How I Use Claude to Run My 7-Figure Agency

Here's how I run my 7-figure agency using Claude. It tracks all my data, made my team twice as fast, builds my internal tools, and even launches my ad campaigns for me. And let me show you all 4. All right, so let's get into it. Now, this is for educational purposes only. This is just kind of like me documenting how I personally use AI. So please just use it with caution. All right. So the number one thing we use it is for data tracking. Running a business at any level will require you to track data. And the bigger your business grows, the more important it is to have accurate and documented data on what's going inside stuff like operations, sales, marketing, finance, all of those things. And using a tool like Claude and uploading all of your data, whether it's on a CRM or Google Sheets, like makes the process a lot more faster. And this is very true for us because let's say, and I don't know if this is kind of like a flaw in our business or is this like it's not necessarily a flaw anymore because AI pretty much consolidates all the information for us because we track a lot of our data on different sources, whether it's Google Sheets, ClickUp, Airtable, because as we were growing, we just added more sources and we never necessarily consolidated into one specific platform because it didn't necessarily make sense because all of them are departmentalized. But using a tool like Claude or any other tool where we can just collect all of our sources, like use API for ClickUp, use API for Airtable, connect Google Sheets to it, it can just kind of like scan through all those sources and just answer questions for us. Like, hey, how was the churn this month? Hey, what is the status on this specific client? Who signed this specific client? So it can just pull up information for all types of different sources. Now, another thing before we kind of like get too deep into this, I want to understand how we use AI as well and kind of like what I tell my team on the mindset you need to use AI on. Because some people I've spoken to, they use AI in a very 1 or 0 mindset. Either it's 0% done or it's 100% done. Like they just want the entire thing done for them. They don't necessarily want to have any human input. So what— how we use it is, let's say a task is 100%. We don't want AI to take over the 100% of a task. We only want to use AI for 30 to 80% of it because we still want the human touch, especially as sensitive as data to be there. So we can either verify the data, we make sure that, you know, it's properly allocating it from the right resources. So we still need to do some sort of quality check. And what this allows us to do is, let's say if something takes us 2 hours to do, okay, with the help of AI, can you reduce that to at least 1 hour, which is 50% of your time back? And this allows you to be a lot more efficient. Because if you do it this way, like AI is not necessarily going to just take over and, you know, just do whatever it wants. And sometimes it You know, just puts in inaccurate data. Just happens is how AI is nowadays at this point. And again, like I have added some stuff like are there some tasks that can be done 100%? Sure, but data should never be one of those things. So you need to keep this in mind and whatever you do, however you implement AI, it's extremely important and efficient to keep this in play. I'll give you another example. So for example, for some of our funnels, especially recruitment, we like to use a conversation bot. Now what a conversation bot does effectively is as soon as a new recruitment comes, like The conversation bot is going to take over and just talk to that specific person. But at one point, we've added some triggers where it cannot answer a question. Like, it needs to tag in a human. It can't just force itself to carry the conversation if it doesn't necessarily have the answer. Now, can we train it? Of course. Like, we have trained it. We trained it as much as possible. But we still want the human element in there to be there so it doesn't necessarily sound like, you know, this person is running in circles and getting the same response. So that's just kind of like one example. Okay, perfect. So it is a must to have an eye on data. Like without data, it becomes hard to make any decisions and grow predictably. And this becomes more important as you grow. Also, like if you don't necessarily have accurate data, you are just not coachable. So let's say if you're joining a coaching program, if you have kind of like a coach, a consultant, whatever it is, if you don't have accurate data, whether it's on operations, sales, marketing, fulfillment, like nobody can help you. And this is something you have to kind of like get in program with is without proper numbers, you are not coachable because like, what do you want the coach to do for you? Like, just think about it. Like, do you want them to just kind of give you airy-fairy answers? Like, you need to have all of your numbers properly documented and I can kind of like help you do that in a much more efficient manner. So what do you need to do to kind of like prepare this? And this also depends on the business, but I have kind of like certain metrics that you can track pretty much right after this, right after this video. And you can of course access to this document as well. So in operations, you need to— for let's say if you have kind of like a, you know, client who joins in your program, You can do start date, end date, package purchased, locations, progress status, sold by, deliverables, churn, churn reason, additional notes. So things like these, like it's useful for AI when the list builds up. So for example, let's say you have kind of like, you know, 3 people churning on a month-to-month basis. On the 6th month, you know, when you have around 18 people churn, you can actually look into the churn reason if you've been documenting this properly on like, hey, what are my, what are my churn reasons? Like why do people churn the most? So AI can pretty much kind of like scan through all of that and give you kind of like your top 3 reasons. So that's just one of those things where as soon as your list of customers goes up and of course the churn builds up as well, which hopefully doesn't, but every business has churn, it can pretty much give you feedback on what exactly is going on and why people are churning. So this is just one of those additional useful things that you can tap into to kind of like reduce churn because that's the reason we're tracking it in the first place. So the second one would be, of course, sales. So stuff like this should already be documented within your CRM, stuff like lead name, company, date booked, date of call, booking source, which funnel event or setter booked it, who was it set by, who was it closed by, what was the call result, package pitched versus package sold, cash collected versus total contract value, any objections/reasons not closed. So this could be a kind of like another one. This could go into the notes section and then of course a follow-up date. So this lets you see the show rate, the close rate, and revenue per closer at a glance. So you can literally just, you know, as long as you're tracking this properly, you can connect all of this into AI and be like, hey, so what are this closer stats? Which package are we selling the most? What are the most common objections we're getting? All these type of things and also kind of like what is my, what's it called, sales cycle. All these type of things it allows you to pretty much tap into at any given time. For marketing, you can track stuff like, you know, ad spend, impressions/views, leads generated, cost per book call, cost per lead, calls booked, show rate by channel, which content/ad the lead came from, your cost per acquisition. Now, of course, you can use a tool like Hyros to pretty much do this for you. However, if you are already tracking this on a much more manual basis, let's say if you don't necessarily want to spend for whatever reason on a tool like Hyros, this would allow you to pretty much tap into your marketing metrics at one glance. Like, hey, what was my cost per call this week? What was my cost per call within the next, you know, like what's the trend? What has the trend been over the last 3 months? So all these things is something you pretty much get for marketing. Now on the finance side of things, this could be, you know, monthly revenue, recurring versus one-time revenue, payroll, contractor costs, ad spend, software tools, refunds and failed payments, gross margin per client, outstanding payments receivables, profit, and overpay. So all these type of things, of course, you know, again, connected to AI, and you can get a lot more accurate information and also do projections as well on how the next month is going to look like. Because if you're not tracking all of this, then, you know, just having things a little bit more predictable for your business becomes a lot more difficult. And then common denominators. So these are fields that repeat across everything. So things like client name/company, contact info, email, phone, Date, source, owner, package offer, status. So the common denominators would pretty much help you connect all of these data because let's say if you're, you know, doing sheets or you're doing sales on a Google Sheets, you're doing marketing on a Google Sheets, you're doing operations on ClickUp, like it would just allow the AI to pretty much kind of like read through like who is this lead, when did they close, where are they from, who owns it, and what did they actually buy. So every record and every sheet should pretty much answer all of this. So why it matters is because let's say if John Smith is in your Google Sheet or in your sales sheet and J. Smith in operations and Smith Co. in finance, like no tool can connect them like human or AI. But if the same client name, source, and package appear in all 4, meaning the similar name at the very least, you can just ask questions like, what's my churn rate on clients for Meta ads? Or which closers' clients stay the longest and actually getting answers. So the common denominators are what turn 4 separate lists into one business you can actually see. So having this information in AI can help you get answers fast rather than, you know, just going through every single data source all the time. So which you should anyways, but this pretty much makes it less frequent and gives you specific answers. And I've also written down some tools you can use for this. So the tools we use are, for example, ClickUp and Airtable to track pretty much almost everything. Marketing, you would need to, of course, track some stuff manually. Also using a tool like Hyros, like I said, can help. And it doesn't necessarily need to be Hyros. I'm pretty sure there are other tools in the market. We've just been using Hyros, but Yeah, you can use of course Hyros as the API can now also directly connect as well with Claude. And then of course you can use QuickBooks for finance or if you use Mercury Bank, I'm pretty sure they have a lot of things. But these are kind of like the tools we use and we recommend. Cool. The next one. Now the next part is going to be team access to AI. And literally the second line says spend the money. So at a base level, what you can do is whatever Claude subscription you have, GPT subscription you have, you can just get a team account and then add people who you think may be able to use this specific software. Now it also depends on the business. Business. And of course, what you guys do, I can give you my example as a content agency. So for example, let's say scriptwriting and research takes us a lot of time. Like, of course, that sometimes takes us days, sometimes even weeks to kind of like get every single thing processed and make sure everything is ready to go for one specific client. So we use a tool called Poppy, which also has, of course, Claude built into it. And we'll upload every single thing that we know about that client, whether it's a website, whether it's a sales call, onboarding call, onboarding document, just so we can always maintain relevant context. Per team member. So the account manager, they're just going to plug all the information in. They can just ask it questions and then just start the research phase way more efficiently than they would before, where they had to manually kind of like watch the call, go through the website. Like this makes their process a lot more faster. So now before it might take days or weeks, now it literally takes them a couple of hours and they can work with more clients as well, which also increases their earning capacity because they're able to do more work more efficiently. Now, if you need to figure out kind of like what your team even needs to do, you can actually do that by just helping them write down all of their tasks their team does within kind of a 72-hour time audit. So within like literally 3 days, meaning they write down every single thing they do in working hours for 3 days. So this is where you can pretty much kind of like see where they're spending the most of their time on. And if you kind of like notice that, okay, they're spending 4 hours on, you know, just calculating XYZ, you can ask yourself, okay, how do I turn these 4 hours into 2 hours? Or how do I kind of like completely eliminate this? Which of course we don't want to, but I'm just giving you an example. How do I turn this into maybe 1 hour so they can have more of their time back and maybe they can handle more projects and earn more? So these are some of the questions I would ask and actually help your team do so you can have more insight of what's going on and what they actually need help with. Of course, asking them is another way, but 72-hour time audits by your entire team usually kind of like gives you a lot more answers. And also you see what's taking the most time in their task and just try reducing that time using AI. Like don't necessarily force replace the entire task by using AI, but like People always think in ones and zeros like we talked about. And again, like it's very important that we understand the goal is to reduce the task by 20 to 50%, which allows your team to be more efficient, which means they can take on more work/projects and earn more. You don't necessarily have to increase staff cost. So by definition, if you just spend a little bit more money on team access to AI and depends on depending on how big your team is, it might not even cost that much. But it also you have to look into a way where you don't necessarily have to hire more people to kind of like do that new work for you as well. You can just give it to your current team and your current team is going to be more happy because they're earning more too. So super, super important. And also look into your current softwares. They will have some AI features that you can most likely look into and kind of like save more time too. An AI software I recommend for businesses is— and if this is, of course, if you use Slack and Teams, if you don't, I don't necessarily think it will work— it's called Victor.ai. Like I have a whole video dedicated to this specific software, which you can pretty much kind of click on the screen. And it's literally like hiring an operations person. Like don't necessarily use it as an ops replacement. Use it to make your current ops more efficient. And if you haven't noticed by then, like we're not necessarily replacing anything. Like we're making things more efficient and we're just trying to increase profit margins because that's how you need to be using AI. Because people, like I said, they think in ones and zeros. Either we just completely replace this position or it doesn't work at all. Like if you look into it in a mindset of like, okay, how do I decrease the time it takes to do an X, Y, Z task? Like, and you keep stacking those up, then you'll open up more capacity per team. You'll have— you'll save more, what's it called, costs on hiring new team members, training them, you know, spending more time on doing that. You can just increase your profit margins if you just are more efficient on the way you use AI. Agentic workflows. Now this one's a tiny bit advanced as it does require some additional tools and also sometimes even cloud code. However, learning to automate the most repetitive task, and this is related to the same things we talked about earlier with the 72-hour time audit, like you can literally build agents so you can speed up a specific workflow. However, learning how to automate the most repetitive task, and this is kind of like the same thing we talked about earlier with the 72-hour time audit, you can literally build agents around some solutions to speed up the workflow. Now, of course, we know we've talked about some softwares, we've talked about cloud, how to use it, but sometimes you need a custom solution for a specific problem. And if you have that knowledge and if you pretty much know how to do it, you can pretty much do wonders with it. An example I have is in sales, for example, we use a CRM called Close and we wanted to implement an iMessage service. Now, this iMessage service is on a software called Link, which was recommended to us. But the problem we faced was Link doesn't necessarily integrate with Close directly. So we need multiple tools talking to each other. And the problem is Lync doesn't necessarily integrate with Close directly and we need them to talk to each other because it would just add in another platform for my sales team to use. And I don't necessarily want to add in more tools for them to use because the more stuff you have, the more stuff they have to do, the easier it becomes to forget. So I want to consolidate every single thing into one software. Now, as long as each platform gives us API access, meaning we can actually do stuff outside of their software, we can use something like Cloud Code to connect everything together and configure the workflow exactly how we want it. And now it works extremely smooth because as soon as someone books a call, an iMessage automatically goes out to them. And every message sent and received through Link also shows up directly inside of Close. So our sales team can actually see the entire conversation without leaving the CRM. And not only that, we've also built something custom where they can literally just respond to text messages inside of Close. So this is where as soon as you click on someone's profile card on the sales side of things, you can see the entire conversation over there. And they can also use the software as kind of like a reminder to respond to text. So everything is now pretty much connected. And I have kind of like examples over here on, you know, this is Close software where the message is going and incoming. And this is the actual chat bubble, which is the LingQ software. So access to tools and knowing how to use them unlocks a new set of possibilities for your business. So you have to use them. You can quite literally ask Claude on how to build it. You just need to be patient and know how to follow instructions. That's literally what you need to do. Another thing is, and this is of course through the realms of vibe coding and building softwares. And listen, like I've wasted so much time on creating some useless tools. So you really have to plan out on what you're going to be making. And below are some examples that I've spent hours and sometimes even days creating, which we don't even use. And you can see there's— this is just kind of like a handful of examples. There's a lot more. This is Command Center, this is Sales Dashboard, Metrics Dashboard. This is, you know, this custom Workflow Hub, Sales AI. Like all of these are just garbage. Like we don't use them. I don't think we're going to use them. The Command Center is something we're still working on, but Like these ones we don't necessarily use because the sales dashboard like this is already being done in Airtable and it's very redundant because the information it was tracking was not super accurate. This metrics dashboard again super useless. It was already done being done in Airtable, so it was just kind of like you know me fixing a problem which doesn't exist. However, there are some tools that did work and we use it pretty much on a very consistent basis. So for example, we have a thing called an internal script review tool and we're also building out a potential replacement for Poppy, which is kind of like the software I showed you. You guys early on. So this one is something that's required to be used by my team. And what it basically does is I've trained a bot on all of our successful scripts, what we look for, and what errors to pretty much prevent. And as soon as somebody uploads a script on this specific software, it can pretty much analyze the entire thing and give them feedback right off the spot. Instead of waiting for another person or a QA checker who sometimes are busy, they can just upload it to the software, which is already built on their specific knowledge. So this way we can kind of like speed up the workflow on getting the scripts approved and sending them out. And another thing we built is kind of like our custom-made training portal. Now we could have used something like Kajabi or even Skool, but we wanted some custom— like this was an actual scenario where we wanted stuff to be a little custom for us. Because let's say if we're doing a lot of new hires, I've made kind of like a set of course videos for a specific department where they have to watch all of those set of videos and actually quiz them on it as well. And if they don't pass the quiz, they don't get the job. But if they do pass it, then they have a chance. And then of course, this is just one of those things where we're able to code in specifically, how many tries we give, are they actually watching the video, are they being lazy, how much time is it taking them to actually complete the course. So it gives us a lot of insight on what type of person we're hiring as well. And now we're doing a lot of things with it. We're also involving into an SOP library. We're taking all the videos that we've already made, transcriptorizing them, if that's a word, and then creating an AI where they can actually help them just answer the questions. So for example, we want this platform to be used as a thing where our current team members can visit it again Maybe use the AI or kind of like use the SOP library, maybe watch a video again, kind of like see all the tools that they have access to. So it's kind of like becoming a hub for my team as well. Now, last but not least is Claude Cowork to build the campaigns. Now, I know in this screen I've kind of like shown an example of Manus. Claude Cowork can do the same. And I pretty much do this with Claude Cowork all the time as well. Now, these are one of those things where you can literally ask Manus or even Claude Cowork to take over your screen as long as you have the plugin installed. Like I use Google Chrome, you can literally install a plugin to do this for you and do this on your own accord, by the way. To give a disclaimer. And literally, you can tell it like, hey, can you just, you know, go to my Google Ad account and launch me a Google Ad? Or if you've already set up a Google Ad account, Google Ad account, and you tried setting up a Google Ad yourself and it's not necessarily as good, you can just go to Cowork and also Manus and be like, hey, can you just help me build this properly? And then another thing you can do to pretty much take this level above is you can teach them everything there is to know about Google Ads first. You take all the YouTube videos, You take all of kind of like some popular articles, blogs, whatever there is, and you can train it first and then ask it to take over your screen and just use the best practices on either Google Ads or, you know, any type of text-based ads, which are a little bit more simpler to set up. I personally wouldn't necessarily use this on Meta ads. Of course, do this on your own accord, but with Meta ads, it's more creative heavy. So you need to, of course, make sure that, you know, the videos and stuff you're creating are a lot more better rather than just kind of like some text and imagery. So that is pretty much it for today's video. So these are kind of like 4 ways I use Claude to run my business. All of these are something I think you can just tap into right off the bat after this video. The number one thing I would put emphasis on is data tracking, because with data tracking, it literally makes our life a lot more simpler to the point where I do have some automated tasks as well. Let's say on the CS department, we have 3 sheets that we have to mix together and put into a master sheet. ManageThis does that for me. So every Sunday it'll trigger where it'll go through every CS sheet It'll calculate all the totals, whether it's upsells, revenue, churn, whatever it is. It'll just calculate all 3 of my CS people's numbers and put it into the master sheet and add in notes on exactly who did what. So that level of stuff and that level of data tracking is something you pretty much build with AI if you kind of use it correctly. So yeah, I mean, if you have any questions or if you have anything you need specifically help on, just make sure to comment down below and I'll make sure to respond back to you as well. Other than that, I hope you enjoyed this video and I'll catch you in the next one.