I Started 5 AI Businesses in 2 years, here's what I learned..

0:00 Hundreds of AI business models are being thrown around on the internet right now. AI agencies suck, sell AI operating systems, AI infrastructure, AI growth agency. But in reality, most are just a clickbaity name on the same concept. So, in this video, I'll break down the only five AI business models you can realistically start in 2025, the pros and cons of each, the best one to start with in your situation, and how to transition between AI business models to grow to the next level. If you don't know me yet, I'm Ben. I've run all five of these business models over the last two years. And by the end of this video, you'll know exactly where to start and how to grow an AI business. So, first I'll cover the five AI business models, then the pros and cons of each and help you decide which one might be best for you. And then how to scale each of these business models or transition from one

0:45 into the other. Now, the first one uh speaks for itself. I think AI consulting where all you really do is providing AI guidance to either businesses or people. the whole world needs to be educated on AI and on these no code platforms like naden etc. So this is something I started off with too when I was learning relevance AI. I started offering some of this consulting helping people build out their projects advising people and businesses on how to plan these projects. I for example started with $100 an hour. Uh but many of these calls actually resulted in them asking me to build out projects for them. So that naturally transitioned into the next business model which is an AI agency or an AI automation agency the most well-known I think and if you want to have a good reference you can think of this business model more like a

1:31 traditional development agency or an IT agency because what you're really doing here is designing and building custom automation solutions for businesses and the key word really for this business model is custom. Of course, you can add training and consulting on top of these these implementations, uh, which you generally need to do anyway, but you're still doing custom projects for clients. Now, many of these business models that I see being thrown around on the internet right now are just people building a front end, uh, for their custom automation solution on Loverable or some other tool, which I highly recommend doing if it's necessary, but it still falls under the same business model because you're still building custom automation solutions for businesses. This is a business model I'm still running and we have been able to grow and scale. Uh another great example of this of course is Morningside from Liam Modley who's very much taking this

2:16 approach right he's transitioning more and more uh to also heavily investing in the consulting and training part but still they're offering custom AI solutions and consulting for their clients. Then the next model is what I call AI services which you can compare more to traditional marketing agencies, lead genen agencies, recruit recruiting agencies etc. Because the big difference with this model versus the AI agency or AI automation agency model is you're not selling custom solutions. You're selling outcomes, results, and KPIs to your clients. You're just using AI in your own processes to automate and enhance some of your delivery. And this can really transform a traditional agency business model, right? Great example of this is called IQ, a lead generation agency that heavily uses AI in their outbound email campaigns and can

3:03 therefore get better results for their clients, handle more customers than traditional lead genen agencies, and grow their business to margins that traditional agencies could never scale. We also recently started offering this business model. Basically, we're using our own AI SEO system to deliver SEO services and SEO outcomes to our end customers, which are e-commerce or SAS businesses. And the funny thing we're noticing is that many of these customers, they don't even really care that you use AI. They care about the outcomes. And that's really the key difference between those two business models. I talked more in depth about this business model in my last video. Uh so if you want to learn more about this, uh check it out. I'll I'll link it up here, too. Now, the next one is an AI software business, which you can see sort of more traditionally like a SAS type business. Now, the key thing here is that you're selling a tool. You're

3:49 selling a self-s served software to a business with features. Of course, it can be B2C too, but in general, I recommend going B2B. And of course, this self-serve software can potentially be sold to hundreds or thousands of customers. An example of this is 11X, which is an AI BDR software that companies can use themselves to generate more leads. So, instead of cold IQ, uh being a service provider where they do it for you, uh 11X is a self-s served platform to do it yourself. We've also built one of these softwares, our AI SEO software. So, we basically use our AI SEO software in two ways. We sell the software directly to marketing and SEO agencies and we sell the outcomes by using the system ourselves to end customers like e-commerce and SAS business because these end customers are not necessarily interested in adding another tool in their tool stack. They

4:35 prefer just paying for the outcomes. And the last one is AI education. Basically the whole world is behind on AI developments both businesses and people. So there's a gigantic opportunity uh in AI education too. In this business, of course, you can create either courses, communities or programs and what you deliver really is skills and and AI guidance, right? An example of this is 9X who uh does B2B AI education uh on these no code tools like air table, naden, relevance AI, etc. And of course, I run my own community too where I try to help business owners, but also people on how to start with AI and automation and how to grow an AI business. And of course, there are many other communities and learning platforms that you probably know of. So what are the pros and cons of each of these business models and where should you start? Let me go through it very quick. Now the upside of AI consulting of

5:20 course is that you only need some of the automation basics, right? You only need some weeks of experience with NAND or some of these other no code platforms to really be able to help uh other people plan out projects, helping them debug or design automation solutions. Therefore, the speed to revenue can be quite quick. But as generally you do hourly based pricing, this business model caps out at a pretty low revenue. it will probably be hard to scale this past the 3 to 5k uh monthly revenue mark. So generally what I recommend is to start out as an AI agency right away because as an AI agency your starting skills are very similar to the AI consulting. You only need some weeks of automation experience on these no code platforms. You can watch some tutorials, join a community and you'll generally know enough after a few weeks. Uh but the big upside with

6:06 the AI automation agency model here is that you're now also delivering and building the solution. So you can capture much more value out of a deal than only providing consulting. Uh that being said, it could be an interesting strategy to offer AI consulting uh as sort of a lead magnet or a lead generation method for your AI automation agency. I've seen this working for some other people in my community too. You basically provide low hourly rate consulting first. And because sometimes these result uh in them asking you to build out the the projects, it can be a sort of interesting lead generation strategy. But in general, I recommend going for the AI agency model right away because you can capture more value and all you really need is those basic automation skills, right? Clients will come to you with things they want to automate. You can learn while working together with clients and get paid while

6:51 learning basically. And if you're on your own, you only need one or two decent clients to really be able to sustain yourself, which is a huge step, right? Going in all in on AI and doing this every single day is going to steepen your learning curve a lot. Now generally you can charge either based on a project basis or on a retainer. Uh now the speed to revenue can be a bit slower than AI consulting because first of all the market is not very familiar with these kinds of services yet. They don't really know what to expect or what ROI they can get out of working with an AI automation agency. And second if this is your first business and you're just starting out you of course need to learn a little bit of sales and how to acquire customers etc. So it can take a couple of months. I I see it with many people in my community too. But generally when things start rolling, you can scale this business model up to a very good

7:36 business, especially compared to AI consulting. When I was still alone, I grew this to 20K monthly revenue. And when you're still alone, you can of course see pretty high profit margins, too. But when you really want to start scaling this up, this is very much a service type business, right? The only way you're going to grow and really scale this agency type business is by hiring and managing more people. Therefore, the profit margins in the long run will be much more like a traditional service agency or a development agency which generally are around the 20 to 30% margin mark. Right? So, this is why the model I think is getting a little bit of a push back or some negative comments on from people on the internet. But this doesn't mean it's not a viable business or not a good business. We've scaled my agency to consistently 70k uh in monthly revenue, sometimes with higher months. Now that

8:22 being said, there is much more of a revenue cap than some of these other business models because there are quite a lot of growing pains with hiring, managing, scaling, and getting the right customers. I don't think there are many AI automation agencies right now that have a higher annual revenue than $1 million, for example. And if there are, they're probably in count them on one hand. That that being said, I think the AI agency is an amazing business model to start with because first of all, you can start this business basically from zero by learning uh these no code automation platforms like NAN in literally weeks and second, it's a perfect ramp to potentially go into other business models that are more scalable because you can build domain expertise while building and working with clients to potentially transition into an AI service business. And it's

9:08 also a great method to find potential software ideas and sometimes even get paid to build them out uh for yourself. And this is exactly what happened with me. All my other businesses came out of me running an AI automation agency. Then we have the third business model uh which is the AI services. Now in order to do this, you really do need some domain expertise in order to deliver results to clients. So if you have a marketing background, a content background, maybe an SEO background, lead generation or even recruiting background, you can really consider going this route uh first too because the big upside with this model is you're selling outcomes which customers are much more likely to pay for the sale is made a lot easier and therefore the speed to revenue can be even faster. And while you're delivering the service to more and more businesses, you can

9:53 automate more and more of your service delivery with AI. Meaning that the revenue cap of this model is really really high and potential profit margins can be much higher than traditional agencies and the AI agency model because you can potentially handle many clients with a small team and AI automation. For example, we can handle almost 20 SEO clients with one account manager which is unthinkable for a traditional SEO agency or an AI agency. Cold IQ grew their AI lead generation agency to over $6 million in uh annual revenue in I think two years. And Superside, which is an AI content agency, grew to $60 million in annual revenue with a relatively small team for an agency type business. So these numbers are really, really hard to reach for a classic AI

10:38 automation agency. Now, for an AI software business, you really do need technical expertise uh because you're building a self-s served product here. You probably need to go full code very soon, too. Uh you need domain expertise because you really need to be able to solve a problem uh for people in a specific industry. You need go to market experience because you really need to build scalable acquisition channels. Uh so, if you don't have extensive SAS or startup experience or have big funding connections, etc., I generally recommend starting out as an AI agency or an AI service business which can be a great ramp to eventually get into an AI software business. In terms of pricing, there are many different pricing models to softwares. Of course, an interesting one that we can now do with AI, which is what we're doing too, is selling the entire product, meaning the NAND solution plus a front end uh in air

11:25 table at a one-off price, which was traditionally of course hard to do in the software industry. Uh, of course you can also go the full SAS route and go with recurring pricing, usage based pricing, outcome based pricing or a mix between these. In the AI software era, I think we're going to see more and more outcomebased pricing. Uh, but if you want to learn more about how to build an AI software from scratch, I have a full road map to starting an AI SAS fast uh also on my channel. Now, the hard thing here is of course it's really hard to build good products and get enough customers and the ticket is generally a lot lower. So generally the speed to revenue is by far the slowest of all these business models. But of course the potential upside is you can sell the same software to thousands of businesses meaning the scalability and the profitability are the highest of them all. I think 11x is valued at 350

12:11 million after 2 years and I think uh they're still under 30 employees. Now of course this is a little bit of an outlier but it does show the potential of a software business versus these other types of businesses. Now the last one for AI education you most likely need either technical or domain expertise depending on what you teach. Generally AI education requires also scalable acquisition channels and go to market experience. Plus the key ingredient is really uh you generally need experience in one of these business models in order to do AI education. Nobody wants to buy a education product from someone who doesn't have experience in what they're teaching. So therefore, I highly recommend not starting with AI education, but only doing it after you have some experience running uh one of these other business models. And this is what you see happening a lot, right? People have experience running an AI

12:57 software business, an AI service business or an AI agency, adding an AI education program on top of their existing businesses. Now, it might sound like an amazing business model, these communities, etc. But I can tell you from my personal experience, it's not as easy as it looks uh to run one of those businesses. the speed to revenue generally can take quite long because you really do need scalable acquisition channels. There's definitely also much more of a revenue cap uh than for example compared to an AI software business or an AI service business. But generally, yes, they can run quite profitably. So, how do you scale each of these business models and how do you potentially transition from one into the other? On the top side in the blue bar I basically have how to scale in the same business uh model vertical and on the bo bottom I have strategies to transition

13:43 from one uh business model into the other. Now in order to scale any of these business models you do need to scale acquisition. That's why I put it at the top. I don't want to talk too much about uh scaling acquisition in this video. I'll soon do a video on that. But I want to give you tips here on what to focus on inside of the business in order to scale or to transition. Now, if you're doing currently doing AI consulting, there are really two routes you can take in order to scale. First, you can try to scale your revenue in the AI consulting here or you can transition into the AI agency model. Now, the main way to scale your revenue inside of AI consulting is of course to work with bigger clients, charge higher prices, maybe package the consulting, and maybe eventually scale a team where you let other people do the the actual consulting. Now a far better strategy in my opinion is to transition into the AI agency model because by

14:28 providing the actual solutions you can capture much more value out of the deal in my experience and the way to do that is simple. You just have to add implementations or the actual building into your offer. Now in order to scale an AI agency the key thing in my experience has been to first of all work with bigger clients. It's a it's a big mistake we learned the hard way. Trying to work with more and more small businesses and small projects is really really hard and if not an impossible way to scale your agency. Instead, you should work with a similar amount of clients but bigger and bigger clients that pay you higher and higher project prices or higher retainers. That's really been the key difference for us to hit that next level of scale. And in order to do that, of course, you need to target uh bigger companies, mid to large size businesses or even enterprise type businesses. Uh, another important factor

15:15 is adding consulting and training into your value proposition because this is first of all what businesses really need and especially these bigger businesses need and it really allows you to charge a lot more by adding this into your offer, right? And consulting, you don't have to think about Mckenzie type consulting. It's just a weekly call or a bi-weekly call and maybe you can add in a training once a once a month for their team, etc. Right? don't have to over complicate it but it really uh improves your value proposition. And from there the third key factor uh to scaling an AI agency is really have to get good at hiring, managing, having SOPs and scaling up a team in order to scale an AI automation agency. Now once you have experience running this model, this is the moment you can decide to add an AI

16:00 education product. This is what I did. This is what Liam Mley did with Morningside. Uh what the guys at 9X did. uh they previously had an automation agency and later completely transitioned into uh B2B AI education but again it came out of experience. Now an alternative way to scaling is to actually transition from an AI agency into an AI service business which is of course a much more scalable business model and it is important to think about where you eventually want to head because you should focus your current strategy on things that will allow you to make that transition. Right? So for example, if you potentially want to transition to an AI service business, you really need to build that domain expertise in the specific area that you want to deliver outcomes for. So that means you can really try to target, for example, agency type businesses for your

16:46 AI automation agency because agencies generally have a lot of domain expertise. They know how to deliver outcomes and by working with those clients, you can really build your own domain expertise and potentially can be building outcomebased solutions that you can later use yourself to become an AI service business. I, for example, worked with multiple SEO agencies that really helped me to eventually get to this AI service business. And the second route of course is to go and transition into an AI software business, which is what my agency also allowed me to do. As I was building more and more SEO automations for SEO agencies, eventually we were able to build a self-s serve SEO software. And really, the way you can take this approach while you're in your agency is by really, first of all, niching down on one specific industry because you're really going to understand their biggest pain points and

17:32 build solutions around them. Second, you can also, which is a strategy we took too, focus on only accepting potentially productizable projects, meaning solutions that can likely provide value to other businesses in that niche too. Then you can try to resell it to multiple business in that businesses in that niche. Even if you still have to customize it and eventually you can learn so much that you can build it into a fully self-s served product and go the AI software route. This is also approach uh Yonas, a friend of mine and a member of my community took with Aura AI. He built voice agents and he niched down into voice agents for hotels. So he started selling custom voice solutions to more and more hotels and eventually he learned so much that he could build a fully self-s served AI SAS for the hotel industry that's now going quite well.

18:18 Now this might sound difficult building a self-s serve software but it's actually a lot easier now than you think with these no code tools. Uh I also have a full tutorial on my channel uh which I'll also make sure to link in the description below on how to build a self-s serve uh AI software with no code. Now if you have an AI service business or maybe a traditional agency service business the way to scale this uh is by first of all providing better outcomes as you generally get paid for better outcomes. You get more referrals etc maybe more commissions. So the better outcomes you can provide the higher prices you can charge. The second factor here is really try to automate more and more of your processes or your delivery because the more you can automate, the more clients you can handle. And the third of course is as this is still a service uh type business, you do need to get good at

19:04 hiring, managing SOPs and scaling up a team. But again, you can grow this business to really really interesting numbers. And then again with the experience you have running this uh AI service business you can then decide to add an AI education business on top of this. This is exactly what called IQ and Navo has have a very similar value proposition as cold IQ have done too. Uh they teach other lead genen agencies now but also end customers like SAS businesses on how to use AI and these tools uh to generate more leads for their businesses. So they added that on top of their service business. Now, you can also decide to transition from an AI service into an AI software and I talked a little bit more about that also in my last video. Uh, and I think there's a huge opportunity and we're seeing this happening more and more. Uh, which is basically the productization of the

19:50 service industry. More and more of traditional services can now be productized with AI, meaning it becomes far more accessible to SMBs and smaller businesses. And the way to do this to is to try and automate as much as possible of your service and of your service delivery and eventually uh productize it and build a fully self-s served product out of it. Now two good examples of this are first of all hopp copy was which was traditionally a newsletter agency. They fully built a self-s served product around it. So now small business owners can build newsletters by themselves at a very low uh price right instead of a a service a newsletter service. Gloria AI is another example which was traditionally a meta ad agency. They made it into a completely self-s served product. So SMBs and tiny business owners can create their own ad creatives

20:37 and ad copies at a very low cost compared to a traditional agency. Now the interesting thing here is that higher budget clients don't necessarily want another software or tool in their tool stack. They are happy to pay for service agencies delivering these outcomes. So it's not necessarily a transition you have to make. It depends a bit on uh on your target. Uh lastly, how do you scale an AI software business? Of course, the best way of scaling is first of all to build a better product, getting product market fit, you probably have to build some hire some engineers here. If you're really going the SAS route, um you can start increasing pricing, experiment with pricing. Uh the big one with AI software of course is really scaling acquisition. Once you have that product market fit, uh again with that experience, you can then also decide to launch an AI education business. Uh this is also what you see more and more

21:22 software businesses investing into, right? building communities and education around their pro products. Generally, they're more like free education platforms together with the product, but it makes a lot of sense because it helps educate people on how to use the tool as efficiently as possible in order for them to use the tool more. Perspective is a great example. It's a funnel builder, but they highly invest in their community to to help people build the best possible funnels on their tool. Now, that's it for this video. I have a full video on scaling an AI automation agency, an AI service business, and an AI software business or an AI SAS. So, if you're interested in one of these three and want to dive a little bit deeper, I'll make sure to link them all in the description below. Also, if you want to learn how to build uh no code self-s served softwares, I also put a link for that one uh in the description below.

22:08 Lastly, if you want to work closer with me and my team, you can also join my community where we try to teach people AI and automation and how to grow an AI business or how to implement AI in your current business. So, if that's interesting to you, you can also check out my uh community in the link in the description below. Thank you so much for watching. If you got any value out of it, I highly appreciate a like and a subscribe. It really does help me. Uh and I hope to see you in the next one. Thank you so much.