Give Claude to two people. One will use it to build a better email. The other will use it to help build a hundred million dollar company. Same tool, different users. And that gap between people who use AI and people who wield it is about to reshape careers and companies and families and lives over the next decade. If you're new here, I've been a CEO, board member, investor in technology companies worth billions. And today I want to share with you how to use Claude better than 99% of people, from beginner to advanced. And it's way easier than you think. Claude is not just one app, it's a whole stack of tools built for a very different kind of work environment.

For businesses, there's chat for thinking, Claude co-work to get work done, Claude code to build things, Claude Chrome to browse with a brain, Claude skills to make it all repeatable, and lots of tools. Once you understand how this team of agents work together, your output can change dramatically. The best way to visualize it is this: five activities, think, remember, execute, build, and browse. Let's go through each one of them. First is think. You go to Claude chat. And this is what most of us do when we think about chatbots. We input a prompt, we get a response. You bring a messy problem, a fuzzy idea, or a rough draft and work through it.

But many of us just use AI to do all the work so they can copy and paste. We have all seen those LinkedIn posts. So, if you go to Claude chat and type write me a LinkedIn post about leadership, you'll get a very polished paragraph that says nothing. And we've talked about how to prompt better with things like aim before actor input mission. Try this prompt instead. You are a business school professor. I am trying to explain why some leaders sound credible and others don't. I'm attaching my notion page with a rough idea that talks about leadership and authenticity, leadership and humility. Your job is to push back on what's weak and sharpen what's real. Find me research that helps me create a strong insight and then we can write a LinkedIn post. You make Claude your thinking partner now. Of course, like all AI, it will not do what you tell it to do, not the first time around at least. And sometimes it's going to be super frustrating because you will give this detailed prompt and just crap will come back. So, there are three things that you can do that can become your power moves when that happens. First, give it rich context as much as you can. Connect the chatbot to your Google Drive, your email, your calendar, your notion app, anything else you can. That way you can attach files to that chat session. It will read it, it'll get context from it. Second, ground your insights with research. You can ask Claude to go out on the web and do deep research. It has already learned the entire internet during its training, so why not use it? But when you get the research back, always ask one more time, is this verified? And third, use AI to build your AI skill in real time. If you are patient and curious, you can literally build your engine while you're flying the plane. For instance, let's say you have no background in corporate finance and you're building a financial model. You get stuck on how to use a specific spreadsheet feature or even a financial concept like IRR or free cash flow. No problem. Just start another chat in another tab and ask away. So, you can learn and do. Do and learn. The cycle just continues.

Now, let's go meet our second member of your team. This one helps you remember and it's called Claude Projects. We all know when we're creating something new or solving a complicated problem, the work doesn't end in one session or even in one chat. And that's why all three AI platforms nowadays have gotten better at maintaining memory across different chat sessions. Claude remembers a world through projects. A project is where you give the chatbot your files, your instructions, your tone, your ongoing work. You're not starting from zero every time you open it. And of course, I would be careful about sharing confidential documents from your office, but let's say you're actively or passively looking for a career change or a new job opportunity. You can create a project. You drop in your resume, your LinkedIn profile, the two or three job descriptions that you like or the company you're targeting, your writing samples, your professional bio, the other resumes that you like, anything you can think of. And you can customize this project with exactly what you need. So, you can add something like, "You're the world's most badass recruiter and resume writer who has placed candidates in top companies in my industry. I am a brand marketer with 5 years of experience at a large agency called WPP. Your job is to help me create a great story for my next adventure, for my next company, and make me desirable for the next role. I write in a direct, confident tone. I don't like corporate fillers. I hate buzzwords. When I ask you to tailor something, match the job description's language without losing my voice. When I ask you to do deep research, make sure you deliver only the relevant and verified data. Don't make it up. And when I ask you for feedback, be honest. Tell me what's weak and what I need to work on. Be my constructive partner and strengthen those areas." This is one way to provide instructions, but you can see that there are 100 ways to do it. And that's the kind of instruction that makes any project your own. Not a prompt that you're copying from some YouTuber, but a brief that you write for someone you actually know very well, yourself. From that point on, every time you come back, Claude already knows what you want and who you are. You can ask it to tailor a cover letter, craft a specific bullet of your resume for a specific role. A project turns Claude from a one-off conversation into an ongoing relationship between you and your AI. This is also why I don't customize Claude at the account level. I like customizing it at the project level, because that way I can give it specific context. Because using AI to plan your vacation is very different from using it to plan your retirement.

Now, let's go to execute. And for that, we have our third team member, Claude co-work. Chat is where the work begins. Co-work is where it gets done. You start in a prompt window, but finish it on your desktop. Co-work is a desktop application for Mac or Windows. Unfortunately, it's available only on Claude's paid plans right now, but with Claude Work, instead of asking a question, you give it a task. When do you use Claude Work versus chat? Well, you use Claude Work when you have a lot of assets already on your computer's hard drive, when the task needs multiple steps, and when you need that output to be on your local drive as a local file. So, for example, you can tell Claude Work to do the following. Take all of these customer interview transcripts from this folder, these spreadsheets of accounts, and notes from last week's product meeting, find top three reasons on why the customers are churning. Why are they leaving? Provide all the supporting data, create a very clean slide deck that I can use tomorrow. Put the slide deck in this folder. And here's what makes Claude Work even more interesting. It can reach out to other AI tools through a connector called MCP, the Model Context Protocol. Think of MCP as giving Claude a set of keys to other AI tools. It can knock on their door, unlock the capabilities of those tools, act with them, act through them to give you results. Let me show you what that means in a real example. Imagine you run a business selling handcrafted candles. You have a small business. You have a terrific product, but you're bootstrapping your business, right? And you don't have a huge marketing budget. In the old world, that would mean that you couldn't create any ads that looked super polished or professional. But now, look at what you can do when you can plug Claude into a creative AI tool called Hicksfield. They are the sponsor for this video. Now, Hicksfield is used by 18 million backed by tier one VCs like Excel and Menlo Ventures, and they've already crossed the billion-dollar valuation. And they were the first creative platform to ship an official MCP connector for Claude. Here's how you connect it. You have to do it once, and that's it. Go to settings, connectors, plus button, paste the Hexfield URL, and you're done. That's it. Now you can go into Co-work, and you tell it to do whatever you need to sell more products. For instance, you could say, "Generate 20 beautiful ad creatives for my new lavender candle line. Three ideas for Instagram, three editorial close-ups for the website, and a vertical video for the launch. Write a tagline for each." And that's it. It just runs. Claude Co-work writes the prompt and passes it on to Hicksfield. Hicksfield generates the images and the video, and the entire set of files land in your local folder without you needing to hire a creative agency and pay them thousands of dollars. That would have taken you 3 weeks of a professional shoot, a photographer, a stylist, and weeks of post-production, and of course thousands of dollars. But now, you can do it in a single session with Claude and Hexfield. This is how AI is leveling the playing field. A solo founder with the right set of tools today can now produce the kind of creative output that used to require a large team and a large budget. There's a link in the description below. They have a free tier. Take it for a spin. Try it. And you could build something very cool and interesting this weekend. Two things to keep in mind. First, the output will be a starting point. You'll still have to go through each and every slide and make sure it's good. But a lot of heavy lifting is done by Claude co-worker already. Second, give explicit permission for Claude co-work to access only specific directory and specific files. Claude is very conscious about security, but please be very mindful about what you give access to. And I usually create a subfolder and move all the files and data into that folder, and that's all Claude has access to. It's called sandboxing, so the rest of your hard drive remains secure. Claude can't get to it.

Fourth is build, and this is where Claude code comes into play. This is the part that makes a lot of us nervous. It made me nervous. The moment we hear code, we're like, "Ah, I don't know." We assume that all that stuff is only for engineers, but that assumption is one of the biggest missed opportunities in AI right now. If you can type in English, you don't need a computer science background to code and build something cool and useful. Let's say you have a consulting business or a side hustle. Simple example, just go to Claude code and prompt, "I want a dashboard where I can drop in my sales pipeline and meeting notes and instantly see which deals are at risk, what's stuck, and what my next move should be." That's all you need. You don't need to think like an engineer to start using Claude code. You just need to be clear about the problem you want to solve and the thing you want to build. And of course, you have to have the right data so that dashboard you build is useful. And finally, the fifth team member is for browsing the internet. And for that, you can use Claude Chrome. And you're going to go, "Wait, what? Don't worry. Claude has this little extension that you can install in your Chrome browser. Google already has something like ask Gemini button baked into the browser now, but now you can let Claude see what you're looking at as well. Help you process it and act inside your workflow in real time. And I think about this a lot nowadays, because most of my work happens inside the browser, right? My files are on Google Drive or a Microsoft OneDrive, my emails, my scheduling, my team meetings, my research, filling out forms, making purchases, planning business trips, all of it is taking place using the browser. And now Claude is right there with you for all of it. You're on a job listing site, Claude can read it and help you tailor your outreach. You're reading a 40-page industry report, Claude can pull the three insights that matter the most in real time. So, that is your team. There are other team members, but these five are the most interesting and useful. This is your stack.

But even with the full team working with you, one gap can make all the difference between the average and the elite mode. That's where we go next. Any AI's results depend on you. Even with the full team of agents working for you, Claude is only as good as the input you give it. But the superusers who get exceptional output from Claude, they direct Claude the way they would direct a smart person they've just hired. And the easiest way to do that is a framework I call Prime. P is purpose. When you prompt Claude, give it a precise goal. Tell it what you need it to do, like help me turn this messy proposal into a five-part organized document that looks like a sharp client memo. R is research. We talked about it before. In many cases, you may need external research to ground the response. Ask it to go out and fetch it for you. Ask it to verify it. And sometimes the raw material may come from you, your notes, transcripts, files, examples, background, constraints, anything that Claude needs so that it can stop hallucinating or guessing. I is for interview. This is the hidden gem. I really like how Claude does it. In a lot of cases, you'll see that it will start asking you multiple-choice questions. And you pick the right choice, and based on what you clicked, it refines its approach to a specific way of responding to you. M is for mechanics. Here, you tell Claude how the output should look like. Do you want bullets or a paragraph? Do you want a document or a table? Do you want it concise or you want detail? Strategic or conversational? Anything you choose. And finally, E is for examples. Show Claude what good looks like, a format you liked or a tone reference or an outline that worked before. Examples are one of the fastest ways to quality. Let's go through an example. You have an important presentation coming up tomorrow. Run it through prime. Purpose. What is this meeting really for? Do you need an approval, alignment? Are you asking for resources? What is the outcome you really want? Research. Give Claude the context. What's the audience like, the risks, the data, what your team cares about, external resources, internal docs, anything you can get. Interview. So, tell Claude, "Interview me before you respond." Now, it will ask you questions that sharpen its own scope and your thinking mechanics. Now you know the output of that deliverable. You want a 10 slide deck and talking points. And finally, examples. Give it a past deck that you liked or style you admire. So that is prime. Once you start doing that, AI stops being a tool and starts becoming your edge. And by the way, if you want the frameworks like these in your inbox, subscribe to my newsletter. The link is below somewhere. I mean it's free.

The real edge is the human using it. Everyone watching this feels like someone else is getting ahead faster. That they are smarter or they are richer or they have better resources. I know I feel like that sometimes. That is the default condition of ambition. The question is not whether you were outgunned. The question is how you outshine when you are. You have been carrying labels for years. I am not technical. I'm not a designer. I don't have an MBA. But those are just labels. For $20 a month, you can hire something that is already better than most MBAs and PhDs now. You can think more sharply with it. You can design better. You can code. You can build. But do you have the curiosity to keep learning? The tenacity to get through that machine friction. The patience to build real capabilities. Or will you slowly replace real effort with convenience and lose your instinct to wonder? AI will give you a lot of things. But what AI can't give you is that innate curiosity that made you press play on this video. And you already have it. As one of my mentors used to say, "You are the one you've been waiting for." Thank you. And I love you.