Transcription
I'm David Andre, and this is how to make money with AI agents. The demand for AI services and products recently hit an all-time high because AI is a growing industry. Unlike other businesses, this one is not easy to replicate because it requires some skill, right?
Also, a recurring model is essential. People think, "Oh, I'll build agents for somebody, and that's it." In most cases, especially if you do a good job, your clients will request updates and improvements through agents. So, let's say GPT-5 comes along. If your agents are running GPT-4, obviously your client is going to want GPT-5. This means this is actually a recurring model.
Here is my promise with this video: I will break everything into small steps, including who to target, how to get clients, and much, much more. You can do this even if you have zero entrepreneurial experience. I'll show you all the steps required, and if you watch this until the end, you'll be ready to start your first AI business.
Now, to make money with AI agents, you first need to learn how to build them, obviously. So, pick an agent framework and master it. Personally, I would recommend CreAI or Agency Form. I mean, I've tested dozens; I've tested so many, right? I've made dozens of videos on AI agents, and these two seemed the best, especially for people who actually build agents professionally.
Honestly, if you are lacking in this part, if you don't have the skills to build AI agents, there are endless tutorials online. Just on my channel, you'll find so many videos on how to build agents in just about any agent framework. So, don't just watch them, though. That's the lazy way; that's easy. Make sure to actually recreate everything yourself. That way, you learn what the people are teaching.
This is my tip: how I do it, right? If I'm watching a tutorial, I recreate everything. I don't just watch it; I actually build what the person built in the tutorial. Only then, after I finish it, do I take it and adjust it to my use case. This is how I learn way faster than anybody else who's just watching and not implementing. Then, like in a week, you forget it because you haven't put it to practice.
Now, to become a 10x agent builder, you have to be truly exceptional. Why? Because 99% of the economics in this industry will go to the most ambitious individuals. If you build terrible agents that don't work and break often, right, and are inconsistent, clients will go to somebody else. Companies will go to somebody else who can just build them reliably and who has a proven track record.
So, if you want to do this, do this fully. It's not a side hustle that you do five minutes a day or, you know, no passive income. That's a myth. This is a serious business opportunity that has potential, and if you commit to it, you can reap the rewards. You'll have to learn both the AI side and the business side.
The good news is that if you're lacking in one of them, you have an advantage over people who don't have the other, right? So, if you have a business background, that's an insane advantage over those who don't have it. But if you have a background in AI or machine learning or building agents, well, you're better than the people who have business but no clue there.
Ideally, you have both, but if you don't have both, just make sure to learn the other one. This video will hopefully give you both sides of the coin. For direct access to me, if you get stuck at anything or if you want access to my software engineers and people who are at the cutting edge of AI and help with building agents and learning how to build advanced agents, join my community. It's going to be the first link in the description.
Now, picking your niche is one of the most important steps, and people are tempted to skip this, which is a massive mistake. People who are new to business just target everyone; they don't choose a niche. There's like, "Anybody who takes my money, you know, I'll do it." That's a mistake. You should actually choose a niche.
What that means is just narrowing down, choosing a certain demographic of people who can benefit the most from your services. As Alex Orosy says, "The riches are in the niches," and he says that because the more you niche down, the more you focus on a single avatar, a single demographic, the more you can market to them. You can run ads specific to 45-year-old homeowners, males, you know, you get the point. The more specific you go, the higher leverage you get on your advertising, on your content, on your copywriting—everything—because you understand your ideal avatar so much.
So, how do you choose? Well, start with your expertise. People are like, "Oh my God, what if I have no expertise?" Calm down; all of us have some expertise, right? You've done something. You had some interests in high school. You have parents that did a certain job, and you just picked up what they did by watching them. Everybody has some expertise or some area they would like to get into. Go there; just start there. Don't go somewhere where you have no advantage.
This is one of the most common places where most people break: "How do I find clients?" Sure, this might be intimidating if you don't have the business side, but if you have the business side, then building agents might be intimidating. So, don't worry; I'll break it down. Define your ideal client—who will get the most value from your service—because those are the people who can pay you the most. We'll get to pricing in a bit, though.
Now, there are many ways of generating leads. If you're new to business, you might be lost for words; you might not know what to do. But there are so many ways of getting clients, right? You can do cold emails. Don't just spam people; write actually personalized emails that you spend 20 minutes on per email. That's how you get people's attention.
Of course, you can run ads on Facebook, YouTube, TikTok, Snapchat—just about any platform. You can also go on Fiverr or Upwork and list your services there. That's how I made my first $15 back in 2018 when I was doing Instagram marketing. I just posted on Fiverr, and I got social proof because I built it on my account, and so somebody bought.
You can also make content on YouTube or any other social media platform, showing how you can add value and your expertise. So, if I was doing the service, if I wanted to be a professional agent developer, I would be doing what I'm already doing, basically, but just marketing my agent-building services. But I don't do that; I focus on my community.
Every niche will have a different strategy that works. The reason that I put "works" in quotation marks is that everything works; you just need to do enough of it to become good at it. That's why you should know the niche also because in every niche, something is different. Right? Like if you, let's say, completely extremes, right? Gardening—you don't target people in the gardening niche the same way you do people in programming, let's say, right? Completely two different niches, and you have to understand the niche inside and out, which will help you target them more efficiently.
Now, volume is the answer when it comes to getting clients. Most beginners grossly underestimate how much volume is required. I mean, Jesus man, don't build half-built bridges. Don't try sending five emails, making two shorts. No, pick one way of getting clients and stick to it. Stick to it for at least three months until you figure it out. Because if you quit before you figure it out, you have to start from scratch in the next way of getting clients.
Do more. This is like the easiest way to get an advantage. If most people are sending five personalized emails a day—which, to be honest, most people are, are sending less than that—just do 50. If you do 50, you will succeed, even if you are new to this, even if your offer isn't perfect, even if your agents have some bugs in them. You will just succeed based on the volume.
Now, how do you craft a compelling page that will attract clients? Well, talk about what they will get from your service, not how you do it. Nobody cares about that. Nobody cares if you use this agent framework or that, or if you use Python or if you use something else. What nobody cares about is what it does for them. Will it save me money? Will it save me time? Will it make me more money? Sell them the vacation, not the airplane ticket.
Now, use case studies. This is huge because if you can show, "I've already done this," that takes away so much of the perceived risk from the client's eyes. You can do this even if you have none, right? So, let's say you're watching this video, and you want to start from scratch. You will actually— you know what? I think I have a slide on this, so we'll get to that in a bit.
Now, the offer is everything. For example, just adding a money-back guarantee—no questions asked—will improve your sales. I'm speaking from experience; I'm not just making this up. When I did this for my community, sales improved because there is no risk. You take all the risk away. If people don't like it, if they don't get the value, just give them the money back.
Then you actually have to do it. Never argue with people who want a refund; just give them the money back, right? It's much easier. So, yeah, if you want to build a great offer, I definitely recommend Horos' book, "$100 Million Offers," because this will decide whether you struggle with getting clients or whether you just get clients hand over fist. If you make an offer that people feel stupid saying no to, they won't say no.
Now, here is how to build a portfolio of case studies even if you have no clients. This is another thing that people are tempted to skip, but if you do this, selling to people is much easier. You start by developing three to five really solid agents for problems that companies in your niche face. This is how you can have case studies even before you get your first clients because you basically do the work as if you were doing it for some company, even if they don't pay you.
You could also do it for an actual company that you select. If you do the work ahead of time and you present the agents and just give them to them, they’ll be like, "Whoa, that's insane!" Then you can charge a recurring fee that you will update them and keep improving them. This would be a no-brainer offer.
Obviously, it takes time upfront, but if you have a company in a niche that you know you could help, you build them solid agents. You put in 20, 30, 40 hours of work, build a solid team of agents for them, give it to them for free, and then say, "If you want, I will keep it up to date. I'll make sure it uses the latest and greatest LLM, and I'll keep improving it for $300, $500 a month, whatever, depending on how much demand you have, obviously."
So, pro tip: put in as much work as you would do if someone paid you five grand. Most of you will not do this; that's why most of you will fail. But if you want a really solid case studies portfolio before you even have any work, just make sure to work as if you were paid for it. It just takes work; that's the secret. This is not passive income, make money quick—none of that. This is a real business model that can make real money if you so decide.
Now, pricing is also where a lot of people make a mistake because they undercharge themselves. The more demand you have, the higher your prices should be. If you are beginning and you have no clients, obviously you cannot charge 10K per agent or whatever. But also, if you charge too low, you seem sketchy, right?
So think of it like this: if your agents can do the job of a $40K employee—let's say a company is hiring, right? They are hiring for a position. By the way, this is another way of getting clients: look for companies that are hiring and see if you can automate that thing so they don't have to hire a full-time person. You can maybe do it with agents; there are just so many ways to do this.
Now, if a company is hiring for a $40K position that you can easily automate with agents, you can charge them 4K, and they will happily pay that because that's just 10% of the full-time position they would otherwise have to hire an employee for. This is called price anchoring. You do a high price and then anchor it to something low. If you just say 4K upfront, that seems like a lot of money, but if compared to a 40K annual price on an employee, 4K is nothing, right?
The goal, eventually, over time, is premium pricing. Never aim to be the cheapest agent developer. This is the danger of starting on Fiverr or Upwork because people want to, you know, they want to go in. They fall into the trap of, "I'm the cheapest. I do a little bit more than everyone else for a little bit less." That is terrible. You want to achieve premium pricing. You want to be the best agent builder; you don't want to be the cheap guy. You want to be the best guy.
Now, in order to be the best guy, you obviously have to deliver excellent results. The key to doing that, the first step, is under-promise and over-deliver. Don't promise. Most people promise the world and then do a mediocre job. No, just tell them, "This is what it will do. It will not be perfect maybe the first time; maybe it will take a few tries, but I will keep working on it until it's good."
That's it. Don't over-promise; just under-promise and then over-deliver by, you know, giving them extra support, by giving them some tips on how to use the agent, how to use more AI. Maybe you give them some ChatGPT tips, whatever. Setting expectations is such a key thing, especially in preventing people from canceling and asking for refunds. If you set solid expectations and you meet them or even over-deliver on them, people will love you.
Now, personalization matters. Tailor your communication and solutions to each client's unique needs. This is important because you have to do this both with the lead generation part and with the delivery. In the lead gen, as I said, don't just spam people with emails. I mean, dude, I'm at a size where I get dozens of emails a day because of my YouTube, right? I don't respond to most of them. In fact, I don't even open most of them because I can see that they have been either written by ChatGPT or maybe written by a person, but just like they're spammed to hundreds of other creators.
If you want to get somebody's attention, you have to write an email that's specifically for them, which takes time. You cannot automate everything with ChatGPT. I mean, at least not yet, obviously. But again, if you spend 20 minutes writing an email for somebody, the chances of that person responding and actually booking a call—a discovery call or a sales call with you—go through the roof. It's like 50 times higher than if you're just spamming emails, and then, you know, they're just going to end up in spam anyway.
The second part is the delivery. You cannot build the same agent for everybody. Obviously, there will be rare scenarios where the use cases are perfect, and maybe you can use part of the code base you've built earlier. But in 95% of cases, you will have to build the agents from scratch. So do not copy-paste code or try to sneak in some automations. No, you will have to tailor the product to each client's needs, to each company's needs—not just copy-pasting stuff and doing the same.
Also, go the extra mile. Offer additional value beyond what's expected. As I said, it could be some cool MidJourney images, could be some ChatGPT prompts—doesn't matter. I mean, obviously, the more relevant it is to that company, the better. So if you see they have some bad branding or, you know, their website has some poor images, just by creating some images with MidJourney, you can already give them insane value.
Again, the more you know your niche, the more you can do this. Now, attacking roadblocks: you should expect problems to arise because there is always the most friction at the start. If you're new to this, you will not get it right the first time. You should expect some problems to arise, but prepare for them mentally and also attack them right away. Don't just let that stop you. Don't let a simple roadblock prevent you from doing this if you want to do this, obviously.
So, if you struggle, for example, with getting clients—which a lot of people do—you either have bad lead gen or a weak offer, or both. So figure out which one it is and then work on that until you fix it. Now, if the development part is taking longer than expected, which is a problem if you have a lot of clients, a lot of leads, but you are taking a lot of time to deliver, then you didn't prepare enough.
What that means is that you didn't either ask enough information before you took the job, before you took that company to build the agents for them, or you didn't prepare enough in terms of getting your skills up. Simply, you are not good enough at building agents. That's why something that should have taken you three days is taking you two weeks. So identify the problem and work on it. Don't get overwhelmed; just, you know, make some progress every day.
Now, scaling: what is the potential of this business model? Because people think, like, people just want to make like $200, and that's like so unambitious. Listen, with this exact model, Vren, a fellow agent YouTuber, scaled his AI agency to 50K a month in just two months. Now, I'm going to be real; I'm going to set the expectation straight. You will not be able to achieve this in two months because he already had the skill sets.
He already was really amazing at building agents, so when he launched it, he had a ton of demand, and he was able to scale fast. If you don't have a skill set and if you don't have his demand, you will not be able to scale as fast, which doesn't mean you cannot reach the scale. You absolutely can; it will just take longer than two months, obviously.
Now, once you master the process, you can actually hire agent developers and train them underneath you. The myth is that you have to do all the work yourself. People think this is just a leverage job, and it might be at the start, right? Which will be a very well-paying job, working your own hours with potential to scale. Don't get me wrong; that is already amazing. But the goal is to eventually build a team, right?
You should only do this after you are really good. Don't hire too early. People think, "Okay, I'm making 5K a month, 7K a month. I'll just hire somebody for 1K a month, and they will code the agents, and I will focus on getting more clients." No. First of all, for 1K a month, you will not hire anybody good—like, not even close. For 1K a month, you will only hire developers who will build agents that do not work.
Second of all, you have to do the thing. You cannot have an agency where, like most agencies, the reason why agencies are hated is because you buy them, the best guy does the work, and after a few months, you downgrade to somebody who's junior. That's why most agencies are so bad and so hated, right? So don't hire too early. Make sure you are drowning in demand first.
You have to have clients that are begging you to work with you. You have to have a full email list, a full lead gen process of people wanting to work with you. Then you might hire, but also make sure you have the skills. Because if you don't have the skills and processes, you cannot even train the people. If you don't know what to look for, you cannot train somebody else to do the job for you.
So don't hire too early. Make sure you are drowning in demand. Make sure you're actually good. Then, once you do hire, set the expectations. Don't promise them that you will do it personally if you give it to your new employee. No, set it like, "I have a team. I'm training them. This is the process. These are the results you can expect," and be honest.
If you want to see more videos like this—because in the past, I didn't really make videos on how to make money with AI, but so many people requested them that I'm open to it. If you want to see more videos like this, just comment down below, and I will make them. Thank you for watching.