6: Using AI for Direct Response Marketing with Stefan Georgi

All right, welcome to Using AI at Work. Today we're honored to have Stephan Georgi, a titan in the world of direct response, marketing, and copywriting. Stefan, known for generating billions in sales and elevating brands like Vshred and Tommy Chong, CBD is now pioneering the use of AI in business. He's the brain behind free ad copy.com, which is a tool transforming how we create advertising copy. So get ready for an episode packed with insights on leveraging AI for customer acquisition and operational efficiency. Let's get started. 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. All right, Stefan, welcome to the podcast. I'm really excited to have you here. I know that you're as much of an enthusiast about this topic even more so than I am, so I plan on learning plenty, as do the. The listeners. So before we get started, why don't you tell everybody what you think they need to know about Stefan?

Yeah, for sure. And thanks for having me, Chris.

My pleasure.

So what should people know about me? You know, my, my background is a direct response marketer and a copywriter, which is somebody who writes ads. Throughout that time, I've been able to partner with a lot of businesses like, and as an advisor, an actual equity partner, whatever it is, helping them to grow and scale, mostly direct to consumer types things. So for example, I helped a company called V shred go from I think maybe around 50 million at the time to 200 plus million per year. Did some work with Tommy Chong CBD, which you might have seen on Twitter now x. But really a lot of different kind of brands like that helping them to grow in scale. And so I've started a lot of different businesses, partnered a lot of businesses and anything. Direct response, marketing and acquiring customers and growing and scaling digitally is my wheelhouse in my jam.

But you also are now educating as well, correct? With your groups. Can you talk a little bit about what that looks like?

Yeah, for sure. So I partnered with a friend of mine back in 2019 and we created like a mastermind type of thing where, you know, we're helping other business owners to grow their companies and scale, get more customers, all of that. Since then, it's morphed into different branches of kind of training and education and you know, specifically since about 20, I guess, like summer into fall of 2022. So a little bit before chatgpt came out, really started pivoting to AI. And so our goal is to essentially like, if you look at our mission statement is the first word, I won't, I won't swear, I'll self bleep myself. But basically it's to help people not get effed by AI, right. And so to accomplish that we have really different things like whether it's software, SaaS, products, courses, all the way up to kind of expensive, like 30, $40,000 a year, kind of masterminds. But we're really trying to help people at all levels have access to AI tools, processes, things like that, that they can become more efficient, more effective, improve workflows, processes. And so they sort of embrace what's happening rather than get caught unaware. And it's a lot more painful because there isn't like, there's not, it's not, not going to happen, right? It's already happening, it's already going to accelerate as far as like the disruption and kind of like a global shift in adoption to, towards AI.

Well, based on that mission, I think you're in the right place as far as being on this podcast. So one of the things that I'm always curious about is. So my experience was I heard people talking about GPT. I'd heard about AI obviously a long time ago, but I heard about GPT and I'm a smart guy. I've been around tech and martech and all that kind of stuff for a long time, but I just assumed, oh that's, that's somebody else's opportunity. That's, that's not for me. How did you go from not being a user to being who you are today? Like what was that journey? Like, what was the first exposure where you had that aha, that oh moment.

Yeah, the first, the first exposure or the first I guess aha moment was in like early November of 2022. So again, chat GPT came out at the end of November. I'd already been thinking about AI a bit before that, but early November I did this like little two day thing with Agora Financial, which is a big financial publisher, lots of people, right. And so it was sort of a collaborative thing. It was not a paid, it was invite only. And essentially the idea was just get a lot of really smart people in the room and talk about like where the future is going, what are the next five years look like, what are the kind of biggest opportunities, risks, threats, almost like a SWOT analysis on the next five years with then the goal being that as leaders of different industries we can go kind of help, you know, guide our communities and the industries towards cool, that direction. And so it was really cool. Yeah, it was a really unique thing. I was the like facilitator of it and it just sort of came about organically. But there was, you know, very, a very eclectic group. So there's you know, guys like Porter Stansberry were there, you know, a woman who is from like the World Economic Forum, which I'm kind of right leaning so they get, you know, oh no davos and all that. But you know, she had different perspectives. It was a bunch of very high level, really interesting people in the range of industries. And when we talked about sort of the biggest shift or change that was coming that was going to catch people unaware, overwhelmingly the consensus of all these smart people in the room was like AI. And that was actually for me that moment because again, I'd been paying attention but not hyper focusing. And then we talked through it more and people kind of shared data and evidence of why they thought that was the case. But I left that two day thing kind of committed to exploring AI really deeply and then trying to be more vocal and outspoken. And so then I really started going and playing with different tools like at the time, like Jasper or writesonic. I went to copywriting and content first since that's a world I know really well. And then I was like, wow, you know, it's pretty good. And then I read this book, it's a pretty famous book called Crossing the Chasm. Right. And it's about essentially how, you know, technologies and innovations, whether they sort of make it to the mainstream or they kind of die in this, this chasm. Right. That there's not this linear sort of smooth progression. And for whatever. Exactly. And for whatever reason this was over Thanksgiving, I kind of had the aha moment for all the training stuff. I'm like, we really should just go all in on AI and lead there. And you know, I think like this is going to only get a lot bigger. And then I had no idea Chat GPT was about crazy time. And then like, yeah, like a week later or whatever it was like, I think it was the 30th of November 2022 that chat GPT came out. And so then it came out and you know, at that point I'm like, all right, so. So I guess I was mentally, I was lucky in a way because I was a weird kind of little confluence of events. Like I immediately embraced it because I'd committed to embracing it because I was thinking about it and you know, which is funny because it's not that much of a head start over anybody else, but it was enough time to be, I guess, because I was mentally prepared to jump in. Like, I jumped right into it.

That timing, that was the universe going, there you go. That was perfect timing. So once you got access to 3.5, kind of, where did you start? What did you, like, what. What was that process of learning how to use the models for you? Was it very similar to Jasper?

Yeah, and, you know, and I was never like a. Like a hardcore Jasper user. I'd played with it, and so, you know, like, I just started asking it questions and then. And then, you know, the prompt engineering part came pretty naturally to me because, you know, as a. As a writer and a copywriter, that was good. The other thing I did is it's still around, I think it's called Futurepedia IO, I want to say. But essentially it's a directory of all these new AI tools come out. But it's actually gotten difficult. There's so many tools now, but even then it was like 20 or 30 tools of note would come out per week, maybe 10 or 15. So I kind of made a goal of going on that every week, going through, looking at the new tools, signing up for accounts, doing trials, just trying them. So I can kind of understand the. A broader depth of the capabilities. Right, because copy's one thing, but there. There were, you know, tools for workflows and all kinds of other stuff. And. And then as of now, I think, like, 95% of the tools are dumb. I mean, not to be mean, but it's like sort of like, you know, a menu planner, and you're like, all right, cool. Like, there's a million. You know, like, that's not going to be the breakthrough killer after. It's not. Yeah, but occasionally you'd find a pretty cool one. Right. So I tried to expose myself to a lot of that as well.

So that's fantastic advice. Besides futurepedia, are there any other places where you go. And I mean, I know of best of AI. I know Matt Wolf with his future tools. Are there any other places where you go to kind of see what new apps and functionality is out there?

So one thing we did, which honestly I think might be worthwhile for other people, is we hired some, like, researchers to kind of like, go and start compiling, like a. Do, like a weekly report of like, hey, here's like, the new trends. Here's like, the, you know, new tools that are coming out, like, and then to Demo them and actually go in and like. Right. Talk about capabilities, the, you know, potential use cases as they relate both to our business and in general. And you can get like a VA or somebody to do that. It's not going to super. You know, we're paying like a few grand a month for a really talented person. And so, yeah, I think that's, that's one thing I go to. Like, you know, the only other one I regularly go to is, well, I'm on like Twitter or X now, so you'll see stuff there. And then like even actually, you know, funnily enough, like Venture Beat, which is obviously a well known publication. Right. But they're a good, kind of like a good marker of where the, the wind's blowing generally. Right. And so agreed. If you go there, like, you know, almost every story is about new kind of large language models or new AI tools or. And so, you know, that's the place I still check regularly. It does give me anxiety because then you see there's all these different generative AI things and they're all getting all this money and you're like, oh God, do I need to try this one? And you're trying to stay abreast of it all. But yeah.

So one of the things that I've seen recently is the suggestion that you don't need to be tech to be a good prompt engineer. You need to have a mastery of the English language or whatever your language is.

Right.

As a copywriter. And for those who aren't familiar with what copywriting is, these guys are masterful at eliciting emotion and action from the written word. They you may not know what it is, but you read copy all the time from every advertisement you see to the rate radio, TV commercials and that sort of thing. So somebody who spent as much time thinking about the impact of like an individual word would have within the context of a whole sales letter. Because I would imagine that you get that granular with that, that process. Right. It seems like you would have an unfair advantage in being able to elicit like good response, good output from the models for somebody who hasn't. I mean, they just consider, you know, communication just dashing off an email or something. Do you have any advice for them that would help them kind of compress that learning curve of knowing how to communicate with GPT, Claude, whatever.

Yeah, I mean, to me, you know, the short answer is that Claude is, is generally my favorite large language model of kind of like the big out of the box ones. Right. And the reason why is because you can just give it referential material and have it like, you know, immediately have that knowledge. And so, like, for example, if, for copywriting, if I want to write great YouTube ads, if I have a training on YouTube ads, whether it's a video and I get a transcript and turn to a PDF or it's like a PowerPoint or whatever it is, right. I can basically take that, go, hey, Claude, like, here's a training about how to write great ads. Analyze this, tell me about the structure. What are the takeaways? Whatever it is. Yeah, does it. All right, cool. Now I want you to write some YouTube ads for, you know, this product or offer or brand. Here it goes. Right? But again, and to not make it just copy only. Like, it's the same thing for, you know, like you do for legal, for SOPs, for SEO.

Right.

I talked to a group of several hundred agency owners here in Miami that were all, you know, fairly successful and mostly doing SEO. And so I literally just like Googled like best SEO practices 2023. And then I think there's interested in hyperlocal SEO. So I'm like, hyper SEO local. Like, took that stuff, gave it to Claude. Hey, here's like, you know, training on hyper local SEO practices, right? Like, based on this, like what's like a list of SOPs, what's a checklist? If somebody's going to do, you know, hyper local, like SEO, what would it be like? There it is, right? And then what's cool? I mean, not to.

What did they think about it?

They loved it. They. They were. I mean, that's the funny thing, right? I have this, I've done this talk a couple times live. Like, I'm not doing a lot of public speaking right now because I'm building so much, but I'm down here in Miami and I moved here. Yeah, it's like when people. If it's local, I'll do it. I'm generally not going to travel for anything right now. But it's funny because I've done a talk a couple times down here where it's really just like how to use quad. I feel like it's just like upload a PDF that teaches quad something. Now have quadrant, you know, kind of do what you just taught to do. And you know, but every time after that training, people are like, or that talk, people are like, oh, you're a genius. You know, like, oh my God. And to me it feels like, not that complicated. But I'm not trying to be like, arrogant. It's just like to Me, I guess it makes sense, but if you've never done it before, you know, so, yeah, they really liked it. And then, I mean, the other part, there's a tool called Agent GPT, which is like an auto GPT type of tool. So essentially for those who are watching, who don't know what that is, like, you know, you have, like, these agents. You have like a, like a, you know, AI basically goes and does, like, different tasks for you, and it does it over several loops. So it kind of iterates and can even improve or get better or go further along. So as an example, if I'm like, hey, Agent gbt, like, you're a hyperlocal SEO expert. We need to find, you know, se hyperlocal SEO content for this ice cream parlor. And Coral Gables is the example I use, right? There's an ice cream place I go to with my family, you know, like, so based on this, like, you know, go to work. And so it's like, you know, it's hard to describe, I guess, but. But basically it's like, all right, give me myself tasks, right? So then Agent GPT goes and assigns itself different tasks, like, all right, go find micro influencers, go find blogs and Instagram accounts that are in Coral Gable. It's like, go find keywords like, find, you know, social media handles. It just starts doing all these different things and like, you know, it ended up finding, like, find events that are in Coral Gables where we could, like, you know, either advertise or hijack, like, you know, SEO related to these events and have that. And I started doing, like, social media posts of emoji cons. Like, it literally can just do all this crazy stuff, and it's doing it on its own in the background. You just gave it one prompt. But the important thing is you kind of gave it the guidelines. So really going back to your initial question of, like, how do you, like, you know, how do you shortcut that process? I think it's like referential material and then like, bumpers or guidelines. You're sort of like, you know, like, here's the instructions I really like. They talk about natural language programming, and all you're really doing is you're, you know, you're giving a set of instructions to the AI and, you know, if you have parameters or direction or, you know, guardrails of, like, don't go over here, don't go over here, stay in here. You'll tend to get, like, you know, pretty good outputs, and you really don't have to know a lot. I Mean, it's pretty wild. I think I basically felt like I ended that, like, I could probably beat most of these people in this room on SEO if I wanted to go start an SEO agency, which I don't.

But, you know, was there, were there any fear or trepidation expressed by the audience when you showed them how easily it was for somebody like you with not, not an SEO agency owner to be able to get those results?

You know, it wasn't like outspoken in that specific instance.

It was a wake up call, I bet.

Yeah, there definitely was. A lot of people were kind of like, you know, generally when I, when I demonstrate, people kind of feel like, all right, yeah, I need to, I've been putting this off, I've been ignoring it. I've been putting my head in the sand. I need to, to step up. But, you know, on a broader level, I mean, adoption is still very low overall. And like, you know, I remember I was reading something in the Economist a few months back and it was sort of like, they're talking about global firms and I'm going to totally make these stats up. But roughly it was something like only 30% of them said that they had used AI within their company, and out of that, only half of them had decided that they were going to keep using it regularly. And so there's this really big opportunity I think right now for smaller business owners to gain a lot of market share and just move a lot faster. Yeah, the timelines, the way I kind of talk about, as a quick note, there's sort of like legacy timelines and there's like kind of today timelines. Right? And it's like the legacy timeline of all these different processes taking weeks or months or whatever it is. And most people want to operate there because they're comfortable and understand that. But the truth is, if you can, you know, generate advertising, collateral images, creatives, videos like sops like you can automate, like we're actually like doing things like automations, like all these sorts of things, you know, in like minutes, then you're gonna win. And so, you know, you have to kind of shift your mindset on these, on the way the timelines with which things get done, I feel.

So this is interesting because, you know, in the old, in the legacy timeline, we felt that there was a, a velocity or a tempo of getting things done right, and our behaviors outside of production or outside of work were built around those. And now I'll give you a perfect example, we're doing some stuff. And I worked a lot over the past Couple days more than I normally do. But I, I think I was able to get about two weeks worth of legacy production done in about two days. And this isn't pure synthetic output. This is synthetic that gets massaged and reworked. Like, this is me participating with it, not just hitting a button and letting it happen, like final output in two days. That opens up. I mean, the first thing, when you were saying that, I was, I was thinking there's going to be a lot of people who, if they start using this, they're going to say, oh, I've run out of things to do.

Right?

Oh, I got it all done. So it opens up this conversation about business owners and employees and consultants needing to be much more intentional about a broader plan. Because if this thing can eat a, you know, your small, your weekly plan in a day, and now you're out of stuff to do, you're not out.

Of stuff to do.

You just haven't thought ahead. You haven't built out your timeline, your strategic execution plan for longer than this week or next week. So is that, did you encounter any of that, or do you have any feedback perhaps on where to go with that? I know it's kind of a. I think of some weird stuff.

No, I don't think it's that weird. I think it's a good question. I think there's, you know, validity to that. I mean, you know, I don't, I think one thing I, I, one idea that kind of struck me early on was that AI doesn't, it doesn't make us less creative because there's. There's this knock kind of right out the gate from people like, oh, we're gonna just, you know, outsource our thinking to AI. We're gonna, you know, we don't like creative types and, you know, don't have to be creative anymore. They tell AI, But I think that's bs. I think it unlocks what I would call like a meta creativity, which means instead of having to like, you know, burn all those calories in your brain about the right word choice or the right image to use or the right. Whatever it is, it's like that part is handled. So now you can zoom out. It's really like the Holy Grail, right? Like, everyone's always been talking about this forever of like the whole, you know, work on your business and not in your business. And it's like, this is like one of the most enabling things to be able to do that. I mean, probably the most enabling thing of our lifetimes. And so, you know, and if you look at it that way, then nothing really changes because it's sort of like, all right, what are your, you know, your goals, your objectives as a business, like your KPIs, where are you trying to go? I mean, maybe you're expiring those timelines. Like what you wanted to accomplish in five years, you now try to accomplish in a year, you know, but to me, like, if you didn't have a clear picture or vision of where you're going and, or your vision was small, I don't mean that in a demeaning way. It's just sort of like if your vision is like a lifestyle business, right, Then you might get there really fast. But cool. Congratulations. Go write the novel you want to write. I mean, as far as I've experienced that, I mean, we're doing a lot of stuff, right. We have two kind of active SaaS, projects, products right now, an agency, we're launching some different brands and things. We about the training and things. But you know, I found myself over the last couple like Christmas, New Year's.

Yeah.

You know, having a good amount of time because I've got a good team and then we've got good processes and we're using AI for all this stuff. And so, you know, I'm doing some like wild stuff. Like I'm, I mean, I'm writing a book, I'm doing essays and things like that. Kind of like a lot intellectual things that I've wanted to do, but I feel like I never had the time, but now it's like I actually have the time to do it, which could then have even more of an impact, right? Yes. And so yeah, I think, I think really that's the way people should, should view it more, you know, then versus like, I guess that's not how. Anything to do. Yeah, yeah.

So inside your team, you mentioned that you're a process driven company. It's the only way to be able to get, you know, as much done as you guys are getting done. Where did you start with the introduction of AI into the process side of your business? We'll be back after a quick break.

Do you want to start your own unsexy business offering a local service or a trade to your local community? Come check out the owner ops podcast. We bring on people like Sam, who has built a $22 million roofing company in five years. I've also brought on people like Brock, who generated $1.3 million of revenue in his first seven months installing septic systems. If you are looking to build Your own local service or trade business. Come check out the Owner Ops podcast. That's ownerops.com o w n r ops dot com.

Yes. I mean, you know, it started with copywriting first and foremost because that's a core competency and it's one of those high value skills. And so, you know, but even that was really interesting because. Okay, like for those who are watching who don't know like what a sales letter is, right, Or a video sales letter, essentially it's like an infomercial script. It's online, but they're like 6, 7, 8,000 words long, very specifically structured, a lot of like, nuance and all of that. And so, you know, we were like, all right, we're going to, we should be able to write a lot more of these because of AI, right? But we had to kind of go through a lot of processes there where essentially, you know, I don't want to get too in the weeds on just copy. So I want to talk about other places besides that. But basically from kind of like, you know, using AI for an outline and some of the research, then getting to like, then we were paying writers $5,000 per VSL or video sales letter where we used to, you know, pay a lot more. But then there, it still wasn't that good. And so basically it's at the point now where we have like an internal SaaS where it's like, put in 20 inputs, push button boop. There it is. Push button boop. It's been copy chief, which means it's been edited and improved, right? And it's like, so there's a lot there, there's like. So it started off with like, how do we do this? Like, what are the prompts? What's like the engineering, all that. And then it became like, how do we automate this? Right? How do we, you know, create like an actual software that, you know, enables us to do this even more efficiently and there's a lot of cool things to solve within that realm. It's actually very difficult to do from my, you know, I don't think anyone's really solved it as well as we have. There's a few other people doing things like, and this is awesome. I'm not trying to say that as a slight to them, but we just, you know, put a lot of work into that. But, you know, then it becomes like, okay, well what about everything else? So, all right, emails, right? Every company, you have to be a copywriter, Your company almost certainly sends out emails. So, you know, if we're going to email our lists, different lists of kind of buyers and prospects every single day. Like, okay, what does that look like? How are we generating emails efficiently and effectively? You know, we. We have, like, two people who are sort of in charge of writing emails, but if they start trying to do it manually, we yell at them. So kind of going back to what you and I talked about a little bit, maybe before we were recording, but it's like. Because the biggest challenge of adoption, I think, is, like, you show people this better way, and they're like, this is amazing, but old habits die hard. So then they sort of start to go back, right? And so, you know, like, another example would be, like, on the sales letter things, it's like I hired two people, taught them my processes. Once they were really dialed in before we had it, like, kind of push button, automated level. And this guy came from a financial background, actually, and it was two, like, health sales letters, two infomercial scripts, right? And in the first week, he wrote, like, basically two and a half in his first five days. And he's never been health, you know, kind of for, like, a supplement type of thing before. And he's like. And they were good. I'm like, yeah, like, we. We sat through it. I mean, I'm like, this is amazing. And then the next one took, like, 14 days. And it's like, why? It's because then he, like. And he had to go travel back to, like, Baltimore to. To pack up his stuff and move down here to Miami. So I wasn't in office with him, and he basically started going down rabbit holes and kind of diverging from the process and all of these things. And so we had to be like, hey, man, like, you know, I get it, but, like, trust the process. You don't. You know, you're now making this way harder than it has to be, right? And so you see that a lot. But then other examples, you know, like, we have a really great coo, and she was kind of. She loves quad now, but it took her a while. But now it's like every single thing that we do, right? Every video, every transcript, every meeting, like, we've got Otter running. There's other transcript, you know, AI transcription tools that, like, I can't imagine. I know Zoom does it, too, but, like, not using something to have a transcript of every single call you do in every single meeting. And then you can put it into, like, again, you know, Zoom can do it. Loom does it if you do a screen recording. But, like, then we'll, you know, she'll put it to Claude like summarize it, you know, an hour long transcript, an hour long call, two hour long call. Summarize it. What are the key takeaways? Like what are the action items here? Who had what, what was said right now, right now Take it and do a slack message to like our leadership team or an email that kind of like summarizes the key points, talks about who took on what responsibilities when they said they're going to like all these things. And so now she's able to like you know, just be so efficient with it. So things like that are just really, really valuable from an efficiency standpoint.

Have you guys taken as far, taken it as far as once that output's there that automations take over and they start doing this for the, the administrative person or whatever?

Yeah, I mean we, we're playing with that stuff and I can't tell you that we have that so dialed in but that's definitely where it's going. There are a lot of like certain, certain things that, yeah, there are like, you know, if it's like a research thing or like generating documents or things like that, you know, there's other processes where we're still, you know, there's some human touch points in it. But I mean realistically that is probably more our fault for not, you know.

Not how much efficiency can we get. You know, it's like yeah, at some pointless. So one of the questions you mentioned that these, these copywriters are going back to their old ways and I would imagine that any business owner or, or employee or whoever, they, they're going to learn this stuff, they're going to come back to the team and say oh my gosh, you got to use this stuff. Nobody's going to look over their shoulder and go oh that's cool. And then they're going to go back to, you know, hitting rocks the old way. Any advice? I mean trust the process was a fantastic bit of advice. But any advice for business owners who may encounter, I expect them to encounter that. But what would you tell them?

Yeah, I mean I think there's a few things, I mean it's, it's tough. It depends on the size of your organization. Like you know, in my case I'm, I'm building this team down here and you know, we're actually hiring for at least three more positions. We have a call tomorrow where you know, we might bring on a pretty big new kind of agency client which will be nice if I hire you know, three more people on top of that. So you know, and I want people in office, which is funny in this age of remote work and AI, but I just find that there's a lot of value to that. But, you know, but I'm. But I am hiring very intentionally for people who are, like, already using AI that are excited about it. So I think, you know, moving forward, like, being proactive about that is really important. Right. You know, as far as, like, the people who have been with you for a long time who maybe are more resistant, you know, it's. It's. It's tough. I mean, I think one clear thing you can do, and I think Eddie Maloof said this, I think Hermosi, Alex Hermosi said something similar, but essentially telling them, like, look, if you embrace this stuff and become more efficient, like, your job is safe. If you don't embrace this stuff is where your job's not safe. Right? Because I think part of the. There's part of a stubbornness of, like, you know, teaching an old dog new tricks, and part of it's like a fear that it's going to replace them. Like, well, if I do a good job of this, they, like, they want it to not be good, because if it's good, it could mean that they're replaceable. And so I think if you communicate to them, no, it is good, and it can make you great, you know, then it could help alleviate some of, like, the fear and concern. Right?

Yeah.

And, you know, and probably there's other positive reinforcement things of, like, if they get stuff done way faster, I'm. You have to be careful of this. You want to set an expectation of, like, they're now suddenly only working half the time, but, like, yeah, some of the times, right. Be like, hey, great job. You know, you use the AI. You did exactly what I asked. Like, this looks really good. You know, like, I want you to see the fruits of your labor, like, take the next two days off or something like that. Or like, you know, if they're like a W2 employee who. That's valuable to them. Right. You know, ways to incentivize them to sort of see, like, the reward of it. Or maybe they have, like, a passion project within the company saying they want to pursue that you don't have the resources or the time. Right. And you're like, hey, this is great. Now that you are so efficient with this. Like, let's talk about that. That thing that, you know, you were even really excited about. Right. And how can we use AI to do it in the most efficient way? I mean, like, that's really valuable. And we're doing that internally for sure. There's, there's things where a lot of stuff I would have said no to as like a leader and CEO of, of, you know, a bunch of different companies, like I now say yes to because I'm like, but the, but the caveat's like, how do we test this out like an MVP in the most efficient way possible, you know, and with AI, we can do that. And then I guess I'm excited. Then they're trying to engage with AI, figure out how to use it, how to use these processes. They're learning more. Even if, you know, the, the test doesn't result in the outcome they were hoping for, now they have these, these new skill sets and now they're thinking that way. So that's another really, really valuable way.

You brought up a couple of good points. First off, that, that, like incentive or rewards program, even if it's only occasional great idea. And when, when we're working with companies, one of the first things we tell them is that, look, if you're going to get this, this AI initiative team together, you better screen them and make sure that they're already using the tools and that they're enthusiasts, because otherwise you're going to get somebody that wants to see failure. They want to keep things the way that they are because of fear, because of misunderstandings. And the way that I've heard what you were saying, the way that I've heard that quote is AI is not going to take your job. Somebody who knows AI will take your job.

Yeah, exactly.

The way that you guys are handling it is fantastic. So outside of work, are there like lifestyle applications? I know you were joking about the GPTs that can make menus and stuff like that, but are there. I mean, I would imagine that an individual who's running a bunch of businesses, that you have resources that need to be deployed. And by resources, I mean investment opportunities or bandwidth for your attention and things like that. Do you use the tools to help you prioritize or explore? Like, of all the roads open to me, which one is the best choice? Do you do anything like that? Strategic?

You know, I've, I've played with doing that. You know, most of the time it's, it's more like, it's like ideation. It's more like about testing ideas and they don't have to be related to business. But like, I'm like, yeah, I'm like an avid reader and so, and you know, right now I'm reading a Bunch of books from sort of doomerists, like the kind of people who are like, very educated, who are very afraid of like AI and biotech and all that kind of stuff. Right. And I really try to read, which I'm not a doomerist. I'm like an optimist. And like, I think innovation solves all of our problems, but not all of them. But I mean, ultimately we innovate out of problems. We don't sort of hide from them. Right. And that's my philosophical orientation to the road. But, you know, they'll elicit interesting thoughts. Right? And there's things where you want to kind of experiment out or you're like, all right, well what would like, what does that look like? What is like the long tail of this idea? And so I'll go to like, Claude or usually it's cloud, and I'll kind of ask it questions and be like, what, what am I not considering? What are like, rebuttals? What are different angles to this? Right. Or if I'm using GPT4 with like a dolly or whatever, like, can you do a visual representation? Like, what is. If we were trying to like, to encapsulate what we just talked about into like a. A picture or an image or a graph or whatever it is that could like really, you know, represent this, what would that look like? And so I, I use it, like, as an intellectually curious person to like, synthesize data, things of that nature. You know, I think, you know, occasionally I've used Agent GPT to go down rabbit holes. I mean, I saw. This is a really bonker movie. It's not bonkers. It was really interesting. Like I saw yesterday this guy had this paper published from. He was a professor at University of. I think it was Louisville. It's basically about under underground nuclear testing and that it's like impact on climate change. And I'm like, well, I've never even heard of that. And so, you know, I kind of found that paper. Then I started googling and looking at other papers and there's a bunch of really interesting stuff there about like, the environmental consequences of like, not nuclear energy, but like, you know, detonating tons of nuclear bombs under the ocean for like, you know, many decades. It's like, it kind of makes sense that that would have some environmental impacts. And so what I'm going to do, I haven't done it yet is, you know, probably have Agent GBT or another kind of auto GBT process. Like go compile like every study you can, and you Know, do an analysis of it for me and like look at the data and so, you know, if I had an investment thesis, I would definitely do the same thing. You know, I probably wouldn't base my investments purely on AI at this point, you know, which by the way though, this is one more, this is a really funny aside, but I'm in like a fantasy football league with a bunch of other marketers and stuff, right? And you know, we have our draft and like this, this year I, I basically couldn't make it. So I had it like the, you know, the. And I'm like, well, you know, AI is getting really good, so I'll let the AI do the auto draft. And I drafted this team and I'm like, I look at the team and I'm like, oh my God, I hate everybody on this team. Like, you know, it's a funny story about how AI is not that good yet, blah, blah, blah. Long story short, last week I won the league and I'm like, like maybe I'm the only person who like auto drafted with like yo Yahoo's AI. Now I made some moves. I would give myself a little bit of credit. But you know, so going back to. Maybe it's the same thing with investment. I'm like, you know, I'll try to outsmart the AI and it would actually be better just if I just did what it told me to do.

You know, it's. It's curious. You've mentioned a couple times that CLAUDE is your default. What. What do you identify as the difference? Because those CLAUDE and GPT. Of course there's Bard and Bing, but most people are going to be familiar with CLAUDE and GPT. What are the, what's your interpretation of the differences between the two and why you would use one over the other?

I just think that Quad's document analysis is like, is superior. Like if I take a PDF of really anything and I put it into GPT4, even like the sort of larger context window, whatever it is, I just find that the analysis and the outputs that I'll get are just really inferior to claude. If I give Claude, it's done some tests on this too, where essentially anything up to. I forget the token limit, but I think it's over 10,000 tokens. I want to say it's 12 or 15,000 tokens, which is essentially 8 to 10,000 words. Like it essentially retains and understands like all of it, right? And then if you, you know, Cloud 2.1 goes up to like 200,000 plus tokens or whatever and Again, for people, tokens are like, you kind of think of them as like characters or words or whatever. It's not an exact equivalency, but it's sort of the way to think of it. After that, it sort of understands a lot of the beginning, a lot of the end. In the middle, it is sort of less concrete on. Right. And there's weird things where at 50,000 tokens, it gets a good amount of the middle and at 70, whatever, but it just tends to retain and seem to understand more completely PDFs and things like that than, you know, GPT does at this point. And so that's the primary reason, you know, for it at this point now and then it just seems like, you know, when I have thoughtful conversations with Quad, like, I'll give you an example. I was talking about. I was thinking about, like, the limit on intelligence. This is not to get super esoteric and philosophical, but people talk about runaway AI and. Right. Like sort of super intelligence and all these things. And I was kind of thinking about how, like, if we're talking about, like, you know, AI continue to get smarter and exponentially improving and kind of, you know, like an ongoing exponential reinforcement loop of it becoming smarter and smarter and smarter, you know, we as humans are gonna. We can't even comprehend it. So all these numerous, you know, books from Oxford philosophers are kind of valueless in a way. Right? So it's like if. If I tell anyone watching or listening to the podcast, like, imagine, you know, what it would be like to be two times more intelligent. Then you can kind of do that, right? Yeah, but if I'm like, now imagine if you're a million times more intelligent. Like, it's. It's meaningless. There's. You can't even do it right. A trillion times more intelligent. So anyway, I shared. I wrote some thoughts. I shared it with, you know, GPT4, like, whatever the most advanced current model of it is. And it's kind of like, well, I can't. I don't have any thoughts. Like, I'm in, you know, whatever, like, all that kind of stuff. And I shared with Claude, and Claude's like, you know, wow, that's really interesting. You know, and it kind of gave me, like, a lot of nuanced, you know, feedback. And we had like, a conversation about it. And so Claude just seems a little bit more dialed in for, you know, and again. But. But I'm not like, it could change. Right? That's the. The thing. It. It sucks. But I'm a. Sorry, Claude. Sorry. Anthropic but I'm a mercenary, so if something comes along that's better, I would just switch to that. Right. You have to be very agnostic right now with these tools.

That was my experience as well, that Claude tended to put out the. The creative output was better from Claude. Yeah. So now it's interesting. I want to go back to the. The. The impact of nuclear testing on the environment out without the agents and without these models, what would that.

Would you.

First off, would you even had the luxury of time to be able to explore compressing that learning curve? And second question, how much time do you think that saved you for you to be able to say, okay, I've gathered enough data to have an opinion now?

Yeah. I mean, in general, like, it. Would it have been possible? Like, you know. Yes, but it would have taken me a lot of time. So how much? Like, I mean, I'd say on average. And I'll compare it to like when I go to write copy, because we do. There's a lot of research involved in that. So it's a really good concrete example. But yeah, the sneaker one works too. But basically, if I was really trying to understand something, you know, you're probably talking about 10 to 20 hours of research that I can typically do in like an hour now. And by me, it's obviously letting AI run the background and I'm doing other things. And a lot of that's also just running prompts and AI stuff. Right. I'm sort of like the conductor, you know, just sort of doing little levers. So, you know, and if you amplify that over, you know, I mean, I definitely save like hundreds and hundreds of hours a month at least, you know, as far as like, efficiencies go and things, it's pretty insane. And so that point, I don't know if I would have had time. Sorry. To finish that thought. Yeah, yeah. If I before, that would have to.

Be like a hobby.

Yeah. And then it's hard to even have the time because it's like if you're, you know, doing a lot of intensive stuff, you're working on the business, you're are working in the, in the business, you're. You're doing all these things, I mean, and then you're exhausted. That's one of the really important thing I would note is like. Because like, I found that like going back to a creative, like a really, like, like really creative tasks, like take a lot of brain power, right. You're spending a lot of mental energy, like your, your brain, you know, it uses something like 80 of the calories you take in. Right. It's like, that's. Yeah. And so you, you'll hit this wall for like really creative tasks after a few hours. But what I found is like when I'm kind of automate of AI and I'm still being creative because I'm like seeing stuff and I'm like, oh, that's interesting. Put that together. But I can go way longer too. That's the crazy thing. So I could, you know, not that I do this every day, but I could go for like 12 hours or 15 hours because it's not as mentally taxing. It's more like, oh, that's interesting. Put that together. Right. So it actually really helps you conserve mental fuel as well.

You know, now that you mentioned it, I didn't even consider that, but that's how I was able to work for 14 hours the other day and knock out a week's worth of work in one day. Yeah, interesting concept because the brain does use most of the calories and all that good stuff. And this concept, that of any type of intellectual query that you may have where you're able to dive in business owners, if you're not interested in, you know, understanding the bigger mysteries of the universe, apply the same thing to a business model or a concept in business or anything like that. You don't have to go get, you know, buy the thousand dollar course necessarily and go through it over two weeks. You can get the general idea and these models will allow you to do the iterations and to the point where after one session you're like, okay, I understand that now. How do I apply it to my business? Oh, how do I apply it to my business?

Right. Let me give one cool concrete example, as I do with copy, that will probably hopefully be valuable. So like I mentioned, we have a couple like SaaS things. One's like a forward public facing, like it's sort of like a freemium model and then ones like that sort of, you know, internal like sales, letter writer thing and we're working on a few others, but you know, so we hired like a cto and he's like a really accomplished guy. He's CTO of two different companies that got acquired for a lot of money, paying a lot of money. And you know, my COO and I were like, kind of felt like there's this sort of gap on like project management where sort of like we're not getting like reporting on these sort of, you know, specific things with like the developers. Well, like, you Know, why are we not getting these clear reports and whatever it is? So we're kind of sitting there talking. I'm like, dude, let's just go look at like. And went to quad. But I'm like, you know, hey, what is like the org? Basically, we're like a small SaaS startup. I put, we didn't take investment because we're self funding, but I'm like, you know, we just raised like a, you know, a seed round for like $500,000 or whatever it is. Like, you know, what would our org chart look like? Right? And it kind of like gave us like this org chart.

Nice.

Like, all right, cool. Right. And it sort of like had these different, you know, tech things. We said, We're a SaaS company and basically then we're like, all right, well, what is like a CTO for a company of this size? Or maybe, you know, this can grow a bit larger. Like, what would the roles and responsibilities be? It's like. And it gave us like the list like, all right. And then we noticed that there was like this other, like, it was like a lead development lead, I think, or something like that. And we're like, well, what's the difference between the CTO and the development lead? And basically we found was that if you looked at the checklist for what a CTO should be doing, again, this is based on Claude, but Claude's data from a lot of places, trillions. We're like, oh, crap, this guy's actually nailing it as a cto. All the things that it's saying he should be responsible for, he's literally doing every single one of these things. The problem is we want him to be doing this role over here that we didn't even know existed until we asked. Now, seeing what this. We just need to hire that person to have them interface with the cto and then our problem solved. And it literally was like this thing that was causing all this stress and did we make the wrong hire with this guy? And all these sorts of things to. Suddenly it was so clear in an instant. And it was just such a valuable. So going back to more practical use cases, even things like org charts, role clarity, things like that are just. It's really helpful for that.

Yeah. So as we're wrapping up here, what would be for those who are listening to this and maybe they've, they've tried it, because I know a lot of people that said, oh yeah, I tried it, it wasn't very good. Right. They didn't have that O moment. What's Any advice for a business owner or an employee or coach, consultant, who's out there that hasn't had that oh moment, what would you advise them to do to. To take advantage of this?

Yeah, I think that, you know, there's a few things, perhaps. I mean, you know, you have to think about it. There's. Okay, there's. There's an intern principle, which, you know, people may have heard of, but I feel like we. We like originally Mario, who, you know, like, I think he's like the. He was the first person I heard say it. And it's become a little bit more popular, but which is essentially treating it like an intern. Right. So if you have an intern and you're like, hey, go bring me a report on xyz. And the intern brings you the report, most likely you're going to be like, hey, all right, thanks. Good first attempt. Here's what I like. Here's where you didn't talk about this at all. I don't understand why can you go back and now address these things? And if you do it that way, I think the fundamental mistake that people are really new to large language models make is they kind of treat it like Google, where they ask a question, expect an answer, and if the answer is not that good, they're like, oh, I guess it's not good. But if you have more of that nuanced, conversational. Right, it's conversational AI. And you're like, hey, thank you. But here's what I was looking for. How do you, you know, can you tweak this? Can you go more in depth about this? Can you structure your answer this way? Right. So getting a little more nuanced will generally get you much better results. So that's. That's number one. And it really is like, the quality of the outputs is dependent on the quality of the inputs. Right. So, you know, generally, if you give it more specificity. This is again, where I like quad. Because you can give it a document like, hey, summarize this. Like, what are the key points? Right? And then use those in your next prompt to give more context.

Nice.

And you'll get a lot better, you know, outputs. And then, you know, there are kind of tools. I mean, I don't. I'm not. You can like, we have like, free ad copy dot com. I didn't. I'm not trying to plug that. Right. But please. The point is. That's okay. It's. It's, you know, it's just like a. Like I said, it's like a freemium SaaS where the whole idea was like, copy free ad copy dot com. Yeah. And we're adding, you know, a lot of stuff too. We've got a really cool market research thing that my partner Mario created called Build a Buyer, where it's like sort of psychographic research.

Nice.

You sign up, you get like 10 credits. You get 10 new credits every month for free. And then new refer people, you can get more credits and stuff like that. But, you know, I bring that up as an example for like, if you want to see like, for kind of a promotional ad, email, whatever, or the. The psychographic research. Like, you know, people are generally really impressed and it's. And it's free. And the goal is not with that to like, we added some paid options for people want more credits because people were asking and we have a bunch of other really cool plans for it. But part of the goal really was just like, give people exposure to what it looks like because we basically have our prompts in the middle that you can't see these templates and stuff. And so you fill out some inputs, it gives you really good ads. But even if it's not free ad copy, find some really valuable, test out some different tools and things like that that accomplish what you're trying to accomplish. And you know, like, now you're gonna have like the whole. The GPT store. Right, the chat GPT store, whatever. Which I'm pretty ambivalent about that myself. But there might be some cool. I think there'll be some cool ones on there where you might get really cool outputs with it. I just don't know that, you know, 2%. I would put my stuff on there. But yeah, there's still. Will be some cool ones. I think most of it will be garbage.

So I don't want you to hesitate to talk about what you're doing because you're doing cool stuff. So do you mind kind of sharing with people some of the projects that you're working on as far as the AI powered content and copy generation? And by the way, full disclosure, I'm actually a client.

Yeah.

Came in off a cold funnel. I mean, I know who you are, but I came in off one of your funnels. I started with the. Maybe the 27 prompt.

Yeah, the copy codes. Yeah.

Yep, yep. So I'd love for you to share maybe some of those opportunities for other people to plug in.

Yeah, for sure, man. Well, I appreciate that. I mean, yeah, I think, you know, free ad copy is a really cool one because again, it's a pretty Easy sign up. It's like it kind of works like magic. You know you put your inputs in, you'll, you'll get ads and emails and whatever. You know we have like some other kind of low ticket things like the 27 copy codes is $27 and it's basically 27 very in depth like prompts where people can you know, plug and play and you know, get for marketing I guess. And then you know we have, we're doing a lot but one of the other cool things we're working on that we'll have pretty soon is this like a AICMO kind of program where going back to automations the idea there is like when it comes to like kind of you know, marketing and whether that's like copy, email list monetization, you know, advertisement, like running ads, media buying, whatever it is like love it. Typically there's some bottleneck there but also typically an automation can solve it. And so the idea is when people come in we help them sort of figure out where those bottlenecks are and then we actually give them like, like the specific playbook with like the specific automations that they can then kind of install into their business that just rapidly solve whatever that bottleneck is so that they can grow and scale more.

So if you need beta testers I'm very interested in that because I love what you're doing and the way that you guys are thinking about applying AI into your business processes. So no, no pressure of course but if you wanted like legit feedback because we're doing some software development over here and using primarily no code and I've never created software before and it's a testament to the ability of an individual like me who's not a software expert to be able to communicate with a one man team that can now go hey boss, what do you think about this? Within a few hours, right, MVP is built so pretty incredible stuff. Will that be a service that like an individual is augmented by AI or will it be I give it my inputs and it spits out like proposed automations and stuff to start.

It'll be kind of get done with you. So it'll be like a 10 week or 12 week type of program with like yeah kind of the coaching training like the step by steps like you know, the help of implementation.

Very cool.

You know probably priced around like 3 or $4,000. We're add more and potentially the price might go up. Right. And then you know we're working on like, you know to be honest I don't mind saying I mean one of our goals on the SaaS side is like essentially really to automate everything though to that, to that point, right? So it's like we have free ad copy where it writes like great ads, right? Well then the next thing is like, all right, cool, turn into a video ad, push the button and it's like, right, AI generated great video ad background, whatever it is like you upload your product and the product is, you know, you know, in there. Like we have a spokesperson. You can clone them like you would have hey Jen. Or something like that. We can talk about that. Like now they're in there right now. Like all right, cool. Like you know, deploy to my, my ads manager on Facebook and you optimize ad spend. Boop. Like, all right, cool. Oh, you know, we need a landing page. Like we need copy, right? Like I really want to create like the full kind of middle. It's not middle layer technically is a different. But call it a middle layer, right, where it's like I'm not going to be able to create the new Shopify, but creating sort of like the ecosystem of basically where you can automate every part of marketing and launching something that then integrates with a Shopify or funnels or whatever platform people use or custom through APIs. That's what we're kind of building towards. I love it.

And for those of you who aren't AI enabled at this point, if you understood what he just said, a one person team now has basically the heavy lifting of the CMO role handled. And if you can replace, I don't know whether HR operations, whatever the other functions are of your business with a tool like what Stefan is talking about related to marketing. I want you to start to see. Oh my gosh, I really like, I don't need other. I'm not. That's a stretch. But I don't need other people on my team necessarily or as many people on my team, which is pretty impressive. So we're going to have all your social handles and everything popping up on the screen and down in the, the show notes. But is there a particular channel where you feel like you're expressing yourself the best, where people should follow you to really learn more about you?

I mean, to be honest with you, I think the two best are like, you know, my email list and, and you know, Twitter, AKA X. I mean, I'll be honest with you on X I'm. I'm all over the place like this sort of my, my cathartic like, yeah, me too. I can go be real so, you know, you'll see me kind of just talk about random stuff there, like, if you, like, you know, I don't know. But that's, that's my favorite platform right now, to be honest with you. I just think it's like there's the flow of information is so fast paced, you know. So, you know, I think probably, yeah, Twitter experts. I think it's just that. Stephan George, I will pop it in there.

To get on the email list, just go to freed copy.com and subscribe.

Yeah, if you sign up for free ad copy, like, you'll be on, like, we have, you know, an email, we set out emails, people who sign up for it. I have like a personal list that if you go to stefanpaul, georgi.com forward/subscribe, you'll get on that one. That one's where I kind of share more of like, basically I write the ones that go to my email list. I don't write the ones that get sent a copy. It's a difference. Right.

I'm heading there. After this recording finishes, I'm going to go subscribe. You'll have a new subscriber, plus all the people who will be subscribing from this. So, Stefan, anything you want to say in closing? I think this was fantastic and I really appreciate your time and being as open to the questions and answering those.

Yeah, no, it's my, it's my pleasure. And yeah, I love talking about this stuff. And, you know, the only thing I would say is for people who aren't embracing the stuff, just, you have to. Right. It's like every. I know it sounds like, okay, what does that mean? But it's like, I don't have to say about saying cliche. Right. Every, every technological revolution, there's people who are resistant, who kind of give you all the reasons why it won't succeed, who, you know, worry about job loss and things of that nature. And every single time it ends up being like, for the better and, you know, being a net promoter, like a net positive and lean to growth and, you know, it's just no different here. And the fear is about, okay, well, artificial general intelligence, people will be out of work. It's like, only if people don't have, you know, big exciting things that they want to accomplish. If they want to have big and exciting things they want to accomplish, then, you know, there shouldn't be job loss, there should be more. Right, right. We're hiring more people, not less now, to be fair. Right. We're more efficient as well. So if copywriting, like, our margins on that agency went from like 30 to like 80%. Right. Which is phenomenal. But that frees it up to where, like now, like, we can pursue all these other things. We can hire other people for these projects. We can do. Love it, this stuff. Right. And so, you know, ultimately, like, it's, it's just for what it's worth, if you thought I said some valuable things and you should really take me on my word that, like, this is not like a fad. There's not like if it froze now, if we didn't improve, if GPT5 stinks, it would still be like, absolutely essential that you use what we have today, like, constantly. Right. And it would be wild for you not to. And you'll have. And there's a window of time where a lot of people are going to be resistant and that will go away. But this is to me, like the most, I don't know, opportunity to go time, you know, of any of our lifetimes, I think, as far as what we can do entrepreneurially. So it's really exciting.

That's great advice. And you guys are on fire over there. I had no idea how busy you were. So, brother, virtual high five.

Boom.

Blessings to you and the family. And again, thank you so much for taking time out of your extremely busy day to share with us. And I'll be in touch.

Yeah, absolutely. Thank you, Chris.

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. special thanks to our producer, Evan Desaunier from making this episode possible, follow us on Twitter at the handle usingaiatwork and visit www.usingaiatwork.com for free resources to help you harness AI in your.