Believe it or not, you're not behind on AI yet. Because the truth is, most people still don't understand how to use it well. By just watching this video, you'll know more about AI than 90% of the people out there.
And if you make it all the way till the end, you'll be in the top 1% of all AI users. I've personally tested thousands of AI prompts across every major platform, built dozen of custom AI tools for my own companies, and helped hundreds of founders turn AI into their most profitable team members. So today I'm going to show you how AI works, the fundamentals you need to use it, how to master it, and how to futureroof yourself so AI makes you a weapon instead of replacing you.
Starting with how it works. Essentially, you've got all the world's information. I'm talking videos, podcasts, documents, Reddit threads, and then AI analyzed it, and then it did this mathematical equation.
It broke all that text into tokens. and then it tries to predict the most logical next token based on what you gave it. It's just like if you say twinkle, twinkle, little, it's gonna say star.
That's why most of the answers that come out of any chat AI have a lot of m dashes. It's not because AI likes m dashes. It's because humans like m dashes and it's trying to replicate what it's seeing us create.
And this is why if you give it really good information, the output is really good because it can use all that input to build context. It's kind of like if somebody tells you all the details of a story and then asks you a specific question, you're more likely going to get it right because you have all the details of the story. That's what AI needs to give you the best outputs and that's where people fall short.
The truth is AI isn't magic. It's just pattern recognition with the fancy suit on. So, now that we've covered how AI works behind the scenes, we need to talk about the fundamentals and how you need to use it the right way.
Starting with the right way to prompt. The more information you give it, aka the prompt, the better the output will look. Did you know you can give it like two, three books worth of information about your question before it even gives you an answer?
Most people are giving it a text message. It's giving it nothing and then you're expecting it to do magic. A few months ago, I was working with a founder that built this really cool AI technology, and he was working on the marketing, and he was having a really hard time getting the marketing to actually communicate what his product did.
He spent like 3 weeks on it, iterating, going back and forth, trying to get the messaging right, trying to get the headline right, trying to get the copyright. All I did was sit down and I gave it the format I'm about to teach you, and I hit enter. And within 12 seconds, it outputed the perfect plan, structure, and strategy for him to get his customers to buy.
It took less than 15 minutes to implement it on his website. And all of a sudden, he's converting that traffic into customers. What did I do different?
Here's how it works. The first thing is the role. You got to tell it who to be like.
It's kind of like act like a professional dancer. Act like a killer musician. So, all I did when I was helping my buddy to fix his prompt, it said, "Act like a worldclass marketing strategist that focuses on conversion for software.
" What that does is it takes all the freaking information in the world that's all been cataloged and structured and it says, "Okay, out of all of this, only focus on this piece. " And then with that, it can then give you an answer that's very specific. If not, it just looks watered down.
The second part to your prompt is the context. I want you to give it all the background information. If you have marketing documents, you have transcripts from a call, you have any kind of standard specification, the more you can give it, it can use that in its little brain for that prompt to guide it in its output.
The third is the command. You have to be very specific. Whatever you want it to do, tell it.
Be very explicit. I always say make the implicit explicit when we're prompting. And the last and the fourth is the format.
Oftent times for me, I don't want these long lengthy outputs. I want bulleted lists. I want it short and concise.
I want it in a table format. I want it in a PDF. The format is how you take that output and then use it in other tools.
You might want to put it in a CSV so that you can import that into another tool, but if you don't tell it that, it's not going to know what format you want. If you do those things, you will notice a complete difference in the output. I'll even go as nerdy in the format where I'll say, "Here's the format, then give it a template to follow, then hit enter, and it can't go outside of those parameters.
" The truth is is the quality of your output will never exceed the quality of the input to your prompt. So now you know how to structure your prompt, but there's this other crucial step that if you don't know, you can't learn AI properly, which is you got to pick one tool and don't switch. The amount of people that come to me and they're like, >> "Hey, I'm using chat GPT and Claude and Gemini and Gro.
" >> And I'm like, "Show me your prompts. " And it's Crayola. It's simple.
They haven't mastered the tool. Learning AI is like learning an instrument. If you tried to learn a little bit about the piano and the guitar and the drums all at once, you wouldn't be really good.
But if you decided to pick one and you decide to be a master at it, the others become way easy after it because you understand music, you understand tones, you understand how to read sheet music. Same thing with AI tools. Pick one and master it before you switch.
Once you understand how one works, the others are way easier to pick up. And I will be honest, the people that keep switching tools all the time, I just call it procrastination. Masters go deep before they go wide.
So know the tools, know which one's right for you, and go deep on it. So, I've already mentioned the top three, okay? We've got Claude, Gemini, and Shad GBT.
I'm going to walk you through each one and tell you exactly which one does what best so that you can pick the right one for your current use case. So, for example, if you do a lot of writing, creative work, deep thinking, code, well, Claude's got everybody beat there. And it could change in a few months, but for right now, I'm telling you, Claude's the one.
The other one, if you do a lot of research, you need up-to-date information, you live inside of Google tools, you can't go wrong with Gemini. It is an incredible language model and it's the one a lot of people that are like Google specific are using. GPT is a great one.
It's the first one to get really popular. It's got a lot of integrations. Can't go wrong with it.
And before you start hammering me in the comments, I understand there's a bunch of others. I live my life on a site called HuggingFace, which is like the nerdiest AI community. So, don't tell me about Grock and all these Chinese models.
I get that. I'm just telling it for most people that are like, I'm ready to be masterful at this so I can become a weapon. These are the three I would recommend.
And by the way, if you choose to go towards Chad GBT, I've got a full freaking master class that I'll link at the end of this video that goes through everything to make you a pro. And if you're a business owner that's trying to break seven figures and you want to go even deeper, I've got an AI implementation playbook, that is what I use in my companies to implement AI, and it shows you everything. Just find me on Instagram and DM me AI business.
I'll send it over to you for free. So, here's the thing. If you made it to this point and you feel like you've got everything covered so far, the cool part, you're in the top 10% of AI users.
But if you stick around and implement what I'm about to teach you, you'll be in the top 1%. So, let's talk about mastering AI, starting with pull versus push prompting. This is where we get a little advanced, but it's fun.
Come with me for the ride. Don't run away. This is going to be good.
Most people use AI the hard way. They do 80% of the work and they let AI finish the last 20%. That's push prompting.
You're telling AI how to do the task. So you want to go from push to pull prompting. And the way that it's different is that you start with the outcome and then you let AI figure out the best way to get there.
Instead of doing all the work yourself, which you're doing today, you let AI do the heavy lifting. The way I like to think about it is push is kind of like when you used to get directions from a friend when they're like, "Come to my house. Come over.
" And then they tell you you got to turn left at the house and then right at the church and then you'll see a red mailbox. That's how most people prompt. Poll is like GPS where you let the AI direct you based on the destination you've given it.
So, here's how you use poll prompting. It's super cool. First, you give it the role in the context like you normally would.
It needs to know kind of who it should act like and what's the information about the scenario. Second is you give it the outcomebased objective, right? This is where you tell it what you need.
So, you might say, I need a sequence that converts cold leads into booked calls. Right? Very specific.
It's like the destination. Third, you say this. Ask me all the questions you need for you to create this for me and then give it back to me in and then you can say the format.
You might say PDF or a table. Then just answer all the questions. My little trick there is I use the voice mode where I can just hit that little microphone and then answer cuz it might give you seven questions and I don't want to type that much.
But you give it all the answers and then let AI do the work. Watch it cook. Now you might say, "Hey, I need to refine this.
Ask me more questions to make it more relevant. " And it'll keep refining. So you keep essentially using that's why it's called the pull method.
You keep having it pull from you exactly what you need and let it go find the solutions. This is how software programmers have been doing it for a while. Okay.
Most coders are sitting there and we're coding through outcomes. It's called spec driven development. That's now becoming the mainstream for how you can prompt.
Now pole prompting is crazy powerful. But if you really want to master AI, you need to make it personal to you. So number five, master prompts for every role in your life.
The way to think about a master prompt, it's kind of like a manual about a part of your life. So if there was like a manual called Dan CEO, what would be in that manual? If there was one called Dan Father, what would be in that manual?
Well, I would assume they would have everything about me, my life, the people in my life, how things are structured, what we sell. When you have these master prompts and you use it to give it to the AI cuz you might switch to different ones, then it'll give you better answers. Cuz right now, if you don't, it's not personalized to you.
It's just giving you generic autocomp completion instead of something that is laser focused to your exact situation. And remember earlier when I talked about giving it context? This is the context.
And if you do it once, you never have to do it again. And it makes prompting so much easier and faster. Every team member in all my companies has to have a master prompt.
They have to show it to me. I got to make sure it's complete. I just can't imagine that they're efficient chatting with AI all day long and it giving great answers if it doesn't have that.
So now let's create one for yourself. The first thing is we're going to practice poll prompting. You can say, "I want to create a master prompt for my role as whatever your role is in life, you know, father, husband, VP of marketing, and ask me all the questions you need to create it for me.
" See how we're going to have it pull from me. Now, it's going to give you all these questions so they can create this document and that's where you use the voice to text feature and just answer each one conversationally. And third, once AI gives you your master prompt, go through it and just refine anything that doesn't feel right.
And you can literally just talk to it and say like update this, add this, you're missing that, you didn't understand this properly, and it'll literally update the whole thing for you. The last part is to save it as a PDF. And I like to just call the master prompt and then the role name.
So I have them on my hard drive and that way when I'm using different AI tools I can always upload it so that it has perfect context on my situation in life. The truth is if AI sounds like a stranger and anybody could have got that output, it's because you haven't introduced yourself. The master prompt is how you say hello AI this is who I am.
The cool part is like I said you can use it in other tools. I personally spend a lot of time in a tool called Manis which lets you create agentic workflows. And when I first loaded up the tool, it didn't know anything about me.
So I used my master prompt to give it all the context so that it could do outputs that were very personalized. If I had to like play with it and tweak it and talk to it, it would have just took me a long time. This gets an AI from stranger to best friend.
So now you got the master prompts. You have a way to give the context to the AI so that you guys are best friends. But what if you're like trying to get an output and you like the output and you want to like repeat it and keep it consistent.
This is where you need the next tool to master AI. Number six, system prompts. Think of a system prompt like a recipe and it has all the structure and you have the ingredients, but the sequencing matters.
How you do what you do in the order will dictate if you have a big pile of mush or you have a beautiful chocolate cake. So if you're chatting with the AI and eventually it gets an output that you're like, "Oh my gosh, this is great. Why did it take me 17 minutes to get to that output?
This is how you solve that problem. Understanding how to create system prompts is the future. So, here's how you build a system prompt using what we've learned so far.
First, you got to start by saying you're an expert AI engineer. I want you to create a system prompt that does X ask me all the questions you need to create. See how I'm using the poll structure.
Next, once it gives you all the questions, answer them using the voiceto text feature. Then, test the response you get back. The third, if something needs to be tweaked, just tell it to refine it.
It's literally you code system prompts with words. Once it's dialed in, then you copy that system prompt into a custom GPT or a Cloud Project or a Gemini Gem. And then you've got a system that works for you on repeat without ever having to give it instructions ever again.
I use this for strategic work. I use it for research work. I use it to create systems for my team to follow where I can give them the link and say, "Look, just give it the topic and it'll do something for us.
" That way, you take the complexity of AI and just make it really simple for everybody else to use. You're basically building little machines that take care of work for you so that you can focus on the fun new stuff. Now, here's a pro tip.
A while ago, there was a leak. All the complex system prompts from the top AI tools, I'm talking Perplexity, Notion, Lovable, and 20 plus others were actually collected and put together. So, if you want to go see what the world's best system prompts could look like, just search system prompts and models of AI tools on GitHub on Google and you will find the link and you can go look at each one of them to understand how powerful these can be created to do literally anything for you like build mobile apps or reconcile financial details.
So, now you know how to use AI better than most people ever will. That's the truth. But here's the thing.
AI is evolving fast. I'm talking every day. And if you want to stay ahead, you need to make sure AI will never replace you.
Which takes us to our last point, but the most important one, which is futurep proofing yourself, which comes down to taste, vision, and care. There are three things that when I sat back and I asked myself like what are the areas of human existence of life that I feel is going to be really hard to disrupt. I feel like things like our taste for what is great and our vision for things that don't exist that should or caring about other people.
What a crazy idea. I really don't think that AI is going to be able to do that. Well, simulate it maybe, but I'm talking heartfelt empathetic conversations that takes you.
Machines can optimize what already exists, but humans imagine what doesn't. First thing is taste. You got to immerse yourself in excellence.
I think of it like standing underneath a waterfall and allowing amazing worldclass things to be poured over me so that I might have the possibility of picking stuff up. So if you're like massively interested in music, go subscribe and use your social media to learn music. If it's code software, I want you to use your feed to feed your mind.
Go and follow all the top people. Go subscribe to the newsletters. The more you consume the best in the world and you get to understand what greatness looks like, the better you build your taste.
The second is vision. Schedule thinking blocks so that you can sit down and think about the future. Visualize concepts cuz your brain doesn't think in words.
I don't know if you know this, but like if I say pink elephant, you just saw it. Okay? You didn't have the words written down in your mind.
So even just doodling, planning, exploring, reading books that'll expand your horizon, that gives you a sense of what's possible or maybe how the world should be shaped around you. And that is your job to shape the AI to help you with the vision. And the last one is care.
And this is why I'm so bullish around AI doing all the boring and dirty and dangerous work for us is then you get to celebrate your clients. You get to celebrate your team. You get to be with your family.
You get to use it to create more because you were born to be a creator. Like that's what you're here to do. Then you have the time to collaborate with people authentically.
Start thinking about ways you can get out in the world and show that you care and contribute and just connect with people. I really think that AI has enabled us to have this deep desire to be around other people. It's why you see events having record number of people showing up and sold out in no time because we have a deep need for connection.
Do these three things and AI will become your weapon instead of your replacement. Now, believe it or not, you're still early. So, stop panicking if that's where you're at.
Most small businesses haven't even started using AI in a meaningful way. So, just by watching this video, you're already ahead of the game. So, here's what I need you to do.
Just pick one daily task and have AI do it. I don't care if it's move information from one system to another. Just 10 minutes of action beats 10 hours of tutorials.
So, just leave me a comment and leave that commitment below. I want to know what are you going to commit to to really upgrade your AI skills. And remember, DM me AI business on Instagram if you're looking to grow your business and I'll send you my AI implementation playbook so you can scale past seven figures.
Like I mentioned, if you want the full chat GPT master class to get ahead of 99% of the people, just click here and I'll see you on the other side.
