How Greg Isenberg would make a sass out of AI.

If I had to grow a SaaS to $100k/month from ZERO today with AI, this is exactly how I’d do it:

I’d start by picking a niche I deeply understand. Maybe it’s ghostwriters, short-form video editors, newsletter operators. People who have recurring, painful, time-consuming tasks that AI can automate. I’d talk to 5 of them, figure out where they spend time and money, and write down one sentence: “I help [this person] do [this painful thing] in 10 minutes instead of 10 hours.” That becomes the positioning.

Then I’d build a rough prototype with something like a Replit, Bolt or Lovable. I’d optimize around one killer feature but then also a brand that feels really warm and fuzzy. Probably scan Dribbble for someone awesome or do it myself. It should feel like magic the first time someone uses it.

Once that prototype lands, I figure out what a customer is worth to me. Maybe $300 a year. That tells me how aggressive I can get on the revenue share side. I’d happily give up 30 to 50 percent to build momentum. Most people think that’s too high. But they’re thinking short-term. You’re trading margin for distribution. You can turn it down lower for future affiliates in a year or two.

I’d probably use Upwork to take my prototype to the next level. I hear Replit bounties are good too. It’d help with security and iterate on a few features. After all, I’m a computer science college dropout.

Now it’s time to build the affiliate ENGINE. I wouldn’t just reach out to 10 A-listers. I’d go wide. I’d build a list of 100 or more B- and C-list creators. People with 5k to 70k followers who already have trust, aren’t over-messaged, and know their audience well.

I’d use GPT-4o to generate custom DMs. I’d use Lindy to send and follow up. I’d use Gumloop to manage replies, track who’s interested, and automate the next steps.

Each pitch is short and specific. “You talk to [audience]. I built something that solves [pain]. You get 50 percent of every dollar it makes. Want a walkthrough?” I’d record a 90-second Loom for each one, showing exactly how it works, and set them up with a white-labeled affiliate dashboard.

Next step is I’d gamify it. The affiliate side becomes a competition. Leaderboards. Milestones.

“Earned your first $1,000. Badge unlocked.”

“Top partner this week gets a custom landing page and a shoutout.”

"Top affiliate gets a used BMW"

I’d automate weekly earnings updates. Spotlight top performers on social. You’re building hype, loyalty, and momentum. And the best part is it’s automated and can scale. You're building a vending machine. Paying out lots but making more.

At the same time, I’m gamifying the product side too. Not just the seller, the user. I’m building streaks. Progress bars. Unlockables. Referrals that feel like sharing a cheat code.

The whole thing should feel like a game where the user is winning every time they log in. I get inspo by playing Fortnite or CS2. The SaaS becomes useful, fun and spreads.

Once users are inside, I co-create with AI. Every new idea gets filtered through Claude or GPT. “Which use cases are users getting stuck on? What are the top 3 complaints this week? What’s the lowest-effort feature we could add that unlocks the most value?” It’s like having a full product team in my inbox. I build what’s already happening organically. That keeps churn low and word of mouth high.

Meanwhile, I’m building artifacts in public. Sharing updates, sharing teardown Looms, before-and-after workflows, even mistakes. I think how can I become the main character of this movie but also be giving value to people consuming this content. If you can do that, it’s a serious moat.

Every artifact builds trust. People start seeing me as the go-to tool for this niche. When I release something new, it feels like the next chapter in a story they’ve already been following. They are supporting.

I’m not building for a big exit. I’m building for optionality. A buyer might come along, and sure, maybe they offer 7x revenue. But because the whole thing runs lean with zero full-time employees, I don’t need to sell. I can cashflow. I can license the playbook. I can spin up a second tool in the same niche with the same engine. That’s the real power of building this way. The exit isn’t the win. The leverage is. The freedom is. And, MAN, it tastes SWEET!

Eventually, I’m at $30k per month, then $50k, then $80k. And it’s still just me, maybe some contractors and a boatload of AI agents. No paid ads. No outbound sales team. Just systems, partners, and product doing the heavy lifting. Feels almost unfair. You can layer that stuff on later if you want.

If you’ve been on the sidelines waiting for a sign, this is it.

You can build something real. Something useful. Something profitable. You can take back control of your time, your income, your creativity. All it takes is one great use case, one small loop, one tiny success that compounds.

More ideas like this on "the startup ideas podcast" wherever you listen to your pods.

You totally could do this with a full-time job or maybe you're still in school.

And no, to be clear STARTUP ARENT easy. It’s not supposed to be. But it’s possible. And WITH ALL these trends (AI, niches, social etc) this time makes this the best shot you’ve had in a long, long time.

I hope you take it. Have a creative day, friends.

I’m rooting for you.