If you're thinking about starting an AI side hustle or building an AI business, watch this [music] before you start because the easiest way to make money with AI may be the worst business model to build. I learned this as a CEO and board [music] member at tech companies that created billions of dollars in value. And they all say the same thing.
Speed matters in the world of AI, but only when it helps you build something differentiated and defensible and figure out how to build a [music] strong business that can make money for years. 40 years ago, if you wanted to do a television show in America, there were basically three doors you could knock on. ABC, NBC, and CBS.
Now, if all three of them said no, your show basically didn't exist. If you have a phone and a Wi-Fi connection, you can launch your show on YouTube today. You don't need anyone's permission.
But at the [music] same time, by the time you wake up tomorrow, another 20 million videos will be uploaded on YouTube. [music] So, YouTube makes creation and distribution easier, but what becomes harder is your ability to make something valuable. And AI is doing the same thing to intelligence and execution.
This makes building a side business easier. But AI itself is not your edge because your competitors have the same tools. Low startup costs, great.
Low differentiation, fatal. That's the artificial illusion. Easy to build doesn't mean [music] easy to win.
The real question is what should you build that's harder to copy? What's that place where your AI side business can succeed and thrive? To find that place, you have to pass through four gates.
I call them deal [music] gates. We will pass through each of those gates and bring each one to life with a different AI business example. Let's go to the first gate, demand.
How do you know you have a good business idea when somebody starts paying for it? In 2006, Elon Musk was trying to do something that sounded insane. His company wanted to build millions of electric cars, but he did not know if there would be any [music] demand for even one of them.
Almost nobody believed that electric cars could compete with Toyota or BMW. So Tesla unveiled the Roadster and offered the first 100 cars as a special signature series. Every buyer had to put down a $100,000 deposit and they didn't get a car.
All hundred were sold out in 3 weeks, but not a single one was manufactured yet. So by August of 2007, Tesla had pre-sold 570 Roadsters. Months before manufacturing even began, he had an answer to a question most founders don't even bother to ask.
If I build it, who would pay and how much? Your customers have a vote. They have a veto.
And it comes down to the same thing. Will they cut you a check? If nobody pays, you should be willing to change the offer, change the customer base, or even change the idea.
And that's why a lot of us build AI side businesses backwards. Someone shows you that you can build um a website with nothing but a prompt and it just takes 10 minutes. So you think, "Oh man, I I can build an AI website business.
" But all you've done here is proven that AI can make a website. You have not proven that anybody wants to pay you to make it for them. Why wouldn't they build it themselves?
So AI can supply solutions for you but it cannot create demand for you. That you have to do yourself. Now how do you do that?
How do you verify demand? You can do three things. First pick narrowly the niche of the niche.
Don't choose a solution for small businesses. I mean that doesn't exist. Choose a solution for dentists with less than three locations in three towns around you.
or pick independent recruiters or Shopify brands that are doing under a million dollars in revenue, something like that. Because the narrower the customer, the easier it is for you to understand what they actually need versus what they think they want. Second, talk to buyers.
Don't interview your friends. I used to do that. They will tell you what makes you feel better.
Your actual buyers will tell you whether there is a pain and where is that pain and what gets in their way and if that problem is painful enough for them to pay for it to disappear. And third now build the smallest test. In the tech world they call it the MVP, minimum viable product or service.
Use AI to build the smallest working version and put it in the hands of your future buyers. Now you can tweak it, change it, test again, keep going until someone figures out that this is valuable enough and they'll pay for it. Now let's take a real world example like AI bookkeeping.
Now here's the competitive dynamic that you will see with almost every idea that you'll come up with in today's world. You're never going to be able to enter a completely empty market unless you're super lucky. For example, in bookkeeping, QuickBooks already owns a large portion of the market, at least in the US, and all these AI native startups are coming after it.
So, your opportunity is not AI bookkeeping. It's a narrow customer with a problem that the existing tools can't solve. I'll give you an example.
After I retired or graduated for full-time work, I was advising private equity firms and serving on boards. And none of the offtheshelf bookkeeping tools quite fit what I wanted to do, what I needed. So, I ended up hiring an independent bookkeeper who then built a custom process around my workflow and used AI to help build it, help run it.
That's the kind of hidden demand that you're looking for. So, I would narrow cast a lot. maybe independent consultants like me or maybe laundromats or something very small and talk to 10 of them and find out what QuickBooks or the AI startups or their current bookkeeper doesn't solve.
Then build the smallest version around that particular pain. The cost of running those kind of tests has collapsed. You can research faster, build a rough version in days.
So AI changes the economics of entrepreneurship completely. But choosing the customer, earning their trust, reading reactions, reading the room, deciding what matters to them, to you, all of that, that's the hard part. And that's still yours.
Now, let's enter the second gate, the edge. Your edge comes from many places. Let's cover three.
Imagination, focus, and context. Let's take imagination first. Many years ago, there was a street performer in Canada.
His name was Guy. So, he would breathe fire, played music, and he had this one interesting dream. He wanted to start a circus.
That was a startup idea. The problem was circus was a fading industry. Kids were going to theme parks instead of watching animals do tricks.
Traditional circuses were losing money or shutting down. But Guy imagined something completely [music] different. He combined circus with the idea of a Broadway show.
So he would bring incredible gymnasts, dramatic lighting, original music, storytelling, no animals. And that became Cir to Sole. Today, CIRC generates roughly a billion dollars in revenue every year.
Sometimes [music] your edge comes from imagination. Sometimes it comes from obsessive focus. I'll share my own story.
When I became CEO of a company, our company had three possible paths for growth. And two of those markets were much bigger than the one we were already in. So it was very seductive.
It was very tempting to build something for those markets. But instead of spreading ourselves across three opportunities, we decided to double down even further. One market, one problem, one product, one type of customer.
And you know that focus paid off within a few years. The company had a great exit. And finally, sometimes your edge will come from context.
Andrew Ang is one of the expert voices in AI and he calls this context advantage over AI. So you [music] have knowledge our relationships our conversations our mistakes our workflow our nuances and that is something [music] AI will not have. Okay but now what if you are 26 and you don't have tens of years of industry experience which means that you don't have that context.
Well that's okay. you can still build context fairly quickly. And it's the same trick.
You go super narrow, get close to people who live the problem, people who have the context, and learn obsessively. Let's say you're passionate about photography. That's great.
You can go in so many directions from there. You could build an AI service for e-commerce brands or you can build an AI workflow for real estate. Or you can have an AI solution for wedding photographers.
All three of them are three very different problems that you can solve. But where is your edge? Now if you take wedding photographers for example, then you have to talk to a lot of them and learn where their workflow breaks, which moments matter most to them.
Where do they lose money? Where do they lose time? What do their clients care about?
And what happens during the wedding season? and what do these guys do when the wedding season is over? All these questions when you go deep into them, that's what creates an advantage that others don't have because that's the context that tells you what to build.
AI can research the niche. It can analyze workflows and build you prototypes literally in one day. And your job is to create that edge through your imagination, your focus, and the context that you can create.
So you can decide what good looks like. Now let's pass through the third gate access. Every product needs to match its distribution channel.
Dollar Shave Club is such an interesting story about distribution. They were selling razors and [music] blades to men. But Gillette was the 800 lb gorilla in that category.
[music] So fighting them in the supermarkets or in drugstores, I mean that made no sense. So Dollar Shave Club found another route to [music] distribute their product direct to consumer. They came up with the idea of subscription-based service.
[music] So you pay every month and then every few weeks or something you get your razor and cartridge delivered [music] to you. And 5 years after launching, Uni Liver bought that company for a billion dollars. That's the power of finding the right path to your customer.
And sometimes your customer decides what the distribution channel would look like. If you're selling software to laundromats, well, the channel might be Google maps and cold calls and literally walking into these stores. The question is where do your product and your customer naturally meet?
All right, let's walk through a real world example. Let's say your AI side business idea is an AI interview simulator. So you could build an AI interview agent that studies the resume and job description.
It looks at the company, the culture and becomes the interviewer. And it would change the questions based on the answers received. But here's the hard question.
Where do you find the customers? You can't go hunting for job hunters one by one. I mean, they're scattered everywhere and you don't know where they are.
But maybe you'll sell through a university career center or coding boot camps or recruiters. Maybe figure out what to do with LinkedIn outplacement firms. There's so many channels now.
Now you started thinking about how to distribute your product. And you'll also notice something interesting that the person using your product may not be the person who pays for it. The job seeker may be the user, but a university or a recruiter or an employer might be the buyer.
But try to find that one relationship that can get you to hundreds of these users who already need what you're trying to build. That's access. Now AI can identify these potential channels for you.
It can research partners and even build prospect lists. But your job is to pick the most effective distribution channel and build that partnership that opens many many doors for you down the line. So if you see that path clearly, then you have access and if you can't well you may have a very good product but you don't have a business yet.
All right. So now let's go to the fourth and final gate loop. Every Formula 1 driver does countless practice runs.
The driver goes out, runs a few laps, comes back in. [music] The engineering team will pull data from hundreds of sensors. They will compare [music] it with what the driver felt, make a lot of adjustments, small adjustments, and then send the car back out.
Then they measure again. Their goal is simple. Every lap should make the next lap smarter.
Your AI side hustle should work the same way. Every customer interaction creates data, right? Every sales call, every support ticket, customer reviews, customer interviews, there's lots of gorgeous insights sitting right inside your business.
Just take all of that and feed it into claude or chat GPT or Gemini and AI will find patterns that you would never be able to find and it will build these learning loops from that data for you. Your loop must make your business more efficient every time it runs. And that learning loop is the most critical thing you can build because [music] the world is already beginning to look like a high-speed car chase to me at least.
Things are changing so fast that your advantage in this lab may not last in the next one. So in that world, we all have to start working at the speed of AI. You can outsource all the grunt work to [music] AI but you cannot outsource learning and thinking that stays with you.
So those are the four gates demand edge access and loops. You know in the world that we're heading towards builders are going to win. So you should always ask how do I cultivate that mindset of a builder?
The judgment that you need to build that you want to build will be built by constantly tinkering and struggling through your AI side business itself. And that benefit will not show up on your balance sheet. Here's one way to think about it.
If I gave you a piece of wood and asked you to make a violin, you would have no idea where to [music] begin. Neither would I. But there is a city called Cremona in Italy where violin making is a sacred art.
The makers typically build only three to six violins per year. They select they shape each piece of wood according [music] to its individual acoustic response. It can take months to make a single violin.
They can pick up a piece of spruce or maple and hear [music] the possibilities in that wood that the rest of us just can't. That comes after years of struggle, building, listening, refining, honing their craft. That's judgment.
It's your struggles that teach you what good looks like. Thank you and I love you.
