it has to, when it knows it has to prompt another AI, it is going down a different line of thinking than it is just trying to solve your problem. Because your problem now becomes the prompt. It's not trying to solve the problem, which you probably didn't give it enough detail to solve. Because you don't you don't think that way, right? Like it's a it's a new David: Right. Drew Donaldson, Automata Intel: way of thinking. So like I I use this anytime I run into an issue where I'm like, I am not getting the output I want. I use it anytime I feel like I don't know where to start on a problem. I use it to transfer data from one session to another. If I'm if I like start a project in one session, but I don't want to keep the project there and I'm like, I want to move this to a different one. Hey, write me a prompt to take everything that we've talked about in this session so that I can dump it into a new session and pick up right where we left off. It's so it's such a dead like this is like I'm teaching you how to make sugar water here, guys. Like this is this is David: Yeah. Drew Donaldson, Automata Intel: you don't need any fancy cooking equipment for this recipe. This is literally just admitting to AI that you're not as smart as it, which is understandable. Like David: Yeah. Drew Donaldson, Automata Intel: it's so simple and it will absolutely it'll fundamentally change the way you work with AI the first time you do it. And then once you do it, you'll never go back to the old way of prompting. You will. You will do this until you become very good at prompting AI and you'll learn through that osmosis of reading how it wants to be told how to do things and how it thinks about tasks. So it's it's the best it's like, you know, working on a car with your uncle. And like, you know, just watching what h how he's fixing stuff. That's that's that's what you're you're essentially doing. David: Huh. Love it. Very strong. Drew Donaldson, Automata Intel: I'm glad David: Man, fun. How how do how do folks find you? Drew Donaldson, Automata Intel: automataintel. com is our web address. You can find us there. That's automata, not with a automated that's a British company. but automataintel.com and then I'm on LinkedIn, I'm on X, you can send me an email. It my email is very simple. It's just Drew at the domain name, automataintel.com. So yeah, you can reach out and I'd be more than happy to answer questions or provide prompt advice or You know, whatever your challenges with AI right now, it doesn't have to fit into a special box. I like I like big, difficult, ugly challenges. Those are my favorite. So send Send me the David: Amazing. Drew Donaldson, Automata Intel: send me the thing everyone else told you was impossible. That's that's what I that's what I really want. I wish I could have like a bounty for those because like those are my favorite ones. It's like, well, I talked to three other people and they said this wasn't ca possible. I was like, okay. That's that's what I can sink my teeth into. I like making David: Love it. Drew Donaldson, Automata Intel: impossible things possible. So David: Magician. Drew Donaldson, Automata Intel: yeah, that's for that's right. David: man. This was fun. Really I really Drew Donaldson, Automata Intel: It was great talking to you, David. It really was. David: loved having you on the show. This was awesome. Drew Donaldson, Automata Intel: any time, any time I'd be happy to talk to you again. David: Alright, well, we may revisit this. Alright, folks. Well, we're gonna wrap it out here. we'll see you on the next one. All right, bye everyone.