How to build Y Combinator’s "AI Agency" (Step-by-Step)

So last week we released a video about Y Combinator begging entrepreneurs to start this kind of business. Now this is called an AI native agency. Think of a marketing agency, social media marketing, web design agency. All of that sort of wrapped into a new model powered by AI. Now this video blew up and this was the high level of what is this? Like why are they telling you to do it? If you haven't seen it, go watch that video. However, in this video, I'm going to show you step by step exactly how to do this. Not from a high level, not from a concept, not from a why commentator said, but from a hey, here's how to watch this video and by the end of it, go do the thing. Now, just for context, we've been doing this for over the last year. And we have students doing between $100,000 profit a month all the way up to a million dollar profit a month. And I know it sounds crazy, but I selected, and you can see by the member count, I selected three random screenshots to show you proof of this across different time periods over the last year. So, we have Agus doing $93,000. Max baits $109,000. And then this was like two days ago. Uh Tyler, two deals, okay, two clients, 30% and 25% rev share. That could be like anywhere between 20 to $50,000 a month for him probably. Now that you know this real, what we're know we're talking about, let's dive in. So, why is Cominator a big firm that invests a lot of capital talking about this business? Well, there's a few things. The old agency, just for context, is a note efficient model. Now, the problem with the old agency was as you scaled, your profit dropped because it wasn't efficient because you had what's called human capital. Now, that's not that great of a business to run if your goal is to cash flow and make a ton of money. Some agencies sell, but it's not really a sellable asset like software is. So, in venture capital's view, in the past, agencies were bad because they're unscalable, high overhead, and low margin. But the new model which leverages AI and removes all of the problems to be honest. Delivery is no longer a bottleneck. Margins look like software because you're running kind of a software model. Revenue scales without the headache and suddenly you are venture backable. Now I'm not telling you guys to go build this to be venture backable. I actually think you shouldn't do that but you should start this business. We've been saying this for the last year. So here's exactly the page they released talking about this. This is their actual article. So for example, now instead of selling software to customers to help them do the work, you can charge way more by using the software yourself and selling them on the finished product at 100 times price. That is exactly what we've been doing and what we've been preaching for a long time. And they even said in their own words, agencies have always been hard to scale, but AI changes this. Instead of selling software, you can build an agency. Now, in my opinion, I run a software company. I run an AI software company. It is called KendoAI. We are currently at a very good revenue number and we're growing quickly and shipping a lot of new features and we're fully bootstrapped. However, it's hard. I've burnt a lot of money, hundreds of thousands of dollars just to get off the ground. And if I didn't have the business acumen that I had, we probably would have failed. So, my recommendation is you start a agency. You start a service-based business. We call this AI operating or AI freelancing if you're a beginner. I don't care what you call it, but you need to start a business where you're arbitrageing on the software tools that we have access to. and the AI tools. For example, a Claude Max plan is $200 a month. I can deliver $100,000 worth of services with that one $200 a month plan. That is the greatest arbitrage of all time. And that's what this model really is at the end of the day. Okay? And why combinator and all these big companies know this. And so what is an AI operator exactly? It is someone who simply delivers a service to a client faster using AI either faster, better, more efficient, more profitable, whatever uh metric you want to throw in. It is not an AI consultant. You're not giving advice. You're not a developer. You're not a dev. You're not building legitimate software. You're not an automation agency. This is a commodity. And you're not a traditional agency where you trade hours for dollars. Okay? It is, however, someone who uses AI as leverage to deliver high value services to business. Okay? So, you're paid on results, not hours. You're a business operator who happens to use AI. You're not a technical person building AI software and all these sort of things. And you focus on outcomes, not technology. because this is the problem that a lot of people run into. And again, just for context, more quick wins. We've been doing this for a long time. We have about a 100 people in AI insiders now. Um, all absolutely crushing. So, the three pillars of business haven't changed. This is where everyone's getting caught up and I see a lot of bad information out there right now. It's still sales marketing systems, okay? Sales system, sales process, sales teams, you got to close deals. It's still marketing, content creation, ad campaigns, funnels, copyrightiting, lead genen, whatever falls under that. and then how to do those two things efficiently and run the business. Workflows, automation, CRM buildouts, data management. AI doesn't change what business need. It just makes delivering and achieving those outcomes more efficient. That's our whole model. That's our whole hypothesis and thesis around this business. So wrong would be, oh, I sell AI automation. It's a commodity. I can replace you off of

Fiverr with somebody in the Philippines. You're competing on price and it's a race to the bottom. The right way would be, hey, I can increase your revenue. I can make you 50 ad creatives a week. I can do XYZ. Insert whatever outcome you want based on whatever business you're interested in and do the work with AI. And think of it like this. Think of the cool tools like Claude, Claude Code, Claude Co-work, OpenClaw even right now. Uh Seed Dance just came out. It is the most insane video model I've ever seen in my life. Nano Banana is insane. Uh image generation. You can make ad creatives, you can make the ad copy, you can do the copywriting, funnel, you can do the web generation, you do literally all of this with AI in like 1% of the time it used to take. Meaning, you can do more work, you're more profitable, and you achieve a better result because the skill of what you're delivering is not really in your hands. It's more up to the LLM and the AI tools. So, you're selling the problem that you solve, not the tool that you use. For example, in the past, I have paid a lot of money for VSSLs. What's a VSSL? It's a video sales letter VSSL for AI insiders. If you watch this video, this is basically an hour and a half argument on why you should start this business. It's kind of like what I'm making right now, except for this YouTube video is educational. This is more to get people to book a call and to join AI Insiders because we have a great program. This whole video I scripted out. — I can cut your lead. I'm going to cover all those — with AI. Okay? I scripted it out. I give it to my design team. They made slides, which we now do AI for the slides. But this whole thing was written by AI. In the past, that would have cost me seven to $20,000. Like realistically take a long time. It's extensive. There's a lot of reviews. It's a lot of time on my end. There's multiple revisions. It's tough. Now, the AI operator way, especially with all the context profiles and the scripts and the templates that we give people, so you can literally just plug and play. You can do this in like an hour or a day or a couple hours just depending how efficient you are. Okay. So context profiles, you use cloud with the right instructions. You have the right copywriting knowledge base. You do it in two or three hours. You can do like five of these a day if you really wanted to. And your quality will rival that of a $20,000 project copywriter. This is ultimate leverage. And this is the example I use because it's a tangible thing you can see and that we use. But you can do this across everything, guys. So how do you start an AI operating business step by step? Well, let's dive in because there's a lot here. I told you guys this is not going to be a highle video. So number one, you got to pick your niche. Okay, so who do you want to work with? And there's three criteria that I really look for. So number one, do they spend a lot of money on sales marketing or systems? Yes. Okay. Number two, are they capable of generating at least another $100,000 a year with better systems and insert systems here for sales marketing systems, not just like back-end systems. Basically, if I do X for them, will they make more money? Because we want profit share. We want percentages. We don't want retainer. We want upside in the outcome that we achieve because we achieve good outcomes. Aka, we're good at what we do. And number three, do you have easy access to reach the decision maker? Okay? If I'm trying to sell to the government, I can't reach the decision maker very easily. It's not a good offer. niche. It's not something I should be doing. Now, and I'll skip through these just for uh sake of time here. Info products are great. I know info I kind of hate the industry myself. Um, whatever that that's neither here nor there, but they spend a lot of money. They're highly profitable. They grow really fast and if you do the right things with them, you can really blow them up. Okay, ecom brands, there's a lot of different things that they need. Ad creatives, backend stuff, websites, web work, you know, graphic design, whatever. They have a lot of problems that if you can fix, they make a lot more money. Now, SAS companies, I myself with our SAS company, we have backend, we have marketing, we have sales, we have systems, we have ad creatives, we have funnels, we have VSSLs, we have a lot of things similar to info, but the pricing structure and the LTV is a lot different. So I think SAS is for someone that can do what info does but a bit more advanced. I think that's really good for you. Agencies, you can help agencies do the things they need to do. So if the same problems that they're solving, you could be the freelancer for the agency. contractor a partner for the agency and really do uh any of these things, sales marketing systems. Now, how do you find these people? Well, here's an easy way. There's more advanced ways, but we can't give away our secret sauce uh for free on YouTube. So on Instagram, this is what I call the spiderweb method. So if you go to like my profile and you click this plus icon, you'll actually see all the similar profiles. And so again, Instagram's smart. So who is similar to me? Eddie, big influencer. Uh Fredo, big influencer, hangs out with Brez and all those guys. Um Jordan, obviously big influencer. TJR, big influencer. These are similar accounts to mine. And so you can just repeat this process and you'll map out the entire industry in like a day. You just hit Jordan, hit plus, hit the next guy, hit plus, etc. Uh, Twitter, Twitter is really good in my opinion because you have people posting more insightful, more brainheavy topics. So, you know, oh, this guy's like smart. He's like doing XYZ in business, whatever. Um, LinkedIn, I don't like

LinkedIn personally as a platform, but it's good for business. And then you have things like email. If you're going to do cold email, that's a whole other video to be honest on how to set that up properly to get past spam. but things like Apollo list kit, my boy uh Daniel's company, um building targeting list, etc. So, this video is not a whole outreach master class. We have those on this channel if you want to go watch, but this is from a high level how to do that. Now, number two is how do we choose our service? Now that we know who we want to sell to, you really just pick one of those. What do we do? So, the core principle here is don't be a jack of all trades, okay? master one thing completely because the AI allows you to be good at everything in a sense, but we want to be specialists. Specialization wins and that's how you're going to make the most money. A lot of your services will overlap, but your main focus should still be one thing. Okay? So, for example, if I run ads, I'm probably going to have to mess with the sales team a little bit or the funnels, right? It's just common sense. Now, you want something because of AI that runs off of the software model, which is repeatable equals profitable. Okay? Kendo, my software company is the same thing every time. like it doesn't change. We customize it to the business, but it's the same product. That makes it easy to sell because I can market the same. I can sell the same. My systems are the same. It's repeatable, meaning I can scale really, really fast and aggressively. Okay. Now, some things that you can do, and this isn't the end all be all. This is just a starting point. Number one, creative services. Okay. Add creatives, graphic design, like content, landing pages, things like that. Um, again, I'm not going to read all this for sake of not being an hour long video, but this is a really, really good one. I need these myself even. Don't pitch me. I do this myself with AI. Number two, copywriting. Everyone needs copy via sales, emails, landing page, ad uh sales scripts, webinar scripts, etc. Like the one I showed you before, the example there. Um, number three, sales services. A lot of our guys do this because we just come from a sales background. So, sales management, sales process optimization. This is like what Kendo solves in a software functionality, but you can still do it manually. uh CRM buildouts, sales training. Like my COO, the head coach in our um our AI community, he has sold CRM buildouts for like 30 grand a pop. Um so it definitely works. Number four, marketing ops. So this is broad, but like content creation, ads, funnels, uh lead genen, traffic strategy, you know, media buying, a lot of businesses need this, especially info, ecom, um and SAS. And there's a big upside there. And then number five, you have systems and automation. Okay. So again, workflow automation. This is not an automation agency. These are just services that you can do and you can add that are pretty efficient. Uh CRM setup, custom CRM. You can vibe code these tech stack integration. Again, a lot of that. Uh data management, process, documentation if you're a pretty organized sort of person. Now, those are just examples. Number three is your offer. Okay? So niche plus service equals offer. So you're not going to say this like don't be the hormosy guys that are like I do X and I do like it's not going to work. Okay, this is for yourself to get clarity in what you're selling. So I fix specific problem for a specific business type using spec specific mechanism which typically gets specific outcome. Okay. So, for example, if I were doing this for Kendo, I ramp up and train sales teams more efficiently with this AI tool uh for sales teams basically like we work with a lot of different niches, but uh sales teams just to keep it broad. Okay. So, it's very simple. It's just for you to kind of understand. Number one in breaking this down is a specific problem. Sales issues isn't specific enough. So for us, it's ramp time, training, and then management time. Like we solve those with Kindo, okay? Not marketing problems, but we lower your CAC. Okay? We uh we make your cost per call drop. We're more efficient. Like you want to get granular. Okay? Number two, specific business. You don't work with businesses. You work with info product businesses running paid traffic, right? You don't work with e-commerce stores. You work with DTOC brands doing 5 million annually. Like you got to be granular. Next is the mechanism. You don't want to say using AI unless it's a legitimate like AI product like ours. Using psychologydriven via cells, right? Using proprietary sales processes, using AI powered creative testing, whatever you want to call it. And then number four, specific outcome. You don't want better results. You want we add $300 to $900,000 a year in revenue. Pretty specific. Why? Because you know the math and you know the equation that the typical first three steps you work with usually hit. Like business guys is a game of patterns. It's a game of math. Um and once you do this enough times it really becomes repeatable even at scale like this. Okay. Um typically not more leads you want cut legion cost by 40 50%. So for example and you can steal this if you want. Um offer number one I fix underperforming VSSLs for online coaches using a psychologydriven VSSL rebuild system which increases revenue by 30 to 60% without more ad spend. Now would I run this as my headline? No, you

want a cleaner headline, but this is for you to get granular and like really understand what you're doing. Okay? And there's more examples here as well. Now, the reason this works is because it's specific. You're targeting a specific person. You're using a specific mechanism and your outcome is specific and it is a measurable outcome. It is not a I'll get you more results. I'll make you more money. I'll get you more clients. Those don't work anymore. You have to be granular. So, like if I'm going to do uh ad creatives for ecom, right? Hey, we're going to do 50 highly converting, fully customtargeted based off your ICP and your actual ad data creatives for you every single week. How am I going to do that? With AI. Duh. But like, does that sound really cool and valuable and outcome based? Absolutely. Okay. So, that's the thought process that we want to achieve there. Now, number four is your pricing strategy. Now, there is multiple pricing strategies. Okay? You can price things a lot of different ways, but there's one golden way that we want to do, but you can start and get there with whatever you want. So, you can do one-time setups. This is when there's a clear beginning and end. Okay? So, this is like a contract project based work. So, a CRM buildout, a VSSL buildout, a funnel buildout, whatever. Number two, monthly retainer. So, ongoing deliverables. So, for example, the ad creative thing I could do a retainer for. Like, that's that makes sense, right? 12 ad creatives a month or 50 or whatever emails a month, content a month, ad management a month, right? Then revenue share. This is the golden ticket that you want to try to get to eventually. And this is how our students are making like insane money. Um, but this is when you directly impact revenue. It doesn't make sense to take revshare for making creatives. Doesn't make sense. But if I'm like managing your sales team or if I'm running your full ads and webinars and marketing in VSSL, yeah, then I can do that. Okay. And so your setup fee you can still charge and then ongoing you get a percent. So last year or however many years it's been we did a management uh of a sales team and we did a big launch for a very well-known influencer that just happened to streak the Super Bowl. Okay. Um we charged a setup fee. I think it was $25,000. We got paid to build out the system and then we got a percentage and I'm not going to say like his numbers and our numbers. We did multiple millions in a week managed by our sales team. and we hired 60 sales reps and we made a couple hundred,000 profit from that. Okay, because we made millions upside, we took a percentage. It's very simple. It's the best business relationship of all time because you kind of eat what you kill. Um it's not crazy upfront for the business and you're both aligned incentively of, hey, I'm going to make you more money. If I make you more money, I want more money. Okay, this is the golden ticket of what you guys should be uh optimizing towards. Now, number five, how do you land clients? There's two ways. warm outreach and cold outreach. Warm is who? Your past clients, your friends in business. If you're a sales rep right now, you could start this business overnight and probably get clients, right? You have a network. Um network people you worked with, industry contacts, etc. Now, the approach here is not do you need AI services. It's hey like hey bro, I've been following you for a while or hey like I know we worked on this past offer. I'm doing this now. I noticed you have X problem. Um I think I could solve it. I could achieve whatever outcome. Again, back on the previous steps we went. Um, is it cool if we hop on a call? I take you to coffee? Can I send you a video breaking down how we could do that right now? For your first one, do it for free. Like, I'm not going to be the guy that's going to tell you guys you got to charge for everything. If you're not good at something and you never done it, you probably got to do it for free. Even in Kindo, we still for certain key people, we'll still like wave setup fees and stuff for them just because, hey, not now because we're established, but like in the past, hey, like this is new, this isn't proven, I'll just do this for you, and if you like it, then you pay for it, right? Like build a case study, build a testimonial, uh build a portfolio, and use those for cold outreach. Once you do this a couple times, if you have warm, then you can do this. If not, you just start with this. So number one is build a list. Lists and CRM and actually targeting people is good. You want to be organized. efficient. So again the Instagram thing, LinkedIn sales nav, Apollo or Listkit and Twitter search. And then you're going to throw them. You can do it with AI. You could have Manis do this to be honest. You could have open Call do this. You don't have to do it manually. You're going to throw them in a list and you're going to target these people. And you want to make sure it's the right industry. you have the decision maker and then they have the pain signal or the outcomes from like what deems the three steps of criteria that we went over previously. Then you're going to do research. You're going to figure out like what is going on in their business right now that I can see from the outside and I'm going to compile that into a value first outreach. So I'm going to do some work up front and I'm going to make a VSSL. I'm going to make ad creatives. I'm going to make them something. Now in the past this wasn't efficient at scale because it took so much time. But now I can write a mini three-page VSSL for somebody in like 10

minutes, send it to them, they're going to be wildly impressed and they want me to do a legit one for them. Or the same concept with ad creatives with, you know, like vibe coding or workflows or whatever. Like value first wins because everyone in their mom is doing outreach now. You need to stand out and you need to say, "Hey, I'm good at this. Look what I can do. Here's a visual representation of that. You don't got to jump through hoops to see that I can actually do what I say I can do. Now, do you want me to do this for you? here's what I charge. Very, very freaking easy. Okay? Then you send it to them. Hey, I built these for you. No charge. If you like them, we can discuss more. Like, simple. Okay? Don't ask for calls and not give people any evidence. That's like way 2018. You can't do that anymore. And then you just want to follow up. Um, again, there's a lot of videos of this on the channel. Uh, I don't want to make this a full outreach thing. And then finally, number six is how to deliver. Now, this is broad. It depends what you're selling, but we have a lot of videos on our channel on how to do this um on different services, okay? Like this is a generative AI video. There's full case studies, there's full how to prompt, there's full prompt, engineering, agent skills, vibe coding. We have everything on this channel. I recommend you pick something here, you go watch the video that corresponds with that thing and then you figure it out. That's the beautiful thing about AI as well. If you don't know something, ask the AI, it'll tell you what to do. Okay? So, the bottom line is this guys. Y Combinator and everyone is hopping on this trend. It will eventually burn out. It will eventually become saturated. We are a year early to this. We've literally been doing this for a year. Like I showed you all these results. They're on our landing page on Go Insiders, etc. If you're smart, you will start this business. You will go as fast as you possibly can. You will hit the ground running and you will try to get clients literally today, this week. If you're not smart, you're not going to do that. You're going to get left in the dust and then it's going to be too late once AI gets too advanced. But there's this real arbitrage right now where the skills and the tools and the leverage you have is here and the knowledge of everybody's here. You can profit from that difference. Okay, this is let me level set with you for a minute. This is the greatest business I've ever seen for anyone to make a cash flow business. Now, not equity. You're not going to go sell this probably like I'm trying with software, right? But if you want to make a million bucks, 50 grand a month, you want to make 20 grand, you want to make 10 grand a month, this is the number one best business I've ever seen. Okay? So, go watch this video again. Go study it. Go put it into use. Go watch the videos on this channel. And if you want direct coaching from me and you want to skip the process of learning the hard way and get access to literally like 50 hours of new, fresh, in-depth content, book a call in the description below. And I'll see you guys on the next