Transcription
If I had to start learning Claude from scratch today, this is exactly how I'd do it. Because here's the thing, you're paying at least 20 bucks per month for Claude, and you're probably only using about 2% of what you're actually paying for.
I spent the last 2 years using Claude every single day for my job, for my business, for this YouTube channel. And if I had to start over tomorrow, I would not do it the way I did. I'd skip over about 80% of what I tried, and I'd learn these 13 things in this exact order.
So, in this video, I'll walk you through my step-by-step syllabus for learning Claude from scratch, starting with the one you've already used.
Number one is Claude chat. You can go to claude.ai, ask a question, and get an answer back. You can access it from your browser, your phone, or your laptop. And it's all the same Claude, all the same conversation.
The thing that really separates Claude chat from something like ChatVT is what it can create. Claude creates diagrams, tools, and visualizations right in the conversation. You can click through, drag sliders, actually work through the problem instead of reading about it.
Claude also builds artifacts which go way beyond widgets. You can have Claude build a PowerPoint deck, a Word doc, an Excel sheet, a full website, even a working app. All of that inside of chat. Now, these are static, so they don't update on their own, but there is a version that pulls live data, and we'll get to that.
But first, you need to give Claude access to your tools.
Number two, connectors. This is where Claude starts plugging into your real work. Connectors wire your tools directly into Claude. things like Gmail, Drive, Notion, Calendar, and Slack. Claude can read from all of these apps directly and even perform actions in them for you and on your behalf.
You might have Claude read and summarize a Slack channel or send out a message to your team for you. If you're working on a project, you might get Claude to read all of your documentation in Notion or Confluence so it has the full picture. Connectors work across your whole Claude account once they're connected. And later, I'll show you how you can start automating tasks inside of these apps using routines.
There are hundreds of connectors to choose from with more being added every week. They work using MCP or model context protocol which lets Claude make a request to the connected software to do something. You can configure your own connectors as well if the app you want to connect has an MCP.
Connectors is when Claude starts doing real work, but to keep that work organized, you need the next one.
Number three, projects. Project lets you organize your work into self-contained environments, each with their own chat history, instructions, and knowledge. You can manage the knowledge of a project by uploading specific documents or giving it access to particular connectors. Every time you chat with Claude in that project, the knowledge is automatically there.
You can give it specific instructions or even share the project with your colleagues so they're working from the same setup. Here's my setup for generating proposal documents. Claude has all of the context from the templates, asks me the right questions, and handles everything from there. Each project I have has its own separate memory, too. So, context inside a project stays inside that project.
Projects really start to come alive when Claude can actually act on your real files. And for that, you need to step off the web.
Number four is Claude Desktop. So far, we've been using Claude on the web, but Claudei is barely scratching the surface of what's possible. There's actually an app you can install on your Mac or PC for Claude, and it's not just to wrap around the chat models. It gives you the ability to use some of the more advanced Claude products and features like co-work, code, and routines.
If you've only used Claude in your browser, install the desktop app. The next four things on this list all live there.
Number five, Co-work. And now we're finally getting somewhere. Co-work can see every file in a folder you pointed at. It can open them, edit them, create new ones, organize them, and even run scripts. When you use co-work, you give it a task and it goes and does that task. It can create a PowerPoint for you, save it to your documents, and then go and edit that same document later. It's when AI starts to do real work.
And co-work is built for people who don't code, but still want the power of the Claude code agent hardness underneath. But the thing is, co-work doesn't get blanket access to your machine. It runs in a virtual environment, and you pick the folders it can see and edit. Projects are useful here, too, because each project gives you a natural file system barrier.
This is the rung where most people's jaws drop because for the first time, Claude is doing real work on real files instead of just talking about it. Co-work also unlocks a different kind of artifact, one I promised earlier.
Number six, live artifact. Live artifacts are the same idea as artifacts in chat, but instead of being static, every time you open one, it pulls fresh data through your connectors. This means that you don't have to go and regenerate your artifacts, which burns a bunch of tokens. Instead, you can build them once and have Claude fetch new data from your connectors to populate that artifact every time you open it or refresh it, but now they're permanently useful.
And these are built inside of Claude Co-work. And Claude can even help you build them for you. Just go into the live artifacts menu, hit new artifact, and Claude will build whatever you need to create. You can connect your tools, build a custom dashboard that's actually useful, and even share that with your team. This is one of the first projects I get people to build when they're learning co-work because it brings so much of it together for them. It's a real aha moment because they see Claude building a real persistent business asset.
Number seven, skills. And this is the one that most people will end up using more than anything else on this list. A skill is just a folder, and inside that folder is a file called skill.md with instructions for how Claude should do a specific job. Claude is pretty intelligent, but it doesn't have the domain knowledge required to do real world tasks, especially ones that are specific to your workflow.
That's where skills come in because they package up knowledge, scripts, and contexts so that Claude can invoke them when it needs to do a specific job. Claude has a bunch of pre-built skills, and you can even get them from the customiz menu super easily. One of those pre-built skills is skill creator, and it's a skill that lets you create your own skills.
The part that makes skills different from everything else in this video is that they work everywhere. The same skill runs in chat, co-work, and in the PowerPoint extension. They can even be shared with other people. I used about six of them to create this video you're watching right now. Brand voice, scripting, hook generation, thumbnail creation. Skills are the multiplier. Every other product on this list gets dramatically better the moment you teach Claude how you actually want things to get done.
Good skills are hard to find, and the bad ones will waste your tokens or worse. So, I'm building a vetted marketplace for the skills I actually use. There's the links in the description below if you want to be on the wait list.
Number eight is Dispatch. Dispatch is the bridge between your phone and your desktop. You can hand off a task from the Claude app on your phone and Claude does the work back at home on your desktop with all the same files and all the same connectors. You get the result back on your phone.
And Dispatch is a fundamental redesign of the traditional AI workflow. Because up until now, we've been actively working with Claude to get our work done. But this flips it on its head because Dispatch is designed to give Claude a task remotely and then have the work completed and given back to you when it's done. It's one shared conversation across the two devices. So you can pick the task back up on your computer once you get back from your walk or the cafe.
So co-work and dispatch handle your work on your computer. But if a lot of your day is actually spent inside other apps, that's when these next two come in.
Number nine, Claude in Word, PowerPoint, and Excel. These are enthropic extensions that put Claude directly inside of Microsoft Office. It's the same Claude you've been using, but now it can actually read your slide deck or your spreadsheet while you're editing it live from within the app.
This is great for working on a file live with AI because you might have Claude generate the first draft of a slide deck from co-work and then go and open it up yourself and continue working on it live in PowerPoint with Claude right beside you. There are extensions for Word, PowerPoint, and Excel.
In Word, it can draft, rewrite, and reformat documents in your tone. In PowerPoint, it reads your slides, your fonts, and your color schemes. So anything it creates stays on brand. You can create full slide decks in minutes. And in Excel, it can build detailed financial models, analyze multi-tab workbooks, explain every formula with cell level citations, and update your assumptions without breaking any of the dependencies underneath.
These extensions share context with each other when you have multiple files open. So if you're working on an Excel model and then need to update a PowerPoint, you can do it without reexplaining anything.
Number 10, Claude in Chrome. This is a browser extension that turns any tab into a place where Claude can do work. It can read what's on screen, click buttons, fill forms, manage multiple tabs, and run multi-step workflows from start to finish. You can trigger the Chrome extension from co-work or from code, or you can even talk to it directly from inside of Chrome. Really, it it's whatever fits the task at hand.
The thing that really sets this one apart is that you can record a workflow once and then Claude learns the steps. So, if there's a clicky repetitive task you do every Monday, you can teach it and schedule it so it happens automatically.
Once Claude can move around your apps and your browser, the next thing you'll want is for it to actually create the visual stuff for you. Design is Claude's visual workspace. It can build slides, infographics, posters, social posts, one pages, and even full wireframes. It's the same models, but the output renders in a designable canvas instead of dumping the message into a thread.
I asked Claude for a five slide pitch deck for a fake AI consultancy. It structured the slide deck, the layout, the copy, all of it. And then I can just ask Claude to change it. I can even export these as PDFs, a PowerPoint, HTML, or push it straight to Caner if I wanted to. If you've ever opened a blank Canva template and st it for 10 minutes, that's the problem that Claude design solves. You can use this to vibe Design Apps, too. It can build wireframes, highfidelity frontends, and once you're happy with it, you can hand it off to Claude Code to actually develop the back end.
Number 12 is Claude Code. Claude Code runs in your terminal or inside of the desktop app and it's the most powerful way to use Claude and you don't even need to be a developer to use it. Most people think Claude Code is just for coding and it is incredible at that. But it will happily do almost anything else you can do on your computer and it's particularly good at go to market workflows. Think outbound content SEO and marketing.
Here's the part that matters the most with Claude Code. It has a context window around five times bigger than co-work so you can run long complex jobs without losing the thread and you can see every file it touches in real time so you can guide it. You can give it different levels of autonomy through permissions. Ask me before every step to go for an hour and tell me when you're done. And because it works inside a real project folder, it remembers your standards, your file structure, your conventions every time.
Here's what I actually use it for. Claude Code runs my entire content operation. It finds ideas, writes my scripts in my voice, creates thumbnails, SEO optimized descriptions, and builds lead magnets that go with each video. It also runs my client work after sales calls. It takes my meeting transcript and writes a proposal. It builds assets for clients. It does pretty much everything for me. That's what makes this different from everything else on the list because it's not a tool that I open. It's the operator of my business.
If you want a full walkthrough of how I set up my Claude, drop a comment below and I'll make that video next. But the real leverage with claude code is when you start taking work off of your plate for good.
Number 13, routines. This is the one that runs while you're not there. A routine is a saved tasks that executes on a schedule. When a webbook hits or when something happens in one of your connected apps, it runs on Anthropic's cloud, not your machine, so your laptop doesn't even need to be on.
Here's one of mine. Every morning at 9, Claude scrapes the Anthropic blog, finds anything posted in the last 24 hours, and posts to a summary into my news channel on Slack. But the real unlock isn't scheduled routines, it's event- driven one. I have a routine that fires the moment Fireflies finishes transcribing a sales call. Claude pulls the transcript, extracts the action items, and files them in my second brain and slacks me a summary before I've even closed my laptop. No clock, no waiting, no more opening anything. The call ends, the work happens.
This is the rung where Claude stops being something you use and starts being something that compounds because every routine you build is permanent leverage. set up 10 of them and you've handed yourself an entire workday back every week forever. I walk through exactly how I run mine, the five levels, the schedules, the web hooks, the gotchas, and if that's where you want to take this, that's the next video to watch. Thanks for watching.