Transcription
Here are the 14 best ways to make money with artificial intelligence, even as an absolute beginner. Currently, every second online guru wants to explain to you how to make money with AI. The problem with this is that most of them don't generate any revenue with AI themselves, or at most only by selling some AI courses. Therefore, today we will undertake an honest ranking from flop business models, meaning business models that are really stupid ideas, to top business models, meaning the best business models currently available in the AI sector. And we will classify each business model according to profitability. So, how much can you earn with it, according to the competitive situation? How many people are already doing this, and longevity, meaning is the business model outdated after two weeks, or does it have staying power for the next few years? Ideally, even decades. I myself, with Everlast AI, have not only built the leading German-speaking YouTube channel in the AI field in a business context over the past few years, but with Everlast AI, I have also built the market-leading AI consulting and implementation agency in the German-speaking region. We generate over 1 million Euros in monthly revenue and help thousands of medium-sized companies implement AI daily, and at the same time, we have trained thousands of people over the past few years as a state-approved educational provider in the practical implementation of AI, always with the goal of using it profitably, saving costs, and earning more money. Therefore, this will be a practice-oriented classification that you won't get elsewhere on YouTube. Especially for the German-speaking market, because there are some exciting ways, which I will come to shortly, that are often not even discussed in the USA and don't work that well at all.
Number 1 is AI Voice Agents. Voice Agents can be classified as good. Voice Agents are AI telephone agents that handle any phone calls for companies. The typical example is AI receptionists, meaning telephone assistants who take incoming calls, book appointments in calendars, or answer simple questions for which employees often have no time, or these calls are often outsourced to external call centers, and you can implement such voice agents for companies. This business model is classified as good because it is a) very profitable. So, you can either resell such voice agents as a retainer, typically between 500 to 1500 Euros per month. This makes particular sense if you specialize in a specific industry. For example, we advise an AI agency that specializes only in voice agents for the hotel industry. This makes sense because, eventually, you will know all the industry software systems of the hotel industry and have integrated your voice agent use cases, and then you can implement the voice agents for other companies via a monthly retainer without too much effort. You can also implement such voice agents, which is also the path we take with Everlast AI, as a larger fixed project. This means you take three to six months and then implement more complex niche projects. And for such a six-month AI voice agent project, for example, you can quickly reach the mid to upper five-figure range, because you always have to understand that with a voice agent, you can quickly replace a full-time position that costs a company a good 40, 50, 60,000 Euros per year. And this is the added value you deliver to your life. The competitive situation is low to medium. Voice agents have long ceased to be a new topic, but there is an incredible amount of talk and little implementation. So, there are really few providers who implement voice agents correctly and scalably in the German-speaking region, and many fail at trivial things like data protection. You can quickly see this if you do a few internet searches, for example, just search for "AI telephone assistant for tradespeople," and the first ten websites, without me having to name them, I can already tell you, are not GDPR compliant and would be subject to a warning based on their website alone. That's why I say the competitive situation is not really to be taken seriously, even if there is a lot of talk, and therefore it is still a really good business model, and the longevity is high and strong, as voice agents, once implemented, will be making calls for the next years, if not decades. And the trend will naturally continue. At some point, the question will arise, if voice agents handle all inbound calls, why should I even call anymore? And so, an entire agent economy will emerge in which voice agents call other voice agents. For example, Labs released the Gibberlink framework a few years ago, meaning voice agents develop their own language. But the trends are essentially clear. Voice agents are a good business model, still have high demand, still have high profitability, the competitive situation is medium, you have to know your stuff well, which is why it's good and not top, but certainly a solid business model.
Number 2 is Faceless YouTube and TikTok Channels. This is an absolute flop business model. Faceless YouTube channels essentially mean you build so-called cash flow channels on which you create AI-generated content, build reach, and then try to earn money through reach or affiliate partnerships. Profitability is very low. You naturally have high AI costs initially to produce all these videos. You have to invest a lot of time and effort to build significant reach here at all. And in the end, you earn a few Euros with some affiliate partnerships, and many of these channels are now being demonetized by the platforms YouTube, TikTok, and so on themselves. This means you can't even earn real money through YouTube AdSense anymore. The competitive situation is also very high. You can imagine that many Chinese e-commerce providers are now building entire bot farms and AI armies and AI influencer armies. And whether you, as a career changer from Germany, can really keep up with that is truly questionable, and the longevity is, of course, also extremely low, as you have very low viewer loyalty. You don't build a long-term brand, and you can be replaced at any time by another AI channel that produces videos a bit faster or a bit better than you. And that's why I clearly recommend staying away from this business model. An absolute flop.
Number 3: AI Trading Bots are just as much an absolute flop. Unfortunately, you read it again and again that many people get the idea, if AI is so good, then I'll let AI trade for me on the stock market or trade cryptocurrencies. However, the profitability here is likely already in the negative range, and in terms of naivety, the whole thing is also hard to beat. You are competing with the most intelligent and best people in the world, Harvard students, large hedge fund companies that execute such trades at the nanosecond level. And it is completely naive to assume that Claude or Open AI or any of your own AI trading bots will suddenly bring big money. So, I would completely stay away from that, especially due to the longevity, meaning this is not a sustainable business model with which you build a real company, reputation, customer loyalty; you are always completely dependent, in the truest sense of the word, on the market, and therefore an absolute flop business model from which probably only the trading coaches earn in the end, but not those who actually use such bots.
It looks completely different with business model number 4, namely AI Consulting. This is a top business model. In AI consulting, you use AI audits, for example, and identify the biggest bottlenecks within companies where it makes sense to implement AI. The current situation is that most companies have understood that they need AI, but they are asking themselves, where do I even start? So, where does it make sense to start with AI? Many companies are simply overwhelmed, don't take action, and therefore need an audit, and you can sell such audits for, for example, 5000 Euros, analyze companies in depth, and based on that, create a strategy, create a roadmap with which it makes sense to start with AI. The real project, however, usually arises afterwards. So, a really larger consulting project over 6 to 12 months, for example, again in the mid to upper five-figure range, in which you advise and support companies long-term in AI implementation. And even IT departments often need subject matter experts from the AI context. We ourselves advise IT departments, larger companies on the integration of AI processes purely in a consulting capacity, because they say, technically, we can already implement it. We have access to all our systems anyway. We might not want anyone in there, but we want to learn from best practices. Profitability is therefore, as I just explained, quite high. The competitive situation is low. Especially in niches, there is hardly anyone who does it professionally. So, if you offer AI consulting only for mechanical engineering companies, AI consulting for doctors, AI consulting for the hotel industry, as we just said, and you can go even more niche here, and you currently have really little competition, especially you have little competition that really knows the processes and topics well. It makes particular sense if you perhaps come from such an industry yourself and currently perhaps don't have the capacity to implement a large AI project, for which you would quickly need employees, freelancers at least, or even permanent developers, then AI consulting is also a good entry point. The longevity of the business model is also very high. Many make the mistake of assuming, yes, even Essential and McKinsey and so on, they won't be needed at some point. The mistake here, however, is not understanding exactly why you bring in such consulting agencies, because you usually don't do it because the advice is always the very best and you wouldn't be able to do it in-house or with AI, but because you outsource the responsibility. So, if you, as an AI consultant, audit such processes, collect them in a report, create a recommended action plan at the end, or accompany the ongoing implementation, then companies outsource the responsibility to you. So, when we advise and support companies, we naturally have the responsibility to ensure that everything corresponds to best practices, that everything functions legally and is GDPR compliant, and thus department heads or managing directors can also hand over responsibility, and this is something you want long-term, even if AI gets better, because you can't say, hey, I asked Cloud, Cloud recommended this AI strategy to me. That won't be the best way to explain to the board why something failed. However, if you have a consultant by your side, then you also have someone who takes responsibility, and that's why this business model is actually quite long-lived.
Number 5 is AI Copywriting. Copywriting can be classified as medium. The business model means you write landing page texts, emails, newsletters, social media posts, and it's also a very popular business model in the old world. And copywriting is medium because it still works. So, profitability is still medium. It's okay. It's naturally getting better and better with AI, because you can provide enormous added value to companies that really have no idea about copywriting if you have even a basic understanding of how to work with AI tools. The competitive situation is, of course, quite high in this area, and the longevity is also not particularly high, because AI tools are simply getting better and better. And what most people don't understand is that copywriting is increasingly being solved by context engineering. So, prompt engineering is all well and good, and many companies fail at it, but the better the context is prepared in these AI workspaces of companies, and a whole marketing operating system is created for companies, the more irrelevant the skill of copywriting itself becomes in the end. And this only works if you become an authentic marketer in the long term, meaning if you help companies beyond copywriting, that actually offers more potential, but pure AI copywriting itself is rather medium.
Number 6 builds on this exactly. This is AI Web Design. This can also be classified as medium. This means you create websites with AI tools for companies. So, you don't use WordPress, you don't use Webflow. You also don't have freelancers program websites from scratch, as some website agencies out there still do, but you use tools like Claude Code, use tools like Codex, and build your own systems to create websites in record time. Profitability is currently very high because there is still a large divergence between what companies out there know is already possible with AI and what is actually possible. So, if you can currently build good websites with Claude Code, then there is a very high probability that 80-90% of companies out there will think you worked on this website for weeks. There is, of course, a high competition from web designers in the old world. Fortunately for you, there are also many business gurus from the old world who haven't understood this yet and still recommend the old ways and also recommend customers to offer completely overpriced websites, with which customers will of course only be disappointed in the end. When you come around the corner and offer such a website at a fraction of the price, and then they ask themselves, why did I spend so much money? If you can deliver it faster, better, and cheaper. Longevity, however, is rather medium, because here too, the trend is foreseeable that companies will increasingly create such websites themselves. This means the AI web design path only works as long as the majority of companies haven't understood what's possible with AI. This will continue for a few more years, but from a 5 to 10-year perspective, I would be quite skeptical. Currently, it's a very good way to actually shake up the market, but not too long-lived, and therefore classified as medium.
Number 7 is AI App Development. This is a top business model. This means you develop custom internal web apps or even mobile apps for companies from scratch using agentic coding, meaning with Claude Code, Codex, Antigravity, and so on, you actually build productive web apps for companies. Profitability is very high. You have to consider that classic software development projects in the old world often started in the low six-figure range and had no upper limit. And if you implement a complete full-stack app for a company in the mid to upper five-figure range, then if you do it right, you are extremely profitable, and companies still get an extremely good deal, as the added value of such software is simply enormous, and the alternative in the old world is even more expensive and even takes significantly longer than the AI app development that you can implement. Competition in this area is currently also very low. So, there are really hardly any providers who specialize in developing AI apps for companies. I personally know several agencies that are overwhelmed with orders. It's similar for us, so we even have to turn down orders because we have far too many inquiries to develop apps for companies. I also know several other agencies that have full order books and can hardly keep up themselves. And therefore, competition is currently simply very low. There is little supply, there are few people who can deliver professionally, as many have not yet understood what is possible with agentic coding, and especially have not even taken the time to learn it in depth, and this is also possible as a non-programmer. Yes, the whole thing within, for example, we have the Agentic Coding Masterclass. With it, you can learn this entire skill from A to Z within 40 hours of course material, even as a complete career changer. And the longevity of this business model is enormously high. Many make the mistake of thinking, yes, but if it's so easy with agentic coding, then companies will do it all themselves. That's true. So, more and more companies are building MVPs and prototypes themselves. But what happens here is that demand only increases because in the old world, very many companies did not even consider this option. They wouldn't even have thought of building an app because they would know, I don't have the budget for that, it takes too long, we just won't do it. And through agentic coding, more and more companies are now getting the idea, for example, to reduce their large software licenses, save costs, build their own apps. Many will also approach you with an MVP and have already thought about how they imagine the frontend, but hardly any company will manage to implement this app themselves from scratch. And therefore, the business model nevertheless offers high longevity, especially because more and more companies are using agentic coding. Therefore, AI app development is truly a top business model.
Number 8 is Corporate LM Setups. This is also a top business model. This means you set up corporate AI workspaces, for example, with Corporate LM. This means you create secure corporate environments in which every employee can work with leading LMs legally and GDPR-compliant, and above all, efficiently through agents, through skills, through proper team management of apps, of prompt libraries, and so on. Profitability is high because every company recognizes the added value of such setups, often not knowing how it really works. Often, it fails simply at the point of preparing the knowledge for such corporate LM workspaces. Nevertheless, the added value is simply enormous. Everyone can easily imagine, yes, if every employee works productively with AI tools, then the added value of that is hard to quantify. The competitive situation is extremely low, as simply no one has this business model on their radar. We introduced this in Germany at all, this whole term Corporate LM AI Knowledge Management, because in the USA it's not such a big topic, as they can work with all the US tools anyway. In Germany, however, this doesn't work. Furthermore, we have a much bigger problem in knowledge management than companies in the USA, because we have largely not yet fully implemented digitalization in medium-sized companies. And precisely this is what you solve as a Corporate LM Setup provider, and you can also sell such setups for companies as a partner, for example, as a verified partner of corporllm.de, with best practices, with instructions, so that it's easy for you too, meaning you are guided and can implement it in everyday company life with a click. Therefore, this business model is also extremely long-lived. So, companies will continue to have a need to use AI productively in their companies for the next few years, as it simply continues to develop. So, ongoing support, ongoing workshops are what we always see when you have set up a Corporate LM, that you say, we have a workshop once a month, explain individual areas, how you work with skills, how you work with agents, then there's something new in the AI world again, then employees need to be trained again. This means you can relatively easily stay in the game if you have set up such a Corporate LM once. Therefore, Corporate LM Setups are truly a top business model.
Number 9 is RAG and AI Knowledge Management Systems. This is relatively similar. This business model is good. This means you primarily help with the preparation of documents for AI knowledge stores. So, Corporate LM would be the next step or a slightly easier entry. With AI knowledge management systems, you are often very niche and try to process technical drawings, PDF scans, thousands, tens of thousands of documents for vector databases, for example. With this business model, you need to have a relatively good technical depth. However, if you have it, you will have enormously high profitability here, meaning companies gladly pay you between 10 to 30,000 Euros for an AI-ready preparation of knowledge and documents, just for the preparation of this knowledge for the AI age. Demand for this is also increasing enormously. Many companies understand that this is the basis for almost every further project. So, you need data preparation for, for example, voice agents again, and therefore it is also a very good entry-level path that every company out there needs. The competitive situation in this area is currently very low. So, here too, I hardly know anyone who has really specialized in it, as many business gurus, as I said, in the USA, but also in Germany, simply don't know this business model. We only came up with it after we had accompanied thousands of medium-sized companies in AI implementation and did this repeatedly, meaning voice agents, for example, preparing this knowledge first, and through this, we increasingly realized how big this problem actually is in Germany. And when we started to address it very specifically again and again in hundreds of conversations that we have every week with companies, it really became clear to us, wow, the demand for it is enormous. So, of course, every company understands how important it is to prepare this knowledge, and therefore the competition there is currently very low. It will probably increase soon. And the longevity is relatively high. So, this will also not go on forever, of course. At some point, every company will have prepared its knowledge. So, it should really be understood as a setup and as an entry point. And this will certainly continue to work in the coming years. For 5 to 10 years, you will of course have to position yourself more broadly.
Number 10 is AI Video Creation. This is a top business model. This means you create video ads, creatives, explainer videos, and commercials for companies using AI tools. A top business model because profitability is high. You can sell such AI commercials like traditional commercials for thousands to 100,000 Euros per film. And the added value for companies is crystal clear. They don't need a large camera crew, they don't have a shoot, they can relax. Yes, you essentially take over the entire production and only have a few costs for the AI tools, which are not too high if you use the right tools and not standard solutions like Hixfield, for example, but if you delve a bit deeper into the topic, you will find many ways to do it much more cost-effectively than what you find on the internet at first glance. The competitive situation is currently still relatively low. So, there are of course more and more people showing you how to create any AI video with a few quick prompts. But anyone who starts thinking about how to create a professional commercial will quickly realize that it's not as easy as it often looks on the internet. You first have to build a deep understanding of the brand. You first have to think about what works with which tools, what looks good, what would fail in practice. Even the first ideas that ChatGPT and Claude spit out, you often can't implement them as a film. If you don't believe me, just try it. You will find exactly that for each scene, meaning there are now very successful commercials that are often in the six-figure range, and the biggest costs in the end are the tokens and credits, because for each individual sequence, it often takes dozens, if not hundreds, of attempts until that sequence is really good. And that is precisely why the competition is not so high, because most people can create a quick reel, but not really professional commercials, and the longevity of this is also very high, because these films are naturally getting better and better, and it will become the absolute standard in the coming years that hardly any real videos will be created anymore, and only AI commercials will be created. And how many large AI video agencies do you already know? There are hardly any providers who have positioned themselves so broadly in this area. With Cinetic AI, for example, we are currently the largest in Germany, where you will also find the largest German-speaking YouTube channel on the topic of AI videos, and this ultimately shows that there is still a lot of potential in the entire market to position yourself as an AI video agency accordingly.
Number 11 is AI E-commerce Shops, an absolute flop business model. This means you run e-commerce shops and brands with AI. You let AI run the shop, you let AI give you the idea, you let AI create the creatives, the ads, and this is a flop for several reasons. Firstly, because profitability is extremely low. So, even the largest e-commerce shops and people you know might have a margin of 5 to a maximum of 10%, and even if you eventually achieve millions in revenue in this area, it's ridiculously little compared to what you can earn with the other business models. Competition is extremely high. Here too, you simply have to look at the Asian market, which will naturally occupy every single niche, as well as the large providers like Amazon and so on themselves. So, it would be completely naive to assume that Amazon and all the large e-commerce shop operators will not simply pick the best products themselves, run ads on them, and then completely outplay you with budget. So, this is a race that is very, very difficult to win, especially if you are just entering the market. And therefore, longevity is also not high at all. Currently, there may still be windows in individual niches, especially if you are active in very, very strong niche segments. We also have a few e-commerce shops among our clients. However, these are less AI e-commerce shops and more e-commerce from the old world. For example, we have a client who has specialized in smoke detectors for decades and is the top 1 shop for smoke detectors, including for chimney sweeps, and has hardly any competition in this area. So, niches will always exist, but as a beginner and as a career changer, you can completely forget this business model.
Number 12 is AI Workflow Integration with Make or Zapier. You implement the classic boring recurring processes in companies. This business model is good, and you don't even have to use only no-code/low-code platforms like Zapier. The trend is increasingly towards agentic workflows or even building workflows with agentic coding systems and implementing the workflow with Python scripts, for example, using solutions like Trigger.dev, Make, AWS, Azure, and so on, to implement process optimization. The business model is very good. You have high profitability, especially once you have implemented such niche workflows, you can replicate them for almost any other customer. This means the major effort often arises with the first customer, with the second customer, you have hardly any effort. Especially if you position yourself in a niche, you can often sell them in the four-figure to low five-figure range per workflow. Competition in this area is medium; this is the classic business model that many have used in recent years, including many who haven't really understood agentic coding or AI yet. However, you can also catch up more and more here. So, if you, for example, work on your Zapier workflows with agentic coding, you will outperform many other providers who still believe this doesn't work. So, there are indeed many who believe this doesn't work, and this naturally presents a great opportunity, especially regarding profitability. Longevity is also good, because there will always be workflows in companies, and there will always be people needed who automate these workflows meaningfully for companies.
Number 13 is AI Chatbots. Chatbots are an absolute flop business model. Yes, these are the main claims you see again and again on social media. Hey, take the website of a roofer, put it into this AI tool, and create a website chatbot for them and sell it for a few thousand Euros. This means it seems relatively sensible at first. In practice, however, you will find that this business model fails radically at many points and comes from people who don't sell it themselves. Profitability in this area is extremely low, because you are in absolute comparability, as competition is very high. So, this is the first thing most people think of. I'll just make a website chatbot and push it through some AI tool, and therefore there are hundreds, if not thousands, of providers offering such solutions. Some at bargain prices, some for free. So, it's hard to make really professional money with it. And moreover, it is also one of the first solutions that companies simply do themselves, that companies quickly come up with the idea themselves. Here too, there are niche applications for everyone, of course. So, if you implement chatbots in a Corporate LM, for example, then it is of course a completely different business model. But these classic AI website chatbots are now outdated and not a business model I would start with today.
Number 14 is Personal AI Assistant Setups. This is a top business model, something that many people haven't considered yet. What does this mean? You set up AI agents in an enterprise version or infrastructure for companies, and especially for managing directors, because many managing directors want to use tools like Claude Code. They may have seen frameworks like Hermes or Open AI, but they don't dare to really bring it into the company and use it for their own daily work, for example, for managing emails, for calendar management, for retrieving information from document storage or storing it there, and simply being able to communicate with a central instance, like a chatbot via Telegram, or via Microsoft Teams or via Slack, with access to the company infrastructure. This could, for example, look like setting up Claude Code on a VPS for a managing director, and the managing director can then use Claude Code, for example, via Slack or Teams or Telegram, in a sandbox or in an isolated infrastructure, so that there is hardly any risk, apart from the few things you should consider from a legal perspective. But all of this is solvable. Profitability is very high because the added value is simply enormous. So, imagine, as a managing director, you would need a personal assistant, which many have had in recent years or perhaps haven't found yet, for whom you would certainly pay 4, 5, 6000 Euros gross per month if you want someone truly competent who takes care of all of this, and AI agents can do all of this today. So, the added value is enormous. Therefore, companies are also willing to pay a lot for it. Competition is extremely low. So, I haven't seen anyone to date who explicitly calls this a business model. We see it repeatedly in customer projects, where we simply set this up for our customers in addition to other projects we are doing. And for me, this is also a clear trend. So, we are still in the early stages of OpenAI and all these things that don't yet function really productively for a company, but we are getting there soon, and managing directors want to use such solutions even now. Competition is practically non-existent, as competition is hardly or not at all existent in the German-speaking market. No one has addressed this yet, and hardly anyone sees it. And the longevity of this is also quite high, because managing directors always need personal assistants, and you often can't have enough of them. So, definitely a top business model.
Now I have a bonus business model specifically for the German-speaking market for you. This is the tuning of Microsoft infrastructures and Microsoft Copilot tuning. Companies often already pay for Microsoft 365. Some may have even booked Copilot, but are not even remotely efficient with it. So, I always like to compare Microsoft to a Renault Twingo if you use it standardly, and now you help companies to at least drive a standard Mercedes by setting up Copilot setups, for example, to use agents in Outlook, agents in Microsoft Teams. So, there are many low-hanging fruits in Microsoft that many IT employees themselves don't even know or understand. And since 85% of all German companies work with Microsoft, there are many low-hanging fruits here that you can use to create real, tangible added value for companies, and therefore an exciting business model, a good business model for the German-speaking market. Competition is very low. Here too, there are hardly any who explicitly name this. However, in my opinion, it is not too long-lived. So, you should position yourself more broadly and not put all your eggs in the Microsoft basket, but it is certainly a business model that you should keep an eye on.
Now, please let me know in the comments which business model was most exciting for you, and with which you might take your next steps. And if you now want to know what's important in the next step, meaning if you have decided and want to win your first customer in this area, then definitely check out this video here. And if you want an overview of all industries and niches that are currently most exciting for the German-speaking market as an AI agency, then feel free to look here. I look forward to seeing you there again. Until then, take care, Leon.