How I Automated a $30k/Month Business with AI (Skill Systems)

0:00 If you're a service-based business owner like me, your goal is to increase the amount of revenue you make every month without needing to work more hours to generate that revenue. Now, the problem with service-based business owner is that we have a client capacity that we could serve based on the amount of time we have on our calendar. And traditionally, the only way to get more clients is to either work more hours or hire more people to help with client delivery. Now, what if I told you you can increase the amount of clients you serve without needing to increase amount of hours or hire anyone? but with implementing AI into your business. More specifically, you'd implement a skill system that would extract all of your expertise, your intellectual property, and your workflows on how you deliver value to your clients. AI can do that for you. Now, this is completely different from using AI like a chatbot where you need to guide it and do everything from scratch. While a skill system actually does 80% of the work, all you have to do is review the work, sign it off, and send it to your client.

0:55 Now, if you don't know what a skill system is, in my previous video, I explained exactly how you can build a skill system and how it's different from just having normal skills that you would build on Claude. And in this video, I want to walk through a case study of my client Kieran, a fractional CFO, who was spending 40 hours a month on monthly reporting for his clients and got that down all the way down to 10 minutes using a skill system. And I'll break down exactly how Kieran used that skill system in order to scale his business without needing to work more hours. All right. So, the dream of running a business is ideally over time we want our revenue to go up like so, but we want the hours that we work to remain the same or best yet go down. And that's the goal of AI, right? Everyone talks about AI automation. They sell this dream that, hey, we're going to implement this AI automation and your revenue is going to go up and the amount of hours you're going to work is going to go down. Now, very few business owners actually experience this dream because in order to get AI automation, like I always talk about in all the videos on this channel, it's going to require a lot of complexity to eat because AI automation doesn't happen in one weekend and you get to live this dream of oh, you start stacking unlimited amount of revenue, you don't need to work at all. That is something that is not true. In order to get AI automation to work, you need to spend time building a skill system, which we're going to go over in a moment. And if you don't know what a skill system is, in my previous video I went over that. So I highly recommend you to watch that. So what actually happens for a lot of business owners is because they don't know how to implement AI the right way, which is implementing skill systems, is the they might increase the amount of revenue they make like so, but the amount of hours they need to work also increases with them because every client requires their input or they're just using AI like a tool. So if you use AI like a tool, you're just going to work faster. Yes, you might be able to serve more clients, but because you're working more faster, you're going to work more hours, more hours. Now, the perspective shift we want to make here is we want to have work have AI work as an employee, AI employee. And when we work as an AI employee, that means AI is able to be smart enough or we treat AI like an employee. We're going to be meeting it like a human employee. You'd share it with more context. You'd share it with your intellectual property. You'd treat it like someone that learns over time and improves their skills. So as AI gets better and learns more about your business, ideally the amount of hours required for you goes down. Right now in order to get AI to work as an employee, we need to build skill systems. Skill systems. I'm going to do a quick recap over here. Skill systems is a series of skills that chain together that allow AI to run workflows. And as you improve the skills in the chain, that is kind of like treating or improving the training of a human employee in your team. Let's say you you help them improve their um their you improve an SOP. They can do an SOP better. They improve their skill.

3:45 They actually know the difference between a good quality output and a bad quality output. And that takes time. No one gets that right at the first shot. So AI needs to learn that over time. So once we have skill system that allows us to then have AI employees and then that allows us to eventually generate more revenue in our business while working less hours. Now, the big problem is a lot of business owners don't do this.

4:09 They treat AI like a tool or like a product. They think about, hey, I'm just going to work on AI this weekend. I'm going to build this fancy thing or this elaborate sophisticated system and it's going to make me tons of money. Very rarely does that happen. So, meet Kieran. Kieran works with SAS companies as a fractional CFO and he basically sends them monthly reporting and consulting to help them make smart decisions with their financing in order to grow their revenue or to potentially exit their companies. Now, as you can imagine, as a fraction CFO, there's a lot of number crunching, a lot of documentation, reporting, putting things together. It does take a lot of time and effort. And what Kieran was stuck at before was he was spending a lot of hours trying to do all this documentation and stuff. He'd even have an analyst working for him on his team. So Kieran was a good case study because when Kieran came to me, he was working with around seven clients and he was capped because each of his clients would take around 68 hours a month to deliver. Now he was making a good amount of money. He was making around 24k per month all the way up to 30k or in some cases 50k per month with his clients. So he reached a point where he hit his client capacity. So he hit client capacity and some clients actually wanted to work with him and he had to say no. So, can you imagine saying no to clients, saying no to money? That's really hard for a lot of business owners. You'd feel really bad about that. So, in this position, Kieran was actually thinking about hiring another analyst because he already has an analyst on his team. He's thinking about hiring another analyst in order for him to serve more clients so he can scale his business. Now, Kieran, Kieran is not a type of guy who wants to manage a big team. He likes to keep things small, leveraged, and lean. So, that's where I came in and helped Kieran with building a skill system. So after building the skill system, what happened was within two weeks of implementing the skill system, he was able to serve 15 clients. What would take him around 68 hours per client for monthly reporting would take him around 10 minutes. Take him around 10 minutes. Now he can make around 52k per month and he doesn't need to hire anyone because he was able to treat AI like an employee in his business that he would train over time on how he delivers work uh to his clients. So how did Kieran's skill system actually look like? So look like so. So Kieran would then split up a whole workflow into separate skills or separate phases. So step number one for example would be load the configuration for that specific client. So he he'd have a YAML front matter that would show claude where to find the brand colors and the basically the configuration to create that monthly report. Then AI would pull in the data. It would use Python and Power Query query to find the data. Then it would calculate the metrics, arr moves and everything. Then it would validate the data before it moves on to anything. It would look for any anomaly detection or mistakes. Then it would generate the charts. Then it would generate the commentary. Then it would create an HTML or in a PDF after that.

7:02 Then it does a quality assurance review. And then Kieran comes into the picture. So all of this stuff is 80% of the work. And Kieran comes in and does 20% of the work. or this could be even like 90% and he does like 10%. And then the output is a client deck and then a briefing that uh Kieran can share with a client. And by the way, if this is your first time seeing uh infrastructure or skill system look like this, I have another video on my channel. My previous video just posted, check that out to explain to you how skill systems actually work. But this is actually what you want to do.

7:36 And this is what you want to do if you want to get AI to actually generate revenue in your business. You want to break down your workflow into smaller task, create skills and then chain them together into a skill system so it can actually create an output that is 90% quality or more. Now this is what AI automation looks like. AI automation looks like so it is this going through all these skills step by step and creating an output that is 90% quality.

8:02 Now in order to get to this level you have to treat this like an employee again, right? Each of these skills can be improved over time. You're gonna add more edge cases. You're gonna add more guidance. You're gonna add more intellectual property, more mental frameworks to get AI to think more like that you do. Now, potentially Kieran can get to a point where it does 100%. But that again, that would take a lot more time because then AI will need to get the context from these calls. So then Kieran can potentially feed the calls uh to AI to run the full thing uh A to Z.

8:32 But that's something that I wouldn't recommend to do right now. I think maintaining your judgment is really important uh with this stuff because you need to be on top of the strategy. So Kieran knew that the reporting was a bottleneck because he was spending 68 hours on reporting, sending monthly reports. As a CFO, you do have to send reports on the finances and give some insights to his clients so they can make better decisions um in order to grow the business or to ex uh or to exit the business. He was specifically working with SAS companies. So he knew that for months and he he said he would keep fixing it but he just couldn't find the time. So so he was in a difficult position because he was already way too busy delivering the work and he knew if he had to hire someone he has to do more handholding and wait for them to learn and he didn't know how to break it apart. He actually need to figure out how to break up his workflow so that the skill system can execute the work for him. So the first thing we did was when we hopped on a call was we mapped out his business so we can understand how his business works and we can find clarity on how we can get AI to do the work for him in his business and which skill system would be relevant for him.

9:33 So we started by mapping out a business map what I like to call and then we identified that what was his problem was delivery not lead generation. Kieran thought his problem was was lead generation but it was actually delivery. So we then focused on that and that was evident to us because I like to use something called a bow tie funnel which kind of maps out your business from the lead genen side all the way up to client delivery side and we saw that client delivery side was where we could find good AI opportunities to help him scale his business. So what we started was we started with developing a scale system for one specific client at first. So, we started by going over reporting scripts, creating some Python scripts, creating some deck templates, creating the brand files for that client, and then uh putting it all into a trigger called run monthly that um Kieran can run every month for that specific client. So, Kieran worked on that and he got it to around 90% quality to get it get it to work. So, he got it it got to a point where it was he was satisfied with the result to 90%. All he had to do was add the last 10% and ship it out to clients and clients were impressed with that. So with this in place, Kieran was then able to then pretty much copy paste this process, tweak it a little bit for other clients because other other clients were a little bit more custom. So he just needed to to change maybe 20 to 30% of the process. So now he can extend this to all of his other clients. So he had seven clients and he was able to extend it to all of his seven clients like so.

11:00 And that started by creating one full pipeline or skill system built. And this took him around 40 hours of training AI where specifically he would do the report and have AI shadow him while doing it. This allowed him to not only deliver to his current clients without uh firefighting and letting things go while still getting AI to develop uh the skill system for him. And the best thing about a skill system is that for Kieran to onboard a new client to run through that monthly reporting process, it would take him around 10 hours to onboard just to get the API key, set things up. So that in turn would save him 10 hours every single month. So instead of needing to spend like 8 to 10 hours every month doing the monthly reporting himself like so now he just spends 10 hours one time to onboard them and then AI does all the work from here. So this is a skill system doing the work for him. So you can imagine how much more time is he going to get get over time.

11:52 So this compounds, right? He's avoiding spending 96 hours a year on a specific client. So that's a lot of client capacity you can get back. 96 hours back per client is a lot. So it onboarding the client. So spending around 10 hours to onboard the client pays itself over the months so on because he doesn't need to spend as much time. So next thing about the skill system is that uh Kieran didn't let AI do all the work or 100% of the work. he actually let AI do 80 to 90% of the work uh which is most of the repetitive and execution but when it came to his judgment which is identifying the um the diagnosis and giving advice to his clients he would take care of that because that's what the clients appreciate. So AI would build out these reports like so and then Kieran would come in and do the the small uh adjustments or the small uh commentary to guide the client because Kieran has all of the context. You know AI doesn't have all the context. If you are the service delivery uh business owner, you you are on you're the person on the calls. You're the one who understands the emotions. You're the one who understands what the client needs and everything. So AI can help you probably make some commentary, but at the end you want to put in the last 10% because it is an emotional connection or relationship with your clients. So the principle here is AI should do most of the work, but you should do most of the judgment. So you should not let AI do all the judgment. So previously Kieran would take 3 to four hours writing the commentary. AI would do some of it and he would just come in and do the last 30 to 60 minutes to ship things off uh to the clients. So with this in place, generally most service business owners have like three paths in order to scale their business. Number one, they could just do nothing, keep keep the status quo. And for Kieran in this case, he would just stick to serving seven clients and he would just continue working uh normally. Number two, he could hire someone. So for Kieran would be like to add an analyst, but that would add training, review, payroll, and that would just add more stress and handholding that he will need to do. Or third path is you could build a skill system or an AI employee. AI employee and you can increase the client capacity you have like Kieran did. He increased it from 7 to 15 clients and you don't need to hire anyone and it takes and it works on a $200 a month AI subscription on cloud code as an example. I don't know about you but I would definitely pick the skill system option because the gains here are really high and the cost isn't that much. you know, you're paying $200 a month and you might spend a little bit time upfront to learn how to use cloud code and how to use how to create a skill system, but that skill will pay you really well over the long term because not only would you create one skill system for this, you can create skill system for other parts of your business. Lead generation, sales, delivery, everything. And skill systems stack up on each other. So that's a goal for me as a business owner and that's what I do with I help with my clients.

14:44 So, if your goal is something similar to this where you want to build skill systems in your business and you want to have AI employees that do 80% of the work for you in your business, I am launching a community next week on May 18. It's called the chief leverage officer community. It's called that because as chief leverage officers, our goal is to grow our business based on leverage. That means we look to implement digital assets systems, skill systems, and AI employees so we can scale a business without working more hours. Okay? And if you want that, I'm going to put more information in the description below. So, next thing about Kieran is that his calls with his clients became more productive and became more valuable to the client.

15:20 Before he would recap numbers, he would read from the deck and he would tell him, "Hey, this is what happened to numbers and so on." Now, what Kieran does instead is because AI does most of the number work and everything, Kieran just records a loom video. A loom video. He he does what he would do before on the call. He sends a loom video and then on the call he focuses on things that that is decision-m stuff, things that requires judgment. So the client watches a loom. The call is more strategic conversation. You know, Kieran's more energized because he's not doing a boring call explain numbers, but he can actually talk about decision-m and strategy. And clients appreciate that more because now they're spending their time on stuff that they actually value more. So they actually talk about more important discussions. And again, if you want to build skill systems in your business so you can get 80% of the work done done by AI, I do have a community.

16:08 I'm launching on May 18.