Transcription
I want to go straight into the framework, and I want to give people the playbook. And so, step one, and I—and I’ll read it out, and I want to get your reaction—is identify a painful workflow. Find a problem worth solving. The foundation of any AI startup is identifying an inefficiency that AI can automate or enhance. Instead of chasing broad AI ideas, focus on specific, tedious workflows that people are willing to pay to fix.
Can what did you mean by that? Identify a painful workflow. I have this thesis that when you’re in SAS software and you see an export button, that every time you see an export button, there’s a $100 million a year AI business that could replace that export button. Now, the reason I mean that—I, I’ll explain what I why—and the reason why is when you’re exporting data traditionally, you have some sort of analysis, like the reason you’re exporting it is you’re doing some analysis around, um, the data itself. So, for example, I’m filing my—I’m doing my taxes this week, and I exported data from my bank because I wanted to categorize some of the expenses. Now, the categorization today is being done by human beings. But going back to our earlier conversation, what happens when you can have human intelligence on tap? When you can essentially humanlike intelligence on tap, when you can essentially turn on intelligence, um, and analyze data. So what I mean by find a painful workflow is it painful for me to go and categorize, um, my financial data? Like, yeah, kind of, like to be honest, like I’d much rather be like hanging with my baby or, you know, pushing code or designing or being a bit more creative. And I think, um, what people need to do is in their day-to-day lives when they see things like export buttons and when they are a part of a workflow that they feel tedious, prompt yourself and say, “If this could be done via AI, what would that look like?”
You know what? Can you actually—cuz I want to make it so real for people—can you give some examples of companies either that you’ve built or you’ve seen other people build these AI companies where basically they identified almost like a pretty niche pain point and they’ve managed to build a tool around it or a service that is now generating millions of dollars for them in their business?
Yeah. I mean, how much time do you have? Like there’s I, I, I’ll just tell you one that I was looking at literally right before, uh, this this call. There is a company called Icon. I’ll share my screen. Can I share my screen? I think so. Yeah. So, anyone who’s created Facebook ads or Instagram ads knows that it’s a—it’s tedious. It’s very tedious. Um, it’s you have to find an actor. And by the way, I’m not involved in any way in this company. Um, I just thought I—I just thought it was cool. Um, I just, uh, yeah, you have to like find a creator, you have to manage the creator, you have to edit the videos, you have to write copy. And the best, you know, people who know Facebook ads know that the best people create hundreds, if not thousands of ads, and test the best ones, and they always need to be creating new ads. How how tedious does that sound? It sounds super tedious, right? Yeah. So, um, Oh, here even on their website they say, coming up with 100 unique 30-second scripts is painful. Um, creating three permutations of each of the 100 ads makes the pain even worse. So, their solution: first AI ad maker, think ChatGPT plus CapCut, but for making winning ads with AI in minutes. AdGBT gets your ads to 80–99% complete. Make edits with AdCut for ads and prompt AdGBT until you’re happy. Three-person creative teams make 30 ads per month with Icon. They make 300 and get more winners. That’s a big deal. And by the way, these—they look at these like these are AI. This is AI, dude. Um, it doesn’t look like AI, but these are AI. So the actual creative itself is AI.
Yeah. Yeah. This is See, they have AI-generated creators. That’s what they call it. You know, I’m just thinking about the person at home that’s—that’s listening. And so I like what you mentioned of like anytime you see an export button, there’s a potential pain point that you could monetize. And so the thing that—that I would not push back, but the thing that I would be thinking is like how does someone figure out the difference? How does someone basically figure out the pain point that they should actually act on? And by that I mean it’s not just like a minor inconvenience that no one would pay for. It’s actually a pain point big enough that you can monetize it. Like how do I—how does someone make that decision?
I think a lot of people watching this video work at jobs which they don’t like that much or that they think that they can unlock their potential a little bit more, and they listen to this because it fires them up. That’s my thesis based on like the comment section that I saw on this pod—on this—on this channel and stuff like that. So that means that—so you might look at this this business, uh, Icon, and be like, how would I, you know, it doesn’t seem that—it doesn’t seem that painful. It doesn’t, you know, I don’t really see it. But if you’re a marketing manager working at an e-commerce startup and you’re creating 20 ads a month, that’s—you probably know it’s pretty painful. You probably don’t like doing it. So, if anyone has good ideas for AI startups, it’s people who work at jobs they don’t like and doing painful stuff. So, if you’re watching this and you can take stock of here’s the two or three things that I do that are repetitive that I don’t like doing, chances are there’s a good chance that you can plug an AI startup to go and fix that workflow and make it a lot more efficient with a—they call it a human in the loop. Um, you know, maybe it’s not fully AI powered, but you can still have a human review the ad. You know, Icon is an example of a human in the loop business. It’s not like it’s fully automated where you just say, “Hey, I need 10. I need 300 ads, and it’s going to go and optimize this uh click-through rates and the cost per click and everything.” No, you still have a human who looks at the ads, who posts the ads, stuff like that. Um, so that—that’s what I would say to—to those people.
Yeah. You know, I just—I just think it’s so good. And it’s almost like the—it’s almost step one is like just do a—an audit of your own work life and like those two to three areas, and we all know them, right? Those two to three things that you do every single week that is so painful and so tedious, there, your AI business idea is right there. It’s right there. These—and you know, you think that the biggest AI business ideas are somewhere in Stanford, but they’re right here because these people have intimate knowledge around these workflows better than anyone else, right? So you’re competing—people who are creating AI startups, they’re competing—they’re going to compete with the people who listen to this, and they’re going to try to figure out how do I get some insider—like the people listening to this are the insiders basically. Funny enough, because these people know the workflows.
Yeah. Okay. You know what? Okay. So, step two, hack V1 fast. Launch an imperfect MVP quickly. Rather than spending months refining a product, quickly launch a minimum viable product to validate demand. Can you talk through that?
Yeah. The mi—the mistake that a lot of people make is, uh, they find a tedious workflow, and they’re like, I just—they’re like, I know that this is going to be a big business. I’m going to go and create a—a big brand around this. I’m going to go create a deck around this. I’m going to go and spend the next six to twelve months building this fully featured product because that’s what this audience needs in order—for it to be very valuable. And the goal isn’t, you know, you know, the goal is to create a minimal lovable product. People talk about minimal viable product. It’s—it’s actually to create a minimal lovable product. Um, and, uh, thinking about how can you use tools like Replit and Bolt and stuff like that to just get something out there that you can, uh, just gain some—like have a foundation. Think of it as like a foundation. Um, you don’t need to build, you know, Rome in one day. And—and that’s what I mean by that. Just get something out there, um, and iterate—iterate from there. You know, I actually, Greg, I have, um, I have a confession to make, which is, um, the first two businesses that I ever tried to start, they were t-shirt lines in university. Spectacular failures, like complete flops. And the thing, the biggest thing that I did wrong is I tried to develop this perfect product. At the time, it was like when the kind of athleisure—athleisure trend, uh, was starting, and I was like, I’m gonna build the next Nike. I’m gonna build the next Air Jordan. And we just never ended up releasing a product cuz I would just keep trying to make it perfect before I would show anyone. And it speaks exactly to what you’re saying, which is you don’t need the perfect product. You need the minimum lovable product. I think that—that’s so—that’s just so good.
Well, I’ll tell you—I’ll tell you why a lot of people create—want to create the perfect product because it’s actually similar to why people don’t post on social media. So, the reason why people don’t post consistently on social media and build audiences is because they—a lot of the time they—they feel like they’re getting judged by their friends and family. And it doesn’t feel good to put out a tweet and get zero likes, especially after your 40th tweet. And it isn’t fun to post on Instagram and only—and see no likes and just seeing a lot of shares because you’re like, the reason why people are sharing this is they’re sharing it. My friends might—or my friends quote-unquote are sharing this, and they’re making fun of me, and they’re—they’re like, “Why is Greg posting about AI coding tools? That’s so lame.” Or something like that. Um, products are the same way in the sense that we want to be proud of the products that we ship. We don’t want to put something out in the world that we’re not proud of. And that’s why like a lot of us are Steve Jobs-like in the sense—it sounds like you were kind of like Steve Jobs-esque in the sense that you’re like, I need to—I know what I want, and I know what needs to get out there, and I have a—a high barrier for what makes an incredible product, and that’s why I need to perfect it. But the way to perfect is actually just consistency and iteration. So the way you actually get perfection is to put something out there, get feedback, iterate from there. Um, and I wish more people would just amp up the quantity of things that they put out there. Um, and realize that as you amp up quantity, you will increase quality. Cuz I know that there’s someone listening to this and—and they’re saying, “Oh my god, he—his—his philosophy is just put crap out there and something will stick.” And I’m not saying put crap out there. I’m just saying that what you think is the product is usually, you know, instead of 10 features, it’s usually one feature, and—and then the way to get to 10 features is actually, uh, by iterating your way and adding one feature at a time.
Okay, so if you enjoyed this episode, you’re going to absolutely love what we did with this one right here. Click here.