AI Agents Explained: $0 to $1m ARR in 58 Days | Build Series #2 | ft. Johann Sathianathen
Invest Like a Boss
July 2, 2026
What if you could tell an AI to "make me a million dollar business" and it actually starts doing it?Johann is 22 years old. He's already built and sold multiple SaaS companies. And in just 58 days, he scaled an AI agent business from zero to over $1 million in annual recurring revenue.
Speakers Johann Sathianathen, Sam Marks
Johann Sathianathen (0:00)
Set one person up, they would go tell three people, we'd be able to close like two to three of them, and it's just crazy snowball effect, network effect. I would say 90% of the time, any business is always asking about security, but, and that's OpenClaw is like the changing thing, because brought up like what AI agents can do, when people are seeing the crazy use cases online. Make a personal website for Sam Marks, you can do deep research on him to make sure it's very sharp, and then deploy it to the internet and send me the link.
Six second voice message, and we'll get a link back in maybe like five to 10 minutes.
Sam Marks (0:27)
That's crazy.
Johann Sathianathen (0:28)
And I'll send you that link and we can definitely get up to sammarkz.com.
Sam Marks (0:31)
Oh, that's insane.
Hey everybody, welcome back. We're on the build series here at Invest Like a Boss. I've got Johann and Johann is an AI prodigy at just age 22 He's already built and sold multiple SaaS companies, and now he started in just 58 days, an AI agent business that has gone from zero to over $1 million in ARR. So stay tuned. We're going to talk about that and talk about all the great use cases that you can deploy AI agents in your life and your organization today. Johann, welcome to the build series.
Johann Sathianathen (1:10)
Sam, thank you so much for having me.
Sam Marks (1:12)
Definitely, man. Let's start simple. First, what is an AI agent? And how is it compared to AI tools that people are most common with, say, ChatGBT or Cloud?
Johann Sathianathen (1:23)
Yeah. So ChatGBT and Cloud are essentially like chatbots. And I think a really good example to understand the difference between ChatGBT and Cloud versus an AI agent is that for the past two years, I've always been asked in ChatGBT, make me a million dollar business, and it would give me a long paragraph, this is how you make a million dollar business.
Versus an AI agent, I could tell it the exact same thing, make me a million dollar business. And it would give me the whole plan, but then it would say these are the things that I could take action on. And it could actually go out there and take action, it can interact with the computer, it could go fill out forms, go send emails, it could do a lot of stuff digitally to actually help you make progress on that million dollar business. So, yeah, agents are very addicting.
Sam Marks (2:01)
Okay, okay, all right.
What's the recent evolution been of these things? I remember when I got here to the city, maybe six months ago, as the first time I heard OpenClaw. Can you tell when OpenClaw was launched and maybe who the other major players are, what the landscape currently is?
Johann Sathianathen (2:19)
Yeah, so I think it was around November 2025 That's when Opus 4.6 came out around then, and there was also something called a Pi Harness.
Sam Marks (2:27)
Okay.
Johann Sathianathen (2:28)
This is a little bit technical, but there's Cloud Code, which developers were using, and Pi Harness was essentially allowing them to be running AI agents on their computer.
Then OpenClaw made it very popular by allowing you to integrate with the Pi Harness with Telegram. So your everyday person can go and use something like Cloud Code as opposed to opening a terminal that's very scary. The average person doesn't know what it looks like.
But Telegram, WhatsApp, Discord, everyone knows that. So they could have that running on the back end. They never know what's running. They're just talking to what looks like a chatbot, but it's a chatbot that's going out there and doing all these crazy stuff. I know that there's people using OpenClaw to scan Zillow, check all the houses and lowball everyone. I don't think they had much success, but cool little use case for it, and it blew up very, very fast. I think it was the fastest growing GitHub repository ever, and everyone was talking about it. Even my mom was calling me and saying, Johann, what's OpenClaw? I think that's something you do. You make OpenClaw? I was like, no, I didn't make OpenClaw, mom, but I had to work with it a little bit.
Sam Marks (3:28)
Yeah. So with an AI chat like chatgbt or Claude right now, even globally, adoption is still very low. In our circles, everybody uses it, of course, but globally, it's still a very small percentage of people that are using chat AIs.
But for the people that are using it, it's quite simple. You pull up chatgbt.com or you pull up Claude and you just talk to it, right? Very simple, like typing into a Google search box. But for the AI agents, how are you communicating with them?
49 more minutes of transcript below
Thousands of transcripts fetched by people building searchable podcast archives
Fetch the whole transcript
The demo key returns a sample episode in full, no card needed:
curl -H "x-api-key: pt_demo" \
https://spoken.md/transcripts/1000651996090Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.
From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.
Using your own key:
curl -H "x-api-key: YOUR_KEY" \
https://spoken.md/transcripts/1000775151653