WEBVTT 00:00:00.000 --> 00:00:14.075 If you're a business owner still using AI like a search engine, you're already losing. I went from zero to 2,000,000 followers solo. I built a highly profitable SaaS app with thousands of paying users solo. No team, no agency, 00:00:14.155 --> 00:00:26.830 just me, plus an AI stack that does the work of 10 full time employees. You do not need a tech degree, and you do not need anyone's permission to learn this stuff and make money for yourself. Alex Hormozi 00:00:26.830 --> 00:00:30.670 dropped his recent video, how to win with AI in 2026. 00:00:30.750 --> 00:00:39.685 And today, I'm gonna walk through every single key point he makes with real examples from my own work plus insights on why these approaches work. 2026 00:00:39.685 --> 00:00:53.205 is the year that AI agents got real. They can now connect to your email, your calendar, your CRM, your support system, your social media accounts, your customers. The window is right now to learn this stuff before competitors 00:00:52.920 --> 00:00:59.800 catch up. Now most people teach AI like a toy. I teach it like a stack, context, constraints, and tools. 00:01:00.040 --> 00:01:04.680 Understanding that is the difference between ChatGPT yapping uselessly 00:01:04.760 --> 00:01:22.830 versus a coworker that's actually shipping work while you sleep. So here are the seven key takeaways from Hormozi's video, how to win with AI. But first, hit like, hit subscribe, and hit the notification bell so you don't miss my next training. So the number one point that Hermozy starts with is that the time 00:01:22.990 --> 00:01:42.625 is now. Now I know if you've been in AI for the past couple years, you've heard this. Like, time is now. You gotta really learn this stuff. You gotta really learn this stuff. You may have heard AI will never be worse than it is right now, which is true. Okay? How to think about this? Imagine there's a really, really big hill to climb. Okay? And then you have two stick figures. 00:01:42.785 --> 00:02:18.280 One is running up this hill and the other one is staying at the bottom thinking, I am going to learn this stuff later. Well, the person who runs up this hill is going to be literally drowning in millions of dollars in revenue while this person who keeps waiting and keeps saying I'm going to figure this out later is not. They're just falling behind. They don't even realize it honestly. Like, most people I talk to just don't even realize all of the things that AI agents can do and help automate and help you be more productive. So in my particular example, right, so I've gone from zero to now 2,000,000 00:02:18.280 --> 00:02:20.200 followers on social media solo. 00:02:20.280 --> 00:02:34.195 I'm now the number one AI educator for entrepreneurs in the world. And then on top of that, while I was doing that honestly as a part time hobby, I was also building a SaaS business, which is what takes up most of my time that now has thousands of paying customers. 00:02:34.435 --> 00:02:39.955 Okay. And this honestly takes up most of my time. It's a ton of work. I don't ever wanna say that it's easy. 00:02:40.195 --> 00:02:55.080 But because of AI, I've been able to build leverage for myself in ways where I would have traditionally had to hire a really big team just to even get started or scale to this level. So the first thing I want you to do, especially if you're new to AI, is open up ChatGPT 00:02:55.080 --> 00:03:04.125 or Claude and type this prompt. Like, even if you're really well versed in AI, the prompt is list three ways AI puts my income at risk 00:03:04.205 --> 00:04:00.465 and three things I should learn right now to stay ahead. And you should do this even if you're a beginner in AI, even if you are pretty inter if you're intermediate or advanced in AI, just see what it says. Just list three ways that AI puts your current income source at risk and the three things you should learn right now to stay ahead. Okay? Go ahead and try that. So Hormozi's second point is there has been there's never been a better time in the history of all startups in the world to build an AI first business. Now I typically use the term AI native. And what we mean by that is, like, you think about how to structure a team, how to structure a role and build out workflows, thinking about AI first. Right? So it's like you kinda like map things out. You how do you architect to the business and where are you going to put AI? You're thoughtful about that instead of kind of everybody else where you just kind of do things manually first and then you slot in, hopefully, AI afterwards. 00:04:00.465 --> 00:04:19.730 Um, so an AI native business is really, like, structured so that AI can play a productive role in every single department. And there's never been a better time to start this company. So just think about it this way, like a really big company with thousands of employees. These are supposed to be people. And then there's you, like tiny little startup. It's it's well known in the startup literature. 00:04:20.025 --> 00:04:39.990 Like, the biggest advantage a tiny company has is speed. K? Like, it's not resources, obviously. It's not people. It's not brand. It's you'd really don't have anything except speed. If you have ever worked at a big company by the way, my first startup was acquired by a company with over 5,000 employees, over 1,000,000,000 in revenue. K. So big company. I was a director there. 00:04:40.550 --> 00:04:42.230 Like, it moves slowly. 00:04:42.550 --> 00:05:06.070 It's cool when something happens because it can, like, hit the ground running. Like, it can go from zero to like, new product can go from zero to 10,000,000 very, very quickly. Right? But to even get it out the door this is a door. To even do this, to get it out the door, there's a ton of work involved. And it's a lot of bureaucracy. It's a lot of red tape. It's also just collaboration across many different teams, many different stakeholders, 00:05:06.070 --> 00:05:07.430 communicating updates, 00:05:07.430 --> 00:05:48.055 uh, deciding on trade offs, negotiating trade offs. Then there's politics on top. Like, there's just a lot of stuff going on when you have a big company, and so they naturally move a lot slowly. So if you combine your little company plus AI, then you get, like, lightning speed. Um, and that's really what it boils down to when you hear everybody talk about, like, leverage and AI can help you be more productive. It's really about moving faster so that you get the data you need to make better decisions, and then you go make those decisions faster with the help of AI. Okay. So an example of this that's really gone that's recently gone viral is Anthropix growth marketing departments. It was one person. 00:05:48.535 --> 00:06:17.835 And it's kinda cool. He, like, released his playbook in different ways he uses AI agents. So for example, for ads, like, he'll have an AI agent analyze the performance of his ads, look at other ads as well, come up with new copy, come up with new creative, and then deploy those ads and then wait again. Right? Keep AB testing ads in a continuous cycle. And this is what it's about. Like, whenever I think of AI or AI agents, I always think about, like, this feedback loop with your AI in the center. It's really about, like, can we create 10 different variations, 00:06:17.835 --> 00:06:53.770 test them very quickly, get the data we need, double down on the winners, test new creative, etcetera. It's just how quickly can you actually do this loop. And the old way of doing things, like every single step I just mentioned would be, like, manual and take time. If you're at a big company, approvals as well. Right? But in a tiny company with AI, right, you're augmented by AI, you can do each of the steps I listed very, very quickly again and again, getting the data you need, doubling down on the winners, trying new creatives, etcetera. And that is what leads you to better outcomes faster. Um, but let me give other examples. So AI native startups, we're seeing handfuls 00:06:53.770 --> 00:06:56.330 of them, okay, absolutely shatter 00:06:56.330 --> 00:06:59.450 the timeline to go from zero to 100,000,000 00:06:59.450 --> 00:07:08.365 in annual recurring revenue. So AIR just stands for annual recurring revenue. It means like people will pay you each year. Um, so for example, companies like Cursor, 00:07:08.365 --> 00:07:15.325 Lovable, Higgs Field. Now all of these companies have raised traditional venture capital. K? So they're not bootstrapped, 00:07:15.325 --> 00:07:16.365 uh, $100,000,000 00:07:16.365 --> 00:07:32.890 AR companies. But even so, like, in the traditional Silicon Valley playbook, you would still raise capital, but it it would still take you more than five years to hit a 100,000,000 in annual recurring revenue. I mean, this is an insane number. Uh, most of you here, like, we just want 1,000,000 AR. We don't need, a 100,000,000. 00:07:33.050 --> 00:07:38.725 These startups, and there's a handful more examples that I'm just not listing here, but these startups have compressed 00:07:38.725 --> 00:07:52.010 the timeline. So that's the other word I want to you to always think about when you think about AI and agents. I literally just think this is the image I think of. It's all about like speed, doing things quickly, and compressing timelines and compressing costs. 00:07:52.250 --> 00:07:58.010 This is what makes AI so exciting to a one person business because traditionally, 00:07:58.010 --> 00:08:17.040 to scale to even 1,000,000 in ARR, you would typically have to hire a team. Um, now you can do it with the help of AI in core functions, which we'll get to in a bit. Okay? But, yeah, these startups have compressed the timeline from, five plus years to, in some cases, like, under eighteen months to get to a 100,000,000 00:08:17.040 --> 00:08:24.720 in annual recurring revenue. Like, I can't even fathom that. Like, it's it's very hard to fathom. Um, and you may have heard of the other viral example. 00:08:24.880 --> 00:08:26.640 Um, I think it's like MedVee, 00:08:26.720 --> 00:08:38.935 right, that was selling basically Ozempic. It's like a marketplace for you to hit up Ozempic providers. They're obviously in a lot of hot water for unethical business practices. Okay. Number three, they talks about, 00:08:39.975 --> 00:08:41.255 it's all about, like, skills 00:08:41.655 --> 00:08:51.840 and leverage. So I'm gonna talk about this in detail. I think it's really important. Um, so Hormozi's point is that, like, every skill that you give AI gives you disproportionate 00:08:51.840 --> 00:09:07.775 leverage over your competitors who just haven't figured it yet. And I view it as like kind of like a staircase of learning. Like, you you kinda just start with prompts and stuff, I guess. This would be the first one. And that's good. Like, a lot of my short form content are honestly just like really basic prompts. 00:09:07.935 --> 00:09:08.335 It's cool. 00:09:09.020 --> 00:10:01.355 I like to create prompts from different perspectives so that people kind of understand, oh, like AI will give you a totally different answer if you change a few words in the prompt. Okay. The next is like being more structured and methodical about the context that you're feeding into AI. Right? So instead of just a single prompt, you might create a project that contains a bunch of your context. And then the third level after this, would say is tool slash MCP. So for example, maybe you created a project in Claude and it knows how you like to write email replies. Okay. But the next step here would then be connecting it to your actual email system, not only to learn from past emails that you've written, but also to help you draft replies. Um, and then you can even schedule, like, a recurring task for your AI agent to send you an email brief every single morning or to escalate urgent emails to the top of your attention. Right? So this is generally 00:10:01.435 --> 00:10:25.255 the learning path I see most people go through. So we'll call this number 1. Call this number 2, we'll call this number 3, and we'll call this number 4 if you have proactive AI agents kind of, like, scheduled running in the background or with the webhook set up to trigger uncertain events. But I would bet that most people are still here. Like, you have to pause to think about this. Like, we're here on a Friday afternoon watching AI 00:10:25.335 --> 00:10:30.135 on TikTok. Like, first, that's just weird. Like, that's not a thing that cool people do. 00:10:30.695 --> 00:10:49.515 And then on top of that, like, we're in a tiny bubble of the world already or over here. And then within this tiny bubble, most people are still at number one and number two. That's why I say the opportunity for AI education alone is massive. We're a little dot inside a little dot. But I'm very excited when people kind of progressively 00:10:49.515 --> 00:10:55.275 learn and proceed through these steps. Um, I know the amount of education out there is very confusing, 00:10:55.355 --> 00:10:57.915 um, but, like, this is generally 00:10:57.995 --> 00:10:59.355 the sequence 00:10:59.355 --> 00:11:31.115 that you get. Um, and, like, as you go here to proactive stuff, this is where it gets, like, a bit more technical. Um, but Claude is releasing a lot of, like, utilities where it doesn't have to be so technical. Like, now in Cowork, for example, you can schedule recurring tasks, which is really, really nice. So in the video, Hormozi gives this example that he has now launched AI native companies that have achieved multiple millions in revenue per employee. Now for those of you who've never run a business before, like maybe you don't have context for how groundbreaking 00:11:31.275 --> 00:11:40.210 this is, so I can give you a real example. In my space, there's a company I love. They're a competitor, so don't sign up for them. But I still love them because 00:11:40.610 --> 00:12:29.475 they publicly release all of their metrics. So the company is called Buffer. Okay? And if you Google or ask chat Buffer open metrics, you will actually see all of the revenue and metrics over time, the number of employees over time, etcetera. And for the longest time, their revenue per employee was around like 200, let's say $2.50 k. Recently, it's reached above 300 k per employee. Okay. But this is pretty typical actually for a SaaS startup for the longest time. In fact, I'd say this is on the higher end. When you look at AI native companies and Hormozi has proven this because he's launched his own AI native companies. And they're according to him, they're currently doing millions in revenue per employee because they were structured to be AI native from day one. So the change from this being the gold standard to this being, like, expected, 00:12:29.635 --> 00:12:30.355 quite frankly, 00:12:31.050 --> 00:12:40.490 is pretty massive, and it occurred in, like, such a short period of time. Like, I haven't been an entrepreneur that long. I graduated college in 2013, 00:12:40.490 --> 00:12:44.890 had no idea what I was doing. Don't know why people gave me money, to be completely honest with you. 00:12:46.015 --> 00:13:20.235 But at that point in time and for years after that point in time, again, this was the gold standard. Like, even if you're venture backed, even if you're bootstrapped, whatever, like, this is great. Like, you're you're playing great. Um, but now with an AI native company, this is just, like, kinda normal. Like, why don't you have millions in revenue per employee? Um, but this this just happened overnight. Like, this was not a thing five years ago. Like, not even five years ago. Okay? So if you're watching this and you have business background, you're doing the math in your head and honestly, you're probably not at $2.50 k per employees. 00:13:20.795 --> 00:13:22.395 Most businesses aren't. 00:13:22.875 --> 00:13:43.295 But when you build an AI native company, you really do have the opportunity to massively scale this number. And by the way, that's why I'm personally so excited about teaching, like, solo preserve tiny teams. At the end of the day, like, if I were to teach big companies how to use AI, like, that's great. I'm sure that I could make them billions of dollars. But fundamentally, 00:13:43.295 --> 00:13:58.390 I know that those gains don't trickle down to employees. Like, I know I've worked at a big company, guys. Like, you are just an item in a spreadsheet. The gains like, if a if I were to teach a company how to make a billion more dollars, like, that billion dollars would not even go to you. 00:13:58.790 --> 00:15:14.780 Um, and so that's why I'm passionate about teaching, like, solopreneurs so that you can just build your own income stream and not be dependent on a company that honestly could lay you off, like, whenever they feel like it, whatever they want. And they don't care. Um, companies are not built that way. They are not built to care. You can truly build, like, meaningful amounts of revenue. Now I I'm I've given big numbers, like millions and millions. But, like, honestly, how many people's lives would be changed if you had an extra 20 k per month? Um, the beautiful part is, like, once you hit 20 k per month, you realize, oh, shit. I could hit 100 k per month. Like, I I have never seen, at least in SaaS, I have never seen a SaaS business that hit 10 k per month that couldn't be scaled to 100 k per month. And that's kind of the beautiful part. That's why I focus so much on, like, one person just get to a meaningful milestone for you and your family, and then you will realize, like, oh, wait. I can still use AI to scale this up a little bit further. And maybe I wanna hire my best friend. Right? Because it's fun. So an example of how I personally use this staircase. Right? So let's think about my customer support agents called Blue AI. Basically, whenever a new message comes in, Blue AI analyzes the support message. It reads the help docs. So it has access to tools. 00:15:15.020 --> 00:15:26.735 So help docs, it has access to it has access to logs. And then it tries to answer the person's question with some text. So here's some text. This is supposed to be a chat bubble. Okay. 00:15:27.055 --> 00:15:33.295 Um, but this is all happening, like, automatically. Like, it's happening right now. So if you actually go to potato.com, 00:15:33.295 --> 00:15:51.740 if you have an account, you could ask my support bot some questions, and it's gonna do its best. Now it's not perfect, but it has access to a lot of context. Right? So it has access to all that context through tools. And it can perform actions on your account, like canceling your subscription, restarting your subscription. It can even give you discounts. 00:15:51.740 --> 00:16:13.090 Um, like, can do all kinds of things because it has access to tools that allow it to change or inspect the status of your accounts. And it has access to my latest help documentation. So it can answer, like, product questions, what's the current promo, etcetera. And it does a pretty good job. And this is all proactive happening twenty four seven. Right? It's not 100%, 00:16:13.090 --> 00:16:15.170 but it's about 70% 00:16:15.250 --> 00:16:18.930 of customer tickets I get. Um, the remaining 30%, 00:16:18.930 --> 00:16:21.250 so I ask it to escalate. 00:16:21.330 --> 00:16:24.050 Right? If somebody sounds really, really angry, 00:16:24.210 --> 00:16:24.850 I 00:16:25.090 --> 00:16:27.205 want to make sure, like, I am seeing that. 00:16:27.525 --> 00:16:41.460 Um, if if, uh, the help documentation was not clear, for example, I pushed a new feature or I changed the name of a button and the help docs are inconsistent now. Okay. So it knows to, like, flag 00:16:41.620 --> 00:16:53.460 areas of the help docs that I should revisit and clean up to ensure they're consistent. And also just questions that don't have, uh, very clear answers in the help doc. I basically set a confidence threshold. 00:16:54.135 --> 00:16:56.215 I think it's like 96%. 00:16:56.615 --> 00:17:01.655 Okay. So it, uh, I have another agent that basically cleans up open tickets. 00:17:01.735 --> 00:17:03.575 And if it's 96% 00:17:03.575 --> 00:17:08.990 confident in the answer, it will automatically close the ticket. Okay. And then there's an adversarial 00:17:08.990 --> 00:17:09.710 agent 00:17:10.350 --> 00:17:11.150 that 00:17:11.150 --> 00:17:23.035 double checks its work. So this first agent will, like, close a bunch of support tickets. This adversarial one will be like, hey. Hey. You shouldn't have closed Bob over here. I think he could be angry. 00:17:23.275 --> 00:17:23.835 Okay. 00:17:24.315 --> 00:17:46.855 Um, so it's like, uh, three different agents working together. So this is kind of like the primary one that's running twenty four seven, answering support tickets. This is one that helps clean up any open tickets that can be closed. And then an adversarial one that's like, I don't know about closing this ticket. I think you might, you know, piss off this customer. I think we should leave it open and wait until Sabrina answers it. But overall, 00:17:47.015 --> 00:17:52.375 the this system, um, automatically handles about 70% of customer support tickets. 00:17:52.615 --> 00:17:59.540 Um, because honestly, a lot of questions people ask are in the documentation. It's simple things like, oh, how 00:17:59.540 --> 00:18:15.955 do I, uh, get cohort to use the pre pre signed how do I get cohort to upload my photo to like a question like that. Or like, how do how do I connect my accounts? I mean, you would be surprised how many times a day I get that question. How do I connect my social media accounts? 00:18:16.195 --> 00:18:22.035 Um, okay. So next, the task I want you to do to try something like this is to open Claude. 00:18:22.115 --> 00:18:22.435 Okay? 00:18:23.200 --> 00:18:24.800 Open up your connectors. 00:18:25.280 --> 00:18:29.280 I think it's customize on the left sidebar and then connectors, 00:18:29.600 --> 00:18:31.440 and then connect Gmail. 00:18:31.520 --> 00:18:32.240 Okay? 00:18:32.720 --> 00:18:35.520 Now once you've done that, open a Claude conversation 00:18:35.995 --> 00:18:47.355 and ask summarize my unread emails and flag what needs a reply today. Okay. So go ahead and do that. And in the process, you will have actually made it all the way to step three. 00:18:47.595 --> 00:18:48.715 Now proactive, 00:18:48.715 --> 00:18:53.380 if you wanna take it to step four. Right? So this this is already steps one, 00:18:53.540 --> 00:19:00.420 two, three. Step four in Claude CoWork, um, you can click schedule on the left sidebar 00:19:00.820 --> 00:19:02.580 and tell it to 00:19:02.980 --> 00:19:10.375 send me a daily email every single morning and summarize my unread emails, something like that. But this is proactive. 00:19:10.375 --> 00:19:12.215 Right? So that's like step four. 00:19:12.455 --> 00:19:22.420 Um, so it's it's not actually not crazy to go through steps one, two, three, four very quickly in a single example. It's just most people, like, don't realize it or they get confused 00:19:22.420 --> 00:19:25.220 or they get intimidated by the complexity 00:19:25.300 --> 00:19:28.660 of information. But I promise, like, you can pretty much bucket 00:19:28.820 --> 00:19:30.580 all AI education 00:19:30.820 --> 00:19:36.555 roughly into these four, like, kind of basic prompts. Like, oh, that's cute. Like, 00:19:36.715 --> 00:19:48.235 honestly, with my education with prompting, I just want people to realize, like, you can drastically change the outputs you get just by changing the prompts. Like, I want people to realize they're the ones actually in control 00:19:48.670 --> 00:20:03.255 and that they need to be careful with, like, what they're prompting because it will give a particular type of response. So how do you actually break things down? Like, what is the right way to think about it? In Hormozi's video, he talks about thinking in terms of workflows, 00:20:03.255 --> 00:20:04.135 not roles. 00:20:04.295 --> 00:20:07.655 So don't think of like a Facebook ads person. 00:20:07.815 --> 00:20:19.940 Break down what that person does into concrete tasks and then literally go to AI and be like, hey, how do I use you? Okay. So like, instead of thinking like this is a whole support person, 00:20:20.260 --> 00:20:25.620 okay, just think about in terms of the discrete tasks they do. So they read a message. 00:20:26.020 --> 00:20:29.940 They look up help docs. So let's say search for information. 00:20:30.495 --> 00:20:32.655 Okay. They draft a reply. 00:20:32.895 --> 00:20:35.215 Okay. Then they send the reply. 00:20:35.695 --> 00:20:37.855 And in some cases, they escalate 00:20:37.855 --> 00:20:38.975 to a human. 00:20:39.295 --> 00:20:48.140 Or in if it was a human, they would escalate to a manager or someone more senior. So instead of thinking, oh, how do I use AI to, like, create a support person? 00:20:48.380 --> 00:20:55.420 That's, like, really overwhelming because a support person actually does many things. So instead, break it down into concrete 00:20:55.420 --> 00:21:10.285 tasks. What does a support person do? They read the message. They search for relevant information. They draft a reply, review the reply, they send the reply, and in some cases, they ex escalate to a human. When you break it down into these steps. Okay? 00:21:10.525 --> 00:21:16.140 Then oh, it was too close there. Then the next thing is to go dump this list 00:21:16.300 --> 00:21:17.020 into 00:21:17.180 --> 00:21:20.060 chat or Claude and literally ask, 00:21:20.380 --> 00:21:25.180 here's all the things I have to do. What steps can AI handle today? 00:21:25.340 --> 00:21:31.805 And what tool do I use for each one? Okay. Like, that's it. So somebody who just asked about nonprofit 00:21:31.805 --> 00:21:32.365 fundraising. 00:21:33.165 --> 00:21:35.085 So I would challenge you to, 00:21:35.485 --> 00:21:50.010 um, create this list. By the way, AI can create this list for you. You can say, hey, chat. Interview me to figure out what are the things I actually do every day. Okay? It's gonna make an it's gonna interview, ask you a bunch of questions. It's gonna make this really nice list. 00:21:50.170 --> 00:22:11.220 And then you're going to AI and ask what are the areas you can help me with today? And what is the tool that I need to use for each one? Okay. So a lot of people overwhelmed with AI, especially business owners, they think about it at it too high an abstraction level. Like, how do I use AI to do sales? And I'm like, I can see how that's overwhelming. 00:22:12.180 --> 00:22:19.060 Um, because, like, uh, it's like sales is a complex task that involves, like, a lot of moving parts and you need to continuously 00:22:19.060 --> 00:22:22.420 give feedback, improve the system. It's never gonna work on the first try. 00:22:23.125 --> 00:22:28.085 So just don't think about it at that abstraction level. Think about it like, what does a salesperson 00:22:28.085 --> 00:22:28.725 do? 00:22:29.045 --> 00:22:34.165 So they find leads. Can AI help me with that? They qualify leads. 00:22:34.245 --> 00:22:39.540 Can AI help me with that? Or maybe I should plug in an existing API like Clay, like it already works. 00:22:39.940 --> 00:22:59.865 Um, number three, they reach out to leads. Okay. How can they reach out? Could they DM on LinkedIn? Is there an AI tool for that? Um, could they send or write a personalized email? Is there an AI tool that could help me for that? Then they book a meeting. They have a funnel. Oh, maybe AI could help us write the funnel and improve the copy so we get more conversions. 00:23:00.105 --> 00:23:11.440 Does the funnel have a video sales letter? Oh, maybe AI could write the script for our video sales letter. You can even use AI to edit your video sales letter, even though I think it's probably overkill. 00:23:11.680 --> 00:23:15.600 Um, so just break it down into really concrete tasks 00:23:15.840 --> 00:23:27.135 and then go ask AI, like, AI, be brutally honest with me. What can you help with today and what tool do I have to use for each of these tasks? Okay. So think in terms of workflows, 00:23:27.215 --> 00:23:28.895 not like entire roles. 00:23:29.135 --> 00:23:30.655 And generally, things become 00:23:30.815 --> 00:23:32.335 far less intimidating. 00:23:33.020 --> 00:23:44.220 Because a lot of that fear and, like, confusion is also because you're just you're, like, trying to use AI to replace what a real person does, but, like, a real person does a lot of things. It's not realistic. 00:23:44.380 --> 00:23:47.900 So, like, just think about it in terms of, like, the discrete concrete 00:23:48.085 --> 00:23:52.805 tasks, uh, that you do on a daily or regular basis for that role. 00:23:53.125 --> 00:23:57.765 Okay. Number five is what I call the TCCA 00:23:58.165 --> 00:24:01.445 stack. Uh, and the the idea here is, like, you wanna train 00:24:01.990 --> 00:24:06.630 AI just like you would train a person. We're gonna call this TCCA. 00:24:06.630 --> 00:24:15.030 It's a fancy acronym. So what does it even stand for? That's a good question. So whenever you're talking to AI, like, is useful to have some kind of framework. 00:24:15.455 --> 00:24:19.935 It doesn't matter which one, honestly. Just like whatever you can remember. So 00:24:20.255 --> 00:24:24.895 t stands for the task. Right? So you're going to explain the thing you want. 00:24:25.135 --> 00:24:29.750 Like, uh, I want you to write me an email. C is for context. 00:24:29.750 --> 00:24:35.590 Well, what is the email about? This is a reply to a customer support complaint. 00:24:35.910 --> 00:24:38.150 The other c is the constraint. 00:24:38.550 --> 00:24:45.055 Okay. This could be like our company is this. So only say information about our product, 00:24:45.135 --> 00:24:46.335 not anything else. 00:24:46.575 --> 00:24:58.820 Con constraints can come in all various forms. It could be like, um, I don't want em dashes in my response. Like, that's a constraint. Or, um, I want the reply to be less than 100 words. That's also a constraint. 00:24:59.060 --> 00:25:05.940 And then the last a is my favorite. It's the one I use all the time because I'm super lazy and I can't remember any frameworks. 00:25:06.180 --> 00:25:08.340 Um, ask me clarifying questions. 00:25:10.425 --> 00:25:14.985 This I love this one because you can just kind of skip a lot of the other ones, 00:25:15.465 --> 00:25:36.520 and AI is going to ask you. It's just it's literally gonna ask you, wait. Can you clarify what the task is? Can you give me some more context? Can you specify if you have any constraints? I guess, I'm gonna go through this exercise with you. Um, but, yeah, use some kind of framework like this so that, uh, you're giving AI useful information, sufficient information to be able to do its job well. 00:25:37.155 --> 00:25:45.155 So, yeah, as a next step, so I would challenge you to just pick, like, one task. So just choose one task that you do on a repeated 00:25:45.155 --> 00:25:48.115 weekly basis and open up AI. 00:25:48.275 --> 00:25:55.770 Okay? And then write it out using this framework. So, like, what is the task? K? Explain it. Add some additional context. 00:25:56.010 --> 00:26:12.735 If you have any constraints, if none come to mind, that's fine too. And then ask me clarifying questions. You just append this to the end of your prompt. So go to AI and do that. Like, here is a task I do on a weekly basis. Blah blah blah. Write it out, paste it into AI. 00:26:12.975 --> 00:26:17.375 And then if you like the output, you can save it as a reusable skill 00:26:17.455 --> 00:26:18.255 or 00:26:18.335 --> 00:26:19.135 GPT. 00:26:19.215 --> 00:26:26.020 Okay? So that way you don't have to repeat yourself every single time with all the same context, all the same constraints, etcetera. 00:26:26.180 --> 00:26:31.460 K. So this is a really simple prompt stack that I just like to recommend to people. 00:26:32.500 --> 00:26:35.380 Because, like, so many people are like, am I prompting AI correctly? 00:26:36.995 --> 00:26:38.515 There's no right or wrong. 00:26:39.395 --> 00:26:46.435 What matters more at the end of the day is to to me, is AI making you more money? All this other stuff around, 00:26:46.675 --> 00:27:02.340 uh, like, this is the right way to do it. This is not the right way. You're gonna get better outputs here. Like, yeah, you can get better outputs, but, like, at the end of the day, are you taking action on it? Um, if not, why? Like, was the plan you came up with too complicated, 00:27:02.340 --> 00:27:03.460 too overwhelming, 00:27:03.460 --> 00:27:05.380 not realistic given your constraints? 00:27:05.865 --> 00:27:11.785 K. So, like, ultimately, the only metric I actually look at is, like, is AI helping me scale 00:27:11.945 --> 00:27:22.650 my business or my brand with the only metric that matters in business that's revenue or revenue per employee. And for your brand, uh, that might be followers or something like that or watch time or something. 00:27:22.970 --> 00:27:33.745 Um, so point number six that, uh, Hormozi makes is that the last valuable thing a human can do is take risk. And like he goes on to say, 00:27:33.985 --> 00:27:46.385 like either double down on the AI native approach and build a business that's doing millions per employee per year or just like focus on a business where the human elements will never be removed. Okay? 00:27:47.240 --> 00:27:49.480 In in my case, like, 00:27:49.800 --> 00:27:52.520 I would phrase it as bet on yourself. 00:27:52.920 --> 00:27:53.560 Okay? 00:27:53.960 --> 00:27:56.360 And I briefly talked about this earlier. 00:27:58.040 --> 00:28:11.155 It's just that, like, I'm so passionate about teaching AI to individuals instead of big companies because I see the opportunity right now for one person or tiny team to build a meaningful amount of income 00:28:11.315 --> 00:28:12.035 without 00:28:12.330 --> 00:28:20.650 having to hire a traditionally big team and the headache that comes with that, by the way, um, without having to raise a ton of venture capital. 00:28:20.810 --> 00:28:24.730 K? So I'm a big believer in just, like, bet on yourself. 00:28:25.485 --> 00:28:28.365 And that also means, like, learn as much as you can, 00:28:28.685 --> 00:28:30.685 build income streams for yourself. 00:28:31.245 --> 00:28:33.965 And honestly, once you do this, like, once, 00:28:34.525 --> 00:28:40.285 you you will have the skill to be able to do to get. Like, you can build additional businesses if you want. 00:28:41.110 --> 00:28:56.775 And so I'm just really bullish on that. He gave the example in his video that his friend spun up a division inside his own company with one mission, which was to put the larger business out of business. Um, I think that's a super cool challenge. 00:28:56.855 --> 00:29:05.255 Um, I would just question, like, what you get out of it. Um, because like I said, at a big company, the gains don't really trickle down to employees. 00:29:05.700 --> 00:29:07.460 That's just the reality. 00:29:07.700 --> 00:29:16.340 Um, but they do provide, you know, stability kind of, security kind of. Um, but yeah. Anyway, so that's like, he talks about that, 00:29:16.580 --> 00:29:27.035 um, in terms of, like, building your own AI business or doubling down on the stuff that's very human and, like, just cannot be replaced by AI for, like, live experiences, 00:29:27.355 --> 00:29:29.435 um, like, uh, entertainment 00:29:29.675 --> 00:29:53.685 that involves, like, human actual humans and stuff like that. I'm gonna spin that one a little bit and just make the point, like, I believe you should bet on yourself. That means learning as much about AI as possible, trying to build AI businesses until you find one that you actually like and will stick to. And once you develop that skill, you'll be, like, much more confident so you can do it again. You can build another business if you want. So Hormozi's point number seven is every single day, 00:29:54.085 --> 00:29:58.085 audit what you're doing. K. Calls it daily 00:29:58.085 --> 00:30:19.880 task audit. So for example, just write down what you do every single day. So step one, yeah, literally write down what you do every day. Um, if you can figure out how to take tasks that typically take you thirty minutes and with the help of AI compress them down to two minutes, then that's huge. Right? That's like a huge, huge win w. 00:30:20.585 --> 00:30:21.305 Okay. 00:30:21.385 --> 00:30:28.025 Um, so in the Hormozi in Hormozi's example, he talks about Facebook ads in particular. So, like, if we actually break down 00:30:28.345 --> 00:30:30.745 what a Facebook ads person 00:30:30.905 --> 00:30:42.020 does. Right? There's like multiple steps involved. And I'm learning Facebook Ads right now from the amazing Mitch Barnum. And so it's it's really interesting for me to hear about his Facebook Ads examples. 00:30:42.660 --> 00:30:50.015 So break it down into break down what what it means to run Facebook ads, and you'll see all of these concrete steps. 00:30:50.255 --> 00:30:58.815 And then what you wanna do is, again, go to AI and figure out, hey. How can AI help me with this one? So for example, when it comes to running creative, 00:30:59.660 --> 00:31:06.140 I recently ran some, like, bottom of funnel Facebook ad campaigns and just used Nano Banana 00:31:06.300 --> 00:31:08.460 to make, like, a poster. 00:31:08.860 --> 00:31:25.495 Uh, so I have, like, three posters on Facebook. They're just Nano Banana made. One of them really does not look like me, but they're doing really well on Facebook and they just made them with nano banana. Um, so that's an example, a real example of using AI for creative to drive real revenue. 00:31:25.895 --> 00:31:26.535 Um, 00:31:26.695 --> 00:31:27.095 and then 00:31:27.990 --> 00:31:39.190 you can also use AI for analysis. Like, I'm still pretty new to Facebook ads. So I ask Claude all the time, like, what does this mean? Like, I don't know what this means because I'm in the process of trying to develop my own intuition 00:31:39.270 --> 00:31:52.955 around these numbers. So I'm always having Claude, like, explain, what does this mean again? Like, is this good? I don't know. Like, just just keep asking it for help with all of these things. And, of course, um, I use AI a lot for helping me with copywriting, 00:31:53.035 --> 00:31:55.515 um, whether it's email copywriting, 00:31:55.515 --> 00:31:56.555 headlines, for example. 00:31:57.620 --> 00:31:58.980 When it comes to, 00:31:59.060 --> 00:32:08.660 um, headlines, I just help use Gemini to help me write some email copy to reactivate churned users for my product. Because I am not very good at, like, 00:32:09.635 --> 00:32:12.595 like, revenue generating email copywriting. 00:32:12.595 --> 00:32:14.035 Like, it feels, 00:32:14.195 --> 00:32:30.130 like, sales y to me, and I don't wanna write that way. At the same time, it does convert very well. So it's something I'm experimenting with, and that's where AI is teaching me. Like, I'll write a draft. Then it writes its version. And I'm like, well, why do you think yours is better? And it explains to me like, well well, 00:32:30.610 --> 00:32:38.935 Sabrina, mine is, like, actually harping on an emotion, and yours just describes your tool. And I'm like, oh, that makes sense. Right? Like, emotions. 00:32:39.335 --> 00:32:40.455 That's a good one. 00:32:40.695 --> 00:33:03.690 Um, but, yeah, just like figure out break down what the role is. Like I said before, discrete tasks. Literally go to AI, ask it how could it can help with each one. And you just do this on a weekly basis. So every day, you just write down the tasks, look for something that currently takes you thirty minutes but can be compressed in two minutes. Like if I had tried to make this poster in Canva, 00:33:03.850 --> 00:33:09.455 it would easily take me thirty minutes. Like probably two hours because I'm that bad at visual design. 00:33:09.775 --> 00:33:25.050 Um, but with Nano Banana, it took me literally two minutes. So now I can, like, create lots of different creative and test out lots of different ideas in a fraction of time it would have normally taken. Right? So it goes back to, like, speed and compressing the cycle. Like, it's always about 00:33:25.290 --> 00:33:29.930 speed and compressing cycles of data and iteration and experimentation. 00:33:29.930 --> 00:33:34.650 So if you wanna take the next step here and, like, try something like this, go to YouTube 00:33:35.785 --> 00:33:40.185 and go to my YouTube and look for my Claude CoWork tutorial. 00:33:40.345 --> 00:33:41.625 I think it's called 00:33:41.785 --> 00:33:45.785 I think it's called five insane use cases or something. 00:33:46.265 --> 00:33:47.705 So look for this on YouTube. 00:33:48.490 --> 00:33:51.770 This one was really fun because it basically walked through 00:33:52.330 --> 00:33:55.130 building a personal AI email assistance 00:33:55.210 --> 00:33:57.930 and then setting up an email brief. 00:33:58.010 --> 00:34:06.995 It also walked through creating social media call contents, scheduling out your content, and also making unlimited videos for free inside Claude CoWork, 00:34:07.075 --> 00:34:25.040 uh, using open source video generation libraries. Okay. So this one is really fun to go through, and this one will very much level you up if you're trying to understand how can I use AI in my real world day to day? Okay. So, yeah, that was it. So we covered how to win with AI in 2026. 00:34:25.040 --> 00:34:35.377 Just going through each of Alex Hormozi's talking points and then adding my perspective. If you like this, make sure you hit like, hit subscribe, and hit the notification bell so you don't miss the next training.