Transcription
There's a massive gap opening up right now between people who use AI like beginners and people who use it like pros. And honestly, most people are stuck on the beginner side without even realizing it. That was me just a year and a half ago.
Today, I'm walking you through the six skills that changed everything for me. Skills that'll help you save time, get way more clarity, and honestly, just get better results. So, let's start with skill number one, tool selection.
Okay, so here's one thing I've noticed. A beginner tries to use one AI tool for everything. A pro knows that different AI tools are better at different things. I like to think of it as building a small team. You wouldn't hire one person to do design, marketing, accounting, and engineering, right? Same thing here.
Now, before I go any further, you don't need to learn all these AI tools all at once. That would be pretty overwhelming. The goal is to master just a few. The minimum set of tools that can handle your most frequent tasks. Start small, get really good at those, and then expand from there. And here's why this matters. The tool that you choose can change the quality of your result. ChatGBT might technically be able to generate an image, but Nano Banana, which is Google's image generator model, will do it a lot better.
Okay, so first you've got your brain, aka your large language models. Tools like Chad GBT, Claude, and Gemini. Chad GBT is your everyday workhorse. It's like a jack of all trades and can do a lot of different things like draft an email, brainstorm ideas, do research, generate images, etc. It's versatile and handles most tasks well.
Now, Gemini is your powerful document analyzer and image generator. It's great for when you need to review a long document or report. Now, because it's Google, the web search abilities are amazing, which is great for when you want to do research and if you're part of the Google ecosystem, it gives you seamless integration with your workflow. It's great for when you want to do business related tasks or when you want to create amazing visuals without having to switch tools.
Now, Claude is your human sounding writer and coding partner. Use it for when you have customer-facing content that needs personality like blog posts, social media posts, and marketing copy or when you're building an app with natural language.
Next, you've got your researcher tools like Perplexity and Notebook LM. When you need current information with verified sources, Perplexity is built for that. It's got a source-first design that gives you the option to tell it what type of search result you want, whether it's a web search, strictly academic reports, social media, even finance reports. It's your fact-checking specialist.
Now, Notebook LM is for deeper research and synthesis. It can search the web for sources, but where it really shines is when you upload your own materials like notes, PDFs, YouTube videos, and articles. It helps you search for, summarize, and connect information across all of it. It can locate sources directly within your documents and tell you exactly where it got the information from. Think of it like an AI research assistant that can both find new sources and make sense of everything you already have.
And finally, your creator tools like Midjourney, Sora 2, Google's Nano Banana and v3.1 models for image and video generation. Then you have Ideog for graphics and diagrams, for audio, etc. These are your artists for whatever visual and audio assets you need.
So those are your three categories. Your brain, your LLMs, your researchers, and your creators. Master one or two tools in each category and you're set for about 90% of your work. Now these categories will evolve. New tools are launching every month. So, how do you stay updated without getting overwhelmed? You can check the LM arena leaderboard. Different AI models compete based on real user votes in blind head-to-head challenges. You can see which models rank highest for writing, coding, image generation, whatever matters most to you. Rankings update constantly as new models launch. So, if you're ever wondering, is there a better option? This is a great place to look.
All right, so now you know the categories and you have an idea of which tools to use. But if you want a clear roadmap to actually integrate AI into your work, especially when using Chachib, then I suggest you check out this free resource from HubSpot Media called the ChachiBT at Work Bundle. Here's what's inside. First, a 10-step adoption checklist that walks you from defining your business problem all the way to scaling AI across your team. Second, a content refinement checklist for accuracy, brand alignment, and tone. Third, brand voice templates that teach AI to match your writing style. And it also includes a when to use ChatGBT flowchart which is super practical. My favorite part, the bundle includes over 100 ready-to-use ChatGBT prompts for everything from answering customer questions to creating content to analyzing data. So instead of staring at a blank prompt box wondering what to type, you've got proven templates you can copy, customize, and use immediately. I recommend you download this guide using the link in the description. And thanks to HubSpot Media for sponsoring this video.
Cool. So you've got your toolkit. Now, let's talk about the most important thing, how to actually think about working with these tools.
All right, so skill number two is problem clarification. Here's what separates beginners from pros. Beginners jump straight to AI with a vague idea. Pros spend 30 seconds clarifying the problem before they even open ChatBT. And I learned this the hard way. Early on, I'd spend 20 minutes going back and forth with Chacht and getting frustrated because the outputs weren't what I wanted. And then I realized the problem wasn't the AI. The problem was that I didn't know what I actually wanted.
Think of it like going to a restaurant. If you tell the waiter, "Bring me food," you'll get whatever they feel like bringing you. But if you say, "I'm vegetarian. I want something spicy. I only have 30 minutes." You'll get exactly what you need. AI works the same way. So before you prompt, ask yourself, "What am I trying to achieve? Who is this for? And what does success look like?" Because here's the thing, AI can only be as clear as you are. If your problem is vague, your output's going to be vague, too. But when you clarify the problem first, everything becomes easier. The prompts get clearer. The output gets better. You stop wasting time trying to fix something that was unclear from the start. This is the skill that makes everything else in this video work.
Now, let me show you how to translate that clear problem into effective prompts.
Okay, skill number three is effective AI communication. You probably know by now that how you prompt AI matters a lot. But this is where I see a lot of people get stuck. They write a basic prompt, get a mediocre result, and then blame the AI. But it doesn't have to be that way. Let me show you four techniques that will transform your prompting game.
Technique number one is the six-part framework. Role, context, task, format, rules, and examples. And this is table stakes. This is the core foundation that you need. Now, I've done a full deep dive on this in my prompt engineering video, so I'll keep it brief here, but the core idea is simple. You're giving AI a complete brief. Who should it be? What's the background? What's the job you want it to do? How should the output look? What are the hard rules? And what are the examples that guide the style? For example, when you wanted to write like you, you need to give it examples of what your writing style is like. And here's a quick pro tip. For best results, put your most important information at the end of the prompt because large language models typically put more weight on more recent information. The beginning is useful for context setting, but the end often has the strongest influence on the model's response. And that's the foundation. So now let's build on it.
Technique number two is to show, don't just tell. This works for everything from writing emails to coding to apps. Upload an image, PDF, screenshot, whatever. The difference in output quality is massive. Let me show you. Let's say I want Claude to create a landing page mockup. The text-on approach would be: "Create a landing page for my productivity app and make it modern." Now, let's watch what happens when I show it a landing page that I love and I say, "Here's a style I love. Create a landing page like this, but for my product." Look at the difference. Same request, but now it matches the style and the structure because I showed it instead of just describing it.
Now, technique number three is metaprompting. So, getting AI to actually help you create and improve your prompts. So, you can ask it to act like a prompt engineering expert and build the prompt for you. For example: "You are a prompt expert. I need to write a LinkedIn post about AI trends. What's the best prompt structure?" Or ask it to help you improve your prompt. For example, if your prompt didn't work well, you could say something like this: "I used this prompt, but the output wasn't great. Suggest improvements." AI knows how AI works best. So, this works really well.
And technique number four is to force it to critique its own work. After AI creates something, follow up with: "Now, critique your response. What are the weaknesses? And provide a revised version." You can also ask it to rate its output out of 10 and then to improve it to a higher score and provide that revised output. It's like having a built-in editor.
But even with perfect prompts, there's still one big problem which leads me to skill number four.
Skill number four is verification, or as I like to call it, the truth detector. One of the biggest fears that people have with AI is: "What if I trust it and it's wrong? What if I look stupid?" And that fear, it's valid. Because here's the uncomfortable truth. Your AI co-pilot is incredibly smart, but also a confident liar. It makes stuff up called hallucinating. But here's the good news. You don't have to be paranoid. You just have to be smart about verification.
So, let me give you three techniques to become a better fact-checker.
Technique number one is to fact-check with a specialist. If Chad GPT gives you a hard fact or statistic, take it to Perplexity and ask for the source. Perplexity gives you actual links. So, if it can't find the source, then that could be a red flag.
Technique number two is to interrogate the AI. So after it gives you an answer, ask: "For each claim, rate your confidence: virtually certain (95% and above), highly confident (80-95%), moderately confident (60-80%), speculative (40-60%), or low confidence (below 40%). Explain your confidence level." You'll be surprised at how often it actually backtracks and says, "Well, I'm only 60% confident because I don't have real-time data." Also tell it that if it doesn't know the answer that it should just say "I don't know." Give it permission to admit when it's guessing.
Technique number three is to get a second opinion. Take the output from Chachi BT and paste that into another AI brain like Claude or Gemini and then ask: "Critique this response. What's missing or what's biased? What's wrong?" Claude will often catch things that Chachi BT missed or even Gemini missed or point out where the logic is shaky.
Bottom line is never blindly trust AI. Always verify. Use these three techniques and you'll avoid a lot of embarrassing mistakes.
Now, we've been talking about using AI tools individually, but what if you need to combine them?
Skill number five is workflow orchestration. By now, you know how to use individual tools, but real power comes from combining them. And there's two ways to do this, manual and automated. Let's start with manual. This is what I call tool stacking, where you decide the sequence and switch the tools yourself. Let me show you a real example. I want to create a LinkedIn post with a custom graphic. Step one, I'm going to use Perplexity and say, "Find me five stats about remote work." And now I'm going to get some verified data. Then for step two, I'll go to Chat GPT and say, "Act as a LinkedIn expert and write a three-post series." Now we've got the copy. And step three would be go to Ideoggram and create a graphic that visualizes the main stat. And now you have a LinkedIn post draft with custom copy, stats, and a graphic. That's manual tool stacking. Each tool does what it's best at. So, you can use this method when you need creative control, for when it's a one-off project, or when the workflow is unique.
But what if you're doing the same workflow every week? That's where AI agents come in. An agent doesn't just respond to prompts. It autonomously plans and executes workflows. With manual tool stacking, I orchestrate everything myself. But with an agent, I set a high-level goal. Like, "When someone fills out this form for my landscaping company, the AI agent autonomously prepares a quote for a lawn care service and drafts an email to the customer." Or if you're a freelancer, the agent might pull client info, check your calendar, draft an email, and send a booking link automatically. With manual, you're the director. With agents, you're the executive producer. You set the vision, and the agent directs itself.
Well, you might be wondering, how do I know when to use a manual versus automated? Here's the decision framework. Use agents for repetitive, predictable workflows, weekly reports, customer support, quote generation, and content pipelines. Anything following the same pattern every time. Use manual tool stacking when you need creative control, it's a one-time thing, or every step is unique. The skill is knowing when to automate and when to stay hands-on. Here's a pro tip. Don't automate a messy workflow. If your process is broken manually, automation is just going to make it fail faster. So fix it manually first and then automate.
But even with all this automation, there's still one final skill that I need to talk about.
Skill number six, which is super important, is called the human polish. And here's why. I see people all the time who've mastered prompting, who know all the tools, who can generate content at lightning speed, but their output, it sounds like AI wrote it. And that's a problem because the moment someone realizes AI wrote something, they stop trusting it. AI is fast, tireless, and generates tons of output. But it has zero vision, zero taste, and zero care. And honestly, that's fine. That's not AI's job. Its job is to do the heavy lifting. But our job is to make it sound human. Think of it like this. AI writes the first draft, and you and I make it great.
So, there's three ways to add that human polish.
Polish number one is to inject your voice. AI is pretty generic by default. AI might write something like, "Consistency is key to achieving your goals. Develop a routine and stick to it." Technically correct, but also completely forgettable. Here's how to make it yours. You could say something like, "My old coach used to say, 'You don't get strong on the days you feel like lifting. You get strong on the days you don't.' That's consistency." Same message, but now it's your voice. And that's the vision part.
Polish number two is to trim the AISMs. That's the taste part. AI loves corporate speak. For example, it might say something like this: "In today's ever-evolving landscape, leveraging cutting-edge solutions are paramount." Now, there are ways to prompt AI to use less corporate jargon or not as much corporate jargon, but you still might get output that sounds something like that. And in those situations, this is where the taste part comes in. You can rewrite it to something like this: "Here's what works." Same thing, but way more human. That's the taste part.
And polish number three is adding that human connection, the care part. AI doesn't understand your audience. It'll write something like, "The data indicates that 47% of users experience this issue." But you know your audience. So, you can rewrite it to something like this: "I know this is frustrating. I've been there, too. Staring at a screen, feeling stuck, wondering if you're missing something obvious." That's the care part. You're writing to a real person.
So, AI does the heavy lifting, but you add the vision, taste, and care. And look, I know you can prompt AI to sound more casual, to avoid jargon, to be more empathetic, and you should. But here's what I've learned. That even with the best prompts, AI still maybe gets you about 80% of the way there. That last 20% is the part that makes the content actually connect with other people. And that's always you. Your stories, your voice, and your judgment about what your audience actually needs to hear. That's the difference between content people scroll past and content people stop, read, save, and share.
So, those are the six AI skills. And here's the big takeaway. Mastering AI isn't about memorizing prompts or learning every single new tool. It's about mastering the process. The AI is the doer and you're the director.
If you enjoyed this video, then please give it a thumbs up. It really helps the channel. And if you want to learn more, I've just started my own community called the Innerverse Circle on School where I'll be doing weekly videos, live lessons, and you'll get access to a community of like-minded people who are learning how to leverage AI and automation in their business, work, and life. You can find a link to join in the description below.
Now, if you want to learn more about how to use AI to level up your work and your life, then click this next.