Transcription
Let's talk about what AI business you can actually start in 2026 if you're not a coder, not in Silicon Valley, and don't have a million dollars in funding.
Over the last 6 months, I've asked this question to founders, operators, and investors on this podcast. People like Reed Hoffman, a legendary investor, Robbie Stein from Google, and Mati Stanisfki, the founder of 11 Labs. And what's interesting because they're insiders in the industry instead of vague advice, they kept pointing out the same thing. Very specific, very practical opportunities.
In the past, to build a real business, you needed to build a company with resources, with funding. I remember building our very first website. We needed coders for a very basic landing page. That was 2011. Now with AI, everything is changing and is changing so fast. You basically have a co-founder in your pocket, a lawyer, an accountant, a web designer, a marketer. All of them almost for free. That's why the opportunities have exploded.
Take Samir Vasada, an entrepreneur, the CEO of Vice. I asked him what changed in his company in the last 3 years. And here's what he told me. We had 150 people, 160 people at one point in time, and now we are closer to 40. And we are doing 10 times better from a metric standpoint with 40 people than what we were doing with 160 people. >> They went from 160 people to 40 and their performance improved. That is not a future prediction. That's already happening inside real companies.
In this video, I'll break down six AI business models from the fastest ones you can start this week to the ones that you can turn into a real product. For each, I'll talk about what it is, why it works now, and how you'd actually get your first client. But before we get into idea number one, all six of these businesses, you'll be building them by talking to AI. So, let me show you the one tool I don't work without anymore.
Thanks to Whisper Flow for sponsoring today's video. One of the ways AI actually gives me back time and made my process so much easier is dictation. The thing is, I don't like typing and I'm slow at typing. And when I'm writing a long prompt on Claude to build something, I don't want to sit at my laptop for 2 minutes crafting it. I just want to say it. And as Ellie Miller said, the best way to talk to your AI is to complain. And it's so much easier to complain with your voice. So the biggest game changer for me was Whisper Flow. Let me just show you what I mean.
Okay. I feel really stressed every single morning and I want you to make me a morning brief. I want a full research related to my industry and I'm a content creator running a YouTube podcast about AI and I want you to pull the recent news and press releases uh and summarize you should measure the top three based on what's going on to impress my audience so I can post about it the next day. The second thing I want you to pull most insane AI stories related to my industry. And at the bottom of this recap, uh, can you also go ahead and add three fun events happening in uh, Los Altos in the next four days? And I can actually change my mind mid-sentence. It's going to catch it. And of course, I drop stuff. I say m and it took me what 15 seconds instead of 2 minutes typing.
So basically, Whisper Flow is a voice-to-text tool that actually works in every app on every device. You can use it in Claude, ChatGPT, Perplexity, inside Slack, Gmail, or iMessage. Basically, anywhere you type. And it's not like the voice dictation on your phone, and I speak Russian to my AIs. Only Whisper Flow can make it happen seamlessly. Built-in dictation captures your mistakes. Whisper Flow cleans them up. Built-in dictation gets tripped up on names. Whisper Flow learns them from you. It also works on Mac, Windows, iPhone, and Android. And Whisper Flow is free to try. You can download Whisperflow using my link in the description. If you use the promo code Marina, you get an extra month of Whisperflow Pro for free.
So, let's start from idea number one. Something that I've been talking about a lot. AI consultant for your industry. Complexity: easy. You don't need stack product. And actually, when I'm thinking about recent graduates who are graduating into our market with entry-level jobs disappearing, this is your sweet spot. You are growing up in this AI world. You are the one who has all the time in the world, all the energy in the world to try out those tools, to research them for a particular industry and start consulting. And you don't need anything to start it. You just need to prove that you understand AI and can solve real problems for businesses with AI.
Just look at this number. Why Combinator, the legendary accelerator here in Silicon Valley that funded Airbnb, Stripe, and Dropbox, just bet $36 million in one quarter on AI agent startups. Roughly 60% of every YC batch is now AI native. The companies getting funded are building the tools. Someone has to deploy them into the real world. And great for you is that most companies still have no idea how to use AI agents in their business.
I asked Reed Hoffman this: if you're making an average salary, say 60 to 80K a year in the US right now and you can translate to whatever it is in your country, what do you do to double your income? Listen to the whole thing. >> I think this year is when we really begin to see more of the applications and so like how do I run my business better? How do I do um like my analysis of my supply chain or my financial analysis or my risk analysis or my marketing or my sales? How do I do any of that stuff better? And they're going to start looking at it. And sure, they'll go to some other internal people, but everyone's going to start going, well, I've been doing this job and adopting the jumping to the totally new thing is hard for me. So part of it is like, okay, start jumping to the new thing. >> Love it. So basically the business idea or the double income idea, you get proficient in tools in some area and go help other >> and demonstrate it. Yes. And demonstrate it. >>
The key word Reed used was "findable." That's the entire game right now. Companies know they need AI. They just don't know who to trust, who would be that person who will introduce this world of AI to them. You make yourself easy to find by posting proof. So here's your action step. This is the idea I'd honestly pick. And if I was 21 years old tomorrow, I'd start with this. I'll pick one function I already understand. Say marketing, finance, I understand, study abroad, operations, HR, language learn, whatever it is. Spend the next 30 days posting actual case studies with screenshots and numbers on LinkedIn. Like, here's the process. Here's what's broken. Here's how I fixed it with AI and now it only takes five minutes a day. In the end, here's the cost before and after. Your goal isn't to go viral. It's to show proof to your future customers. And if you do this consistently, your first client will usually come from someone already watching, not from cold outreach.
Idea number two, SEO for local businesses. Again, very easy in terms of complexity. You don't need tech stack, just AI search tools you already use. 33 million small businesses in America rely on Google for customers. Now, their customers are asking ChatGPT, Gemini, and Perplexity, "What's the best dentist in Austin?" "What's the best mechanic in Denver?" The businesses showing up in those answers aren't the ones paying for Google ads. They are the ones mentioned in articles, directories, and blog posts. In fact, top three resources that feed our chatbots are Reddit, LinkedIn, and YouTube. So, a lot of businesses already understand that their clients are migrating from just search to these generative engines. You can be the one who helps them get customers from this new medium that we're using for search.
I sat down with Robbie Stein, VP of product at Google Search, and he explained clearly how it works. So what should I focus on right now to be recommended by and actually my business is recommended by which is because we were doing a lot of content but maybe for you know uh some segments of my business that are not recommended what should I double down on for AI to consider me? Yeah, interestingly the AI thinks a lot like a person would in terms of the kinds of questions it issues. And so if you're a business and you're mentioned, you know, in um you know, top business list or from a a public article that lots of people end up finding, those kinds of things become useful for the AI to to find, you know, >> invest in your PR. That's something I've been hearing a lot. >> So it's not it's not really different from what you would do in that regard. I think ultimately how else are you going to decide what business to go to? Well, you'd want to understand that. But also like sometimes I invest in PR and I ask my friends, "Have you seen that article?" And they're like, "No." But then I ask AI and it really sees the article and it uses that information. So now you're investing in PR not for people to see it before AI. >> That's actually a good way of thinking about it because the way I mentioned before how our AI models work, they're issuing these Google searches as a tool. And so in the same way that you would optimize your website and think about how do I make helpful, clear information for people. So people search for a certain topic, my website's really helpful for that. Think of an AI doing that search now. Yeah. And then knowing for that query, here are the best websites given that question that's now coming will come into the >> uh context window of the model. And so when it renders a response and provides all of these links for you to go deeper, that website's more likely to show up. Yeah. >> And so it's a lot of that standard best practices around building great content really do apply in the AI age for sure. >>
And Robbie is right. I'm seeing it in my own media business right now. And then I went and Googled my podcast couple months ago and it wasn't really showing anywhere and now we started working on it and we just asked Claude to create us a strategy. I have a person talks to Claude who executes everything. She has no background in SEO but we're already rising in different requests that people who come from those LLMs, they are converting much faster. They're just in general easier clients to work with. When it comes to our email newsletter, we noticed that sometimes ChatGPT recommends it and people who subscribe through ChatGPT have 80% open rate, like, and this is really good compared to 40 to 50% that we're getting on average.
So the action step for you this week: pick three local businesses in your city or maybe the area or maybe type of businesses that you know. Go to ChatGPT, Gemini, and Perplexity. Ask each one, "What's the best dentist in your city?" Maybe you want to promote a podcast. Ask about podcasts. And if the businesses you've selected don't show up, that's your pitch.
Idea number three, voice AI receptionist. This is actually medium complexity, but it is so interesting. You will need some basic tech stack, which is not hard to learn, and right now there's a massive gap between this technology existing and businesses actually using it. Ask Matty Sisfki, CEO of 11 Labs, the biggest voice AI company in the world, what he'd do if he had to start from zero and make 10K a month fast. >> It will be voice specific, but I think it's so, so early that I think it's it's a huge one is u there's definitely a lot of the infrastructure being built for the voice agents. We we we build it, but other companies are are too. Um, and I think there is a big gap between voice agents and then actually deploying them in a lot of those businesses and you don't have to have the engineering expertise to deploy those voice agents. The platform now frequently will support a relatively self-served manner of taking it but you can easily take that voice agent and deploy that in a specific domains and most of the businesses in the world still don't know don't not know about it. If it's, you know, not um venture scale uh business and you just want to make good money, I would try to take those voice agents and go to um let's say local doctor's office and help them appointment schedule for for for the dentist so they can take appointments more easily and they can then focus more on their work instead of nurse doing that in between or missing appointments. That's actually one of the most common. I don't know the percentage, but so frequently those those appointments don't get booked because there's no one on the phone and can take them. Uh you can go to local mechanics and help them take appointments. And I think there's all of these require slight variation of the domain piece that you need to know and all of those businesses are in thousands to tens of thousands of dollars per month if you get it to the few. Um, their infrastructure is there. You just need to bring it to to to those domains. >> Yeah. It's like B2B automate businesses with AI. >> Yeah. And small businesses all >> you don't have to be a coder. >> You don't have to be the coder. You just need to spend the time call them and ask or or or go to them. Um, and I think there's like this category which might not be um taken off by by some of the biggest companies that will focus on bigger enterprise uh elements. So like you know classic uh uh this is like small medium businesses rather than than than enterprise segment and at the same time most of those companies just don't know this is possible. So like next year two is just an incredible opportunity to do it and of course, you know, start in English speaking but I think the same is true for so many of the of the countries and languages which which might be uh given so much of that work isn't always localized. I think in our case uh we're doing a pretty good job there. You can you can bring it to a local market and do exactly the same the same work. >>
So here's the action step. Pick one vertical this week. Dentists, lawyers, chiropractors, whatever you like the most. Find 20 offices near you on Google Maps. Then call each one during lunch hour and count how many goes to voicemail. That's your pitch. "I counted. You missed that many calls this lunch. I can fix that for $500 a month." Do a couple for a testimonial and then start charging.
By the way, guys, we work really, really hard on this channel. So, if you're finding this valuable and you like what I'm doing, please subscribe to this channel so you don't miss the next episode. We just hit 1.5 million subscribers, which was a dream when I started this channel, and I'm excited to see how big this can grow.
Okay, idea number four, AI-native ad agency for local service businesses. This is medium-hard complexity. You need creative stack. You need copywriting. You need media buying skill. In this idea, you're adding creative skill on top of tech. You'll run ad campaigns for clients, not just set up tools for them. You can go to real estate agents, med spas, dentists, independent gyms. All of them want more ad variation. Their current agency might be charging them thousands of dollars and delivering only a few ads per month. You're going to offer hundreds of variations for less money, generated end-to-end with AI.
Alex Mashrab, founder of Hicksfield, the company doing $200 million a year, told me exactly when it happened. Listen to this. >> Everyone was missing these camera controls. So that's what we delivered March last year. Then in April we delivered a library of visual effects. And then I think in June the industry completely changed. We saw the emergence of AI-native marketing agencies. So essentially those agencies they completely go end-to-end with AI and very often they try to bypass incumbent tooling like for example Adobe or something else and go end-to-end with AI. On the one side they are very limited cuz AI capabilities back in June were a little limited. On the other hands, they drastically improve their margin profile and they kind of show their clients that they can build ads within days. A lot of brands actually want to have constant content flow on their socials and they want to embrace AI. And then from June to December last year, this um this new industry of AI-native agencies completely exploded. And you want to be in a growing market like that because that means you're going to be growing. And it's not a future trend. It's the market that already reshaped itself while most people weren't looking. And most local businesses in your city still haven't hired one of these agencies yet. They're still paying thousands their local agency.
Okay. Idea number five. A bit harder than the ad agency because the creative bar is higher. You're not running ads. You're producing content at scale, consistent across dozens of videos per month. And again, some fun facts. By spring 2025, nearly half of YC's entire batch were vertical AI agents. And in their 2026 cohort, AI-native and content generation startups now dominate. Here we're talking about content generation for e-commerce. The tools are getting funded. Hicksfield is one of those tools as well. The agency's market they use those tools to serve real brands is still wide open. Ideally, go for customers who can pay, right? It could be skincare, supplements, pet products, kitchen gadgets, fitness gear. Every one of them is burning money trying to keep their TikTok shop and Instagram ads alive. They constantly need fresh UGC, the handheld phone style videos, 30, 50, 100 times a month. Human creators charge $150 to $500 per video. And I'm not saying that people have to stop working with UGC creators. I have been creating UGC content for some time, but also as a business owner, I do realize that with humans, we can test maybe 20 creatives a month because they're filming, we're testing. What we use AI creative for is to figure out a script that's working for us because we can generate thousands of different videos, launch them, do the A/B test, and we're like, "Oh, this is what actually working." And then I go to my amazing creator friends and be like, "Let's try this creative because it's working already." And in terms of the numbers, the sky is the limit. You need to figure out the process, but it really speeds the process up.
Alex Mashrab, whose company makes the technology this is built on, dropped this during our interview >> and we just asked them very simple question like, "Did you see this video?" This was a cool generated video. And they said, "Oh no, how is this possible? What's the cost?" And we say it actually cost maybe less than $500 to make this video. >>
So here's your action plan. Pick one product category. Skincare, supplements, pet products. Language learning apps. Make five sample UGC videos using their product images. DM the founder on Instagram. "I made this for you. Here are five for free. I tested them with Claude. If you like them, I can keep delivering and set a price. I don't know, 100 for $3,000" and see what they reply.
Idea number six, vertical AI product. A lot of people call them GPT wrappers. And this is something that a lot of VCs were funding a lot last year. It's less of a hype in the VC world right now just because anyone can now build a GPT wrapper. But that means the whole market is much more approachable now. In this case, you're going to be building a software product, and it's actually much easier than it ever was. So some people when they hear wrapper think it's just ChatGPT with a different logo. Why would anyone pay for that? And Daniel Priestley, UK Entrepreneur of the Year, gave me the cleanest mental model for why that's wrong. >> The assumption is that there is value in knowing more than the customer about a particular thing and you collect a bit of their data. You mix it with your amazing prompts that you've put a lot of thought into. You then put that into a new UX that's better than a chat UX. And between those three things of capturing data, prompting, you know, a set of much better prompts and also thinking about the UX more deeply, you're now creating something that is a specialized tool that's very valuable. And a good analogy would be that you know 100 and something years ago they created electricity and the electricity was this awesome powerful thing but then people came up with applications for electricity and they said, "Oh, we can channel this electricity and make a toaster." Uh, we can use this electricity to boil water and we call it a kettle. Um, oh, what if we do this electricity into light bulbs? So they channeled the electricity into these specialized applications and then those businesses became amazing standalone businesses. Now imagine someone came up and said, "Oh, that's just an electricity wrapper," >> right? You're just you're just wrapping electricity. >> You know, it's like, "Yeah, I am harnessing electricity, but we're harnessing it in a very particular way in a user experience that's perfect for boiling water and perfect for toasting toast." So LLMs are the new electricity. They do this amazing language generation or image generation or or multimedia generation. And then it's up to us as entrepreneurs to say, can we improve upon that through UX, through better prompting, and through better ways of capturing user data. And if we can do those three things and then wrap that up together, we can create these amazing businesses. Some of these businesses, look, a lot of these businesses might end up doing 4 or 5 million a year and they might have 50% margins and you might make 2.5 million a year with a like a very automated little SaaS beta AI-enabled SaaS. You know, there's just thousands of little opportunities where you could create a 5 million a year ARR business. >>
If you hit this right, this has the highest ceiling in terms of income because it scales without you. For example, there's a company called Chestnut, an AI-powered mortgage lender. Bitboard, AI for healthcare operations. Trapeze, AI for healthcare call centers. Here, each one picked one industry, one workflow, and wrapped AI around specific prompts and a custom interface. That is the GPT wrapper. The ones that I mentioned are much more complex, but there are also some that are easier to build, like we built a product where you upload your resume and it fills in your LinkedIn profile. It didn't take us long. What's important was that we had a person who understands how to run LinkedIn and how to create profiles that are going to get attention. So, she unloaded her knowledge into Lovable and Vibe coded that app.
So, here's your action plan. Pick an industry you know where you have some expertise. Write down three repetitive tasks people in that industry do every week that drive them crazy. Pick the most annoying one. Wrap a model around it with better prompts and a cleaner interface than generic ChatGPT and build it this weekend in Lovable. Test with 10 real people in that industry and see if they're willing to pay. If three pay, you have a business.
Those were my six ideas from easiest to hardest. All of them are real. Some of them I'm trying by myself. All of them are in motion right now. The window to start is open. You don't need to be a coder. You don't need to be a research lab. You don't need millions of funding. We live in this most amazing, exciting time. Pick out one of the ideas, start building, start playing with it. And yes, you might fail. Not every idea fits everyone, but you're going to learn a lot of things about yourself, what you like, and what's the best business for you.
And if you're thinking, "Okay, this is amazing. I want to start building. What tools do I use?" Listen to this conversation with Ken Katan Furush, Stanford AI expert. We talked to him a lot about AI skills that you need and AI stacks that help you develop those AI skills. Thank you so much and I'll see you in the next video.