You’re Not Behind (Yet): How to Learn AI in 16 Minutes

Most people using AI are burning money on expensive inefficiency, which is why it's absurdly easy to outperform them. I've spent 14 years in the digital space, been CEO, investor, board member, along with generating billions in revenue for businesses. And I personally own multiple seven figure companies that are all now powered by AI. And here's what I'm seeing. The gap in AI is getting wider fast. In this video, I'll show you the exact seven step roadmap that I use to to turn AI from frustrating time and energy sapping into your unfair competitor advantage. The best part? You can do all of this in just 30 days, even if you're starting from absolute zero. Let's jump in. Here's what nobody's telling you. The problem isn't AI. It's not the tools, it's not even the prompts. It's that you're speaking the wrong language. Week one starts with learning what I call AI English. And it's nothing like what the gurus are teaching you. Most people talk to AI like it's a human being and it isn't. That mistake costs you hours and delivers garbage results. Why? Because these systems don't understand our language at all. They predict it. And when you understand how prediction works, everything changes. If I said out loud now, and I'm going to twinkle, twinkle little, your brain fired star, you knew, you predicted it. That's not magic, that's pattern recognition. And that's exactly how AI works. But here's what most people miss. Google search has been around for decades. And Google Autocomplete does exactly the same thing LLMs does. It has seen millions of searches, it's learned patterns, and now it predicts what you type in. AI models like ChatGPT and Claude work exactly the same way, but with one critical they don't store answers. They generate everything on the fly using math, not memory. So now that's step one almost complete. You understand why AI can feel so smart, yet so alien from you. Your words are actually tokens. Twinkle is one token, little is another. And they're broken down into numbers, technically known as multi dimensional vectors. The numbers are placed inside a huge space based on math called an embedding space. And inside this massive space, similar words and ideas. They lived close together. So it knows, for example, star, moon, sky, etc. Are closer together but very far away from house or brick or car. So when it's time to generate an answer, AI predicts the most likely next token based on that context. So when it sees twinkle Twinkle little star How I it weighs its options. Love you Possible. See, you likely wonder what you are most probable based off of billions of sentences. The answer gets generated not from memory, but from probability. And here is where everybody gets stuck. When your prompt is vague, AI's guessing from a massive pile of mediocre blocks. But when your prompt is precise, AI samples from the sharpest, most valuable blocks in its entire mathematical brain. That's AI English. And once you speak it, you'll never go back to fighting with ChatGPT or Claude again. So what does a short prompt look like? I. I call it rag. R for role, A for assets and G for goal. The role is where you tell the model what role it's playing, so you don't ask it to review your CV. You say you are a senior CV consultant who's reviewed 5,000 plus executive resumes and specializes in transforming vague responsibilities into quantifiable achievements that get interviews. Give an AI a lens. It then gets to run off to its collection of Lego blocks and pull all of the information from there. When it comes down to the assets, you give it the context and the data. So you paste in your cv. Or you could upload your CV is probably a better way of doing this. Here's the candidate's work history for the last eight years. The target role is a marketing Director. His current CV is this attached and we're trying to target SaaS, businesses and applying to Series B startups. Now AI knows what to optimize and where to focus. And for the goal, what do you want it to accomplish? Rewrite the top five bullet points, maybe from its current cv. Review the whole CV and give me your opinion of what Series B SAS companies would think. Whatever you want the goal to be, just give it the goal and that way you will allow it to be able to do things like make the CV more achievement focused or format it for ATS passing and human impact. That is precision. That is structure. Only when you learn to speak AI's language does it start working for you instead of against you. And now into the next steps. Everyone's chasing 50 different AI tools at once. Everybody is doing it. That's like trying to learn the piano, guitar, drums and saxophone simultaneously while blindfolded with your one arm behind your back. It's insane, right? You wouldn't do that, so why do it with AI? The right approach is to pick one, go deep and master the thinking. There's a study showing drummers pick up guitar faster than complete beginners, even though drumming has nothing to do with melody. Why? Because they train their brain to see patterns and practice with intention. It's the same thing with athletes. That's what you're doing with AI. You're not learning tools, you're learning how to think in systems. So where do we start in the first week? If you want the most mature model, choose ChatGPT. If you're in Google's ecosystem, you probably would be better off with Gemini Google AI Studio. If you want business focused AI, which is the one that I tend to use more often, it's Claude. But you know what? In all honesty, it doesn't matter one iota. The tool is the instrument, the skill is in the music. So in week one, pick one, learn its personality, its quirks, its strengths. And by the end of the week, you should be able to craft a very well structured RAG prompt without thinking about it. It will become second nature, it will be natural. But structure alone isn't quite enough. Even the smartest AI in the world will sound clueless unless you give it context. Every answer AI gives you depends on how it understands the question. We've just covered this. Inside of all of these models, there's nothing but billions of numbers in a massive mathematical space. Context is the map that tells AI where to look. I use this framework called Clip. See conversation, conversation history, notes from previous sessions. You can repaste threads or ask the model to summarize the that's how you build continuity. L is for links and files, files, data, resources, attach your campaign exports, your CRM data, your cv, your competitor, research. These assets ground AI in your reality. Not generic Internet knowledge, but your actual reality and eyes for integrations. The tools that the model can call to do work. The integrations could be search the web or scan the drive or you know, write this code or create a Google Doc. They are extremely powerful these days and P is prompt the instruction itself built using rag that you've already learned. When I analyze Google Ads accounts for example, I don't just say review this, I give it 90 days worth of performance data. Previous optimization notes that I've made clients profit margins, KPIs and then the structured prompt using the RAG framework and the output actionable insights that I can implement same day. Once you're using Rag and Clip together, you've welcome you've joined the top 10% of AI users. But if you want to hit expert level, there's one more thing that you need to understand when you're not getting the right answer. The problem isn't AI, it's your Thinking, you need to shift your mindset. Prompting isn't typing, it's iterating. When output is weak, I assume the fault is mine, because it is. Did I give it the right Persona? Did I give it the right context, the right goal? Sometimes I even ask the model, what did you do? Why did you choose that answer? And it explains all of its logic, really in depth and allows me to be able to reason with it. Right? It shows me it's reasoning chain. And that's when you start learning how it thinks. There's three iteration patterns that I use, chain of four. Think step by step, show your reasoning, then give me the final answer, the verifier. Ask me three questions that will clarify my intent and then combine what you learned and try again. And then it's refinement. Propose two sharper versions of my question. Ask, which I prefer these loops. Teach the model how to understand you. And teach you how to understand the model. It's crazy, innit? How simple it seems. Now, the logic, it makes sense what I'm saying to you. The logic. If you've ever felt like AI was making you dumber instead of smarter and making your job more difficult, hit like or subscribe. But the test is this. Tweak, tune and push until you know why something works. Then when it clicks, you're not talking at AI anymore. You're having a real conversation. But here's what nobody's saying. Everyone's using AI to write their sales copy, their emails, their LinkedIn posts, and you know what it does show? Okay? The truth is AI cannot speak to a human heart. I've seen it. Clients ask ChatGPT to write Google Ads for personal injury lawyers. They output injured. We help fast results. Call now. Ad copy. Okay, it sounds like a robot having a stroke. Zero emotion, zero trust. Compliance nightmare. The real use case is AI isn't here to replace your voice. It's here to sharpen your thinking. But Michael, what should I use it for then? Well, use it to research competitors, analyze data patterns, structure your ideas, generate frameworks, run your whole business from it. But when it's time to connect with a human, that's on you. The best outputs aren't AI generated. They are AI assisted and human finished. If you disagree with that, and I know some people will, if you think AI can write something that moves people, tell me in the comments, because I'm genuinely curious. Because here's what the reality is. The eight people winning in AI aren't the ones replacing humans. They're the ones building systems. Something that is incredibly dangerous as we move through this is AI will sound just as confident when it is wrong as if it's right. You can tell it 100 times. Stop making things up and it doesn't matter. It's generative by design. Making things up is why it exists. And there's five ways that I separate signal from noise. Assumptions. List every assumption you made and rank them by confidence. Ask the AI itself. Sources. Cite two independent sources from each claim. Include title, URL, one line quote, whatever you choose. Counter evidence. Find one credible source that disagrees and explain the gap. That is a really good one because it will go out there and it will find sources which disagrees with what it's actually told you. And then it will come back and explain the reasoning why. If it's still correct, you've got the right answer. If not, that's why this step exists. Auditing. Recompute every figure, show your math, ask it, tell it to audit itself, and then cross. Model verification is a really cute one. So you do all of that in the one system that you've chose. You run this Same prompt in ChatGPT, Gemini and Claude. Feed one model's output into the other and say, verify this. That is a very, very, very good way. Especially once you have a finalized version that you think is legit to go and get an answer. Okay, that's how you turn AI from hallucination machine into a research weapon. By week three, you'll feel completely in control. But here's the problem. The best AI outputs aren't the ones that sound original. They're the ones that sound like you. This is the final boss. And most people use AI like a vending machine, so they never get there. But you've got to develop something that's unique to you and it's taste. Okay? They go to the vending machine, they push the button, they get junk food, they call it done. Okay, you're past that. It's week four. Now it's time to step into the ring. Treat AI like your sparring partner. Argue with it, push back, sharpen its thinking and your own. That's the sharp framework that I use. S is for specific. Okay? Are there concrete examples and numbers? And if there aren't, replace every general statement with a specific example, including at least one metric or case study per point. Force the AI to do it for you. Human. Does it sound like a person wrote it or does it still sound like a robot? If not, rewrite this in a conversational tone. Use contradiction. Vary sentence length. Add one unexpected Analogy A My favorite argumentative does it take a stance you could agree or disagree with? If not, don't edge. Pick the strongest position, State it clearly, defend it with evidence, then address the best counter argument. That's the way that you would nail that. R Relevant. Is it actually useful for your situation? If not, focus only on advice I can implement this week. Cut everything theoretical or long term or what are your three highest impact actions that I should take now and then? P Punchy Is it tight or is it fluff? If not, cut this down by 40% 50% Remove all filler words. Make every sentence earn its place. Start with the conclusion. If you've enjoyed this video so far, please like subscribe Comment Hype it means more than you would ever know. But remember this. Every prompt that you write, every revision that you push, every judgment that you make, you're not just training an AI model, you are training you. AI is here whether we like it or not. But I'm a pragmatist. I don't think AI is here to replace all human work. I think it's here to restore human worth. If anything, what have you learned today? You now know how to speak AI English with the RAG framework. You know how to build context with Clip, and you know how to iterate, verify and develop tastes using the Sharp framework. You're not just a normal AI user anymore. You're a systems thinker, someone who can turn vague problems into clear frameworks. Someone who can build leverage while everyone else is still grinding. Imagine next week pulling a competitor analysis in 10 minutes. That used to take you four hours. Picture your inbox. A lead comes in your AI qualification bot, takes them in 30 seconds, asks three questions, scores them, and only routes serious buyers to your calendar. Imagine writing a campaign brief that practically writes itself because AI cross referenced your last 12 months were for data and found the pattern that you'd never see manually. That's not science fiction. That's what the 1% of AI users are doing right now while everyone else is still googling best AI tools to use. Here's the cost of waiting another 30 days. I can't make you take all of this practical information and use it to improve your quality of life, both personally and at work. You can make that decision and only you. But here's what the cost is. You'll spend 40 hours this month on work that should take four. Your competitors who figure this out will deliver the same quality at 1/10 of your cost. Doesn't matter whether it's personal for you to use AI whether it be business related, you will become significantly more efficient. If you use this framework that I've given you in six months time, you won't be a AI amateur. If you're a business, you won't be premium, you'll just be expensive and replaceable. But this gives you a full grounding in AI to be able to become part of the 1% that use AI in its most efficient manner. If this hit home hit like hit subscribe Hype the video. I'll catch you in the next one. And this video right here shows you how I would make a million in 2026using AI. And I strongly advise that you check it out. I'll catch you in the next one.