In this video, I'm going to walk you through five insane ways to use Claude Co-work. So, obviously, Claude has been absolutely blowing up in the AI space, but not a lot of people know how to use Claude Co-work, which is way more powerful than the regular chat in my opinion. So, we'll go over what Claude Co-work is, why it's different than regular Cloud Chat, and also five great ways to use it in your own life. Before we dive in, you will need the Claude desktop app as Claude Co-work doesn't run on the web. So, yeah, it has to be the desktop app. Second, you need any paid Cloud plan in order to use Cloud Co-work. So, I'm on the pro plan, which is $20 a month, and that's enough to do everything that I'm about to show you. And then finally, you'll want to set up all of your connections in order to have Claude actually connect to all your apps. So, for this video, I'm going to set up Zapier MCP, which gives Claude access to over 8,000 apps instead of you having to hook up every app one at a time. Okay, so let me show you guys how to do that step by step. So, first you are going to open up Claude on your desktop. So you should land on the co-work view and that's the main task screen where everything runs from. So then you're going to click into customize and then hit the connectors tab on the left. This is where you plug in every external app or connection. So we're just going to type Zapier into the search bar at the top of the connectors list and Zap year pops up under not connected. So we'll click it and cloud's going to redirect you to the authorization page. And then now you're on the Zapier MCP authorization page. So you're going to want to make sure the right account is selected in the dropown and then you're going to hit allow. So the button switches to authorizing with a little spinner and then just give it a second and then the browser is going to send you back to the claw desktop app and you'll see a little popup in the top right that says connected to Zapier. So once you see that it is live. So Zapier shows up in your connectors list and also shows you a bunch of tools you've already connected. So, I've already connected many tools for this demo. And yours might look a little bit more empty if you just started using Zapier MCP. So, now let me show you how to add more tools to your Zapier MCP, which you'll need to follow the tutorial. So, you're going to want to click the link below to head to Zapier's website and create an account. And once you log in, click the MCP button in the main page. Here, I have a bunch of tools that I've already connected. If you don't connect your tools from here, your cloud agent can't actually use them. So, yeah, make sure you guys connect all the tools that you need. And you can click the purple add tool button in the top right of the tools panel to do that. So yeah, once you do that, then you've finished setting up your Zap Your MCP. Okay, so now let's talk about the first use case, which is using Cloud to create Reddit market research reports. So let's say you're building a product or you're trying to make content or you're just trying to understand what people actually think about AI tools. AI is definitely one of the best places to find these types of opinions, but of course, no one really wants to sit there to read a bunch of Reddit threads for like 3 hours. So, what I'm going to do is I'm going to ask Claude to pull feedback from Reddit, organize it by product, and then turn it into a clean Google doc. On this Google doc, we're going to have like the main complaints, praises, trends, stuff like that. Now, of course, this probably won't apply exactly to your situation, but watch this and you guys will see it's very easy to just change it up a little bit to your exact need. Okay, so here is the prompt that I'm giving it. It's quite long, but basically, it says, "Go scrape Reddit through Amplify." So, that's the tool that does the actual scraping. and then pull what people are saying about every major AI product and then dump it into a Google doc. And then real quick over on Zap Your MCP, this is what we have. Appify, Google Docs, Google Sheets. These are the three tools that it has access to. So now we're going to see a popup and it says which AI products should I gather feedback on. Has a bunch of different options like ChatG and Claude, but I'm going to say multi-product mix. Then it's going to ask how many Reddit posts per query. And I'm going to select 25 just to keep it a little bit faster for this demo. And yeah, as you guys can see, it checks with me on the stuff. Like it asks me different questions before it burns a bunch of tokens and time going in the wrong direction. So now it is spinning up six keyword queries at the same time, right? Like it's going through chat sheep's problems. It's going through cloud. It's going through like different bugs, Gemini. And then here is the finished Google doc. So we have sections for every prot. And then every claim has a citation that even gives low evidence flags when only one or two people say something, right? That way you can take that specific point with a little bit less emphasis. And so yeah, that's basically like 30 minutes of co-working what would have been a half day for me. It's essentially a research intern that doesn't sleep. It could browse the web using Ampify and it can more specifically browse forms like Reddit. Now before I show you guys the next one real quick, here's one thing about Zapier MCP that you guys should know. If you only need cloud connected to one specific tool, a native MCP set might be cleaner. But the reason why Zapier's MCP is really interesting is because you can connect to thousands of apps from one single place. So for this type of business workflow where I'm jumping between like Gmail, the Sheets, Docs, Notion, GitHub, and everything else, Zapier, in my opinion makes way more sense. Okay, let's talk about the second use case, which is basically turning receipts into a tax spreadsheet. Obviously, every business owner hates receipts and having to deal with them. So, if you have a folder full of PDF receipts, at some point those need to become a spreadsheet for taxes or bookkeeping or expense reports. So, either you're going to do that yourself or you're going to pay someone to actually manually copy everything over. And what people don't know is you can actually use Claude to read the receipts, pull out the merchant, the date, the total you spent, the tax, and the category, and organize everything inside Google Sheets. That way, just give that Google sheet to your bookkeeper or your CPA, and it's way easier. Plus, you'll definitely want this in case of an audit. It's just going to make your life a lot easier. So, here's the prompt, and I'm going to be pretty specific. Currency rules. We have a don't fabricate rule, and then skip anything that's not a receipt. Basically, I'm just telling it all the rules I want to follow. And as you can see, before it does anything, it plans out these four steps. So, extract, create the sheet, append the rows, and verify. And then now, it's reading the PDFs one by one through local commands. And as you can see, it just hit a CSV file that wasn't a receipt. So, it just skipped it and it just did all that on its own. And obviously, this is like sample data, but we have the final answer. So, 21 receipts processed, $5557 total, and then one file skipped. And then we pop that into Google Sheets, and we have every row filled in. It's sorted by date, payment method, category. So, looks really clean. And so, yeah, this is the kind of task where the Zapier middle layer is totally fine. It's just structured data going into a spreadsheet. Okay, so here's the thing though. I don't want to ask co-work to do this every single month. So what if we actually just create a script that does this on its own? And that brings us to the third use case, which is basically turning a repetitive task into a script. So this next one's really important because you don't always want AI to redo the same task over and over. If I'm going to process receipts every month, I don't want to keep paying Claude to think through the same problem again. I'd rather just have it build me a simple tool that I can reuse. So this time I'm going to ask Claude to turn that receipt process into a small Python script. Basically we're going to use tokens to do the hard thinking one time and then it's going to give me a repeatable system that I can run again later. All right. So here is what I asked it to do. Basically build me a little mini app that lives on folder and doesn't mess with anything else on my computer. Give me instructions. Give me the actual code and make it a real thing I can reuse. As you guys can see, it just builds the whole thing and there's a nice little folder with everything organized inside of it. If I go look at the folder, here's what's inside. So, the main script, the instructions file, a list of what it needs to run, and then the final output. So, this part is really interesting. It actually wrote a little heads up note saying, "Hey, this only works for these specific types of receipts I just looked at. It's not magic." So, it's not going to handle every receipt in the world. So, it told me what it can't do. And of course, we could teach it more things, but I'm going to save that for a different video. And then now, I'm just going to hit run. And then the same 21 receipts, $5557 total. Same answer every single time. And then so the next time you have a pile of receipts, you don't even need to go back and forth with AI. Just point it at the script and then say, "Use this." So it's going to be a lot faster. It's a lot cheaper. And you know exactly what you're going to get. So yeah, essentially what you're doing is you're using Claude Co to do something hard once. It's figuring out how to actually do it. And then you can write the script so that you don't pay for the brain every single time. And since it's just Python, your automation can actually be used on whatever model is hot right now. Okay, let's talk about the fourth use case, which is basically updating a CRM from Gmail. Essentially, I want to see if Claude could update a CRM for me. And this is actually a huge problem for a lot of businesses. Your CRM, it's supposed to tell you what's happening with leads and customers, but the actual conversations like usually live inside of your Gmail. So, what I'm going to do is I'm going to ask Claude to read a few client threads, figure out the deal status, and then update a notion CRM with names, companies, stages, and follow-up. Okay, so here is the empty CRM in Notion. We just have a bunch of column headers, and there's not really anything inside of it. So, here's the prompt. Read my Gmail threads, sync them to Notion, fill in the status, the stage, and the deal value. And that's it. And now what it's going to do is it's going to pull five client threads out of my inbox. So, we have Priya, Devon, Allesia, Marcus, and Jordan. And it writes all five contacts straight into Notion. And if we go over to the database, yep, it is all populated. And then we have the summary. So, it says $24,000 closed, $18,000 open, two follow-ups overdue. And also told me that there were two date fields. It couldn't map directly through Zapier. And so, it have to use Notion's API for those. So essentially we can basically take data from one set which in this case is Gmail use a cloud agent to analyze that data and then take that information and put it onto any other platform which in this case is our CRM. So the fifth cloud agent use case I want to talk about is using it to build and publish a team dashboard. So let's say your team is using a bunch of AI tools and you want a simple dashboard to see who's using the most tokens, how much you're spending and where the usage is going. So, I'm going to basically ask Claude to build the dashboard, publish it online, and give me a live URL that I can actually send to my team. So, here's the prompt. Build a single webpage file, load the chart library straight from the internet, dark theme, and then deploy publicly so the team can see it. So, it's just going to go, it's going to plan it out, it's going to write the web page, it's going to set up the project folders, it's going to make a readme file, and then it's going to save everything to my computer. So, let's open up the HTML page. And yeah, it's a solid looking dashboard. But right now, I cannot share it with anyone. So, let's actually make it online. And so, first we're going to upload the project to GitHub. So, it's going to hit the GitHub authorization step. And it says, I shouldn't ask for your credentials. Here are the exact two commands to run on your machine. So, what I'm going to do is I'm going to run those two commands myself. I'm going to use Versal to finish it. And okay, we have a live page, a real URL. It's a a-token-leerboard.ver.app. app and when I share this URL to anyone, they can actually open the dashboard through this link. So, if you look at this dashboard, right, you can see 21.1 million tokens, $334 in spend. There's a daily usage chart. There's a team leaderboard with Maya, Diego, Aisha. And yeah, this is a great type of little tool to build out which holds your team a bit more accountable because yeah, right now with AI, you really want your team to be spending more money, right? It's a little bit counterintuitive, but the more tokens they're spending, the more they're actually using it and likely the higher their output is. So, for example, in my business, we have a dashboard to show everyone how much they are spending. We basically want everyone to spend as much as they can. So, yeah, those are five different use cases for Claude. And as you can see, we just have one Zapier connection to basically connect all of my external apps. And what I will say is like these models, they're going to keep changing. There's going to be a new like best one every few months. But the thing that lasts is the systems that you actually build around them. So what I encourage you guys to do is pick one of these, try building it out this week. And hopefully this will save you hours or dozens of hours every single month. And I know I've created many agents within my business has made such a big difference. And hopefully this video was able to help you guys out. In the description, I'll have a link to Zap Your MCP. Also have a link to our free AI community. So I really encourage you guys to check that out. Yeah, hopefully you guys got some value from this video. If you guys did, make sure to hit that like button, subscribe if you want to see more videos just like this. I do a ton of videos about entrepreneurship, AI, finance, all this stuff. Again, thank you so much for your time and I will see you in the next video. Peace.