How to Start Making Money with AI as a Beginner (From a Founder Closing Enterprise AI Deals)

There's a lot of noise in the "make money with AI" space right now. Half the advice is from people who've never closed a deal. The other half is people repeating the same agency-first playbook from 2023 without updating it for what AI has actually done to the agency model.

This is the playbook I actually run. It's the same one I used to close AI implementation deals with publicly listed companies while building Ordana. It works for two reasons: it covers both ends of the time horizon (cash now, equity upside later), and the two ventures share so much infrastructure that running both costs barely more than running one.

The Two Real Paths

Strip away the noise and there are really only two ways to make money with AI from a standing start. Most gurus tell you to pick one. The trick is to refuse to choose.

Most gurus tell you to pick the agency path because "it requires little capital." That's true. But here's what they leave out:

• The upside is limited — you're trading hours for dollars, even if the dollars are good.
• Growth is slow because every new client is a new sale, not a compounding asset.
• I don't know a single agency founder whose stress level is proportionate to their income. They're all running too hot.
• And the kicker: AI agencies will be the first ones eaten by AI. An implementation I sold a client three months ago is already half-obsolete because the underlying model shipped a new capability. We become less valuable, not more, as the space matures.

So the agency-only path leads to a treadmill that gets steeper over time. But the SaaS-only path means you might spend a year building before you see a dollar — and most people can't afford that.

The answer is to run both. Not eventually. From day one.

Why an AI SaaS Is the Only Real Long-Term Play

Yes, AI lets you build an MVP in a week with tools like Lovable, Cursor, and Claude Code. That doesn't mean you'll have a business in a week. It still takes months of iteration — build, market, get feedback, iterate, repeat — to find something that sticks.

Most SaaS will die over the next few years too — vibe-coded clones will eat the soft middle of the market. To survive, your value proposition has to include at least one of three defenses.

I unpacked these three in much more detail in Startup Moats in the AI Era. Before you commit to a SaaS idea, make sure it has a plausible path to at least one of those three. If it doesn't, you're building a clone-bait product.

The Cold-Start Problem

Most beginners hit a wall here. The SaaS needs months of runway. They don't have months of runway. So they default to the agency path and never come back to the SaaS dream.

This is exactly where the dual play matters. Start the agency for cash. Start the SaaS in parallel. They share more infrastructure than you'd think.

The Trick: Your SaaS Is Your First Agency Case Study

Now the question every new agency founder asks: "How do I get my first case study without a client?"

Your own SaaS is your first case study.

When a client is deciding whether to hire you, they want to see proof of your design quality, your thinking, your delivery speed, the polish of your work. A live SaaS you built and run answers every single one of those questions in 30 seconds. It also answers them better than a slide deck of past client work — because it's something you own, ship, and iterate on, not a frozen artifact.

See yourself as your first client. Your AI SaaS is your first deliverable. You learn the craft on something you own. You can experiment freely without risking a client deliverable. And the same product you're building for proof is your billion-dollar shot.

The Second Trick: Land Your First Client on Revenue Share

One case study isn't enough. Most agencies plug the gap by doing the first one or two engagements for free, in exchange for a testimonial. That works, but you're working without pay during the period when you most need money.

There's a better structure: revenue share.

For the client, the risk is zero. For you, the upside is real — if you deliver well, you earn more than a flat fee would have paid you. And you get the case study either way.

The reason this didn't work for most agencies historically wasn't that the model was wrong. It was that setting it up was a nightmare: contract negotiation, revenue tracking, Stripe access, monthly reconciliation, trust disputes. By the time you'd built the operational scaffolding, you'd lost three months.

Ordana solves exactly that piece. You set up the revenue share contract in a day. Revenue is tracked automatically from the client's Stripe account. Payouts calculate and disburse on a defined cadence. Both sides see the same numbers in real time. The trust gap closes because the math is no longer up for debate. I went deeper on why this pricing model is now structurally inevitable for agencies in Why Agencies Should Charge Revenue Share Instead of Retainers.

Why the Two Ventures Reinforce Each Other

Once both engines are running, the synergies start compounding:

What This Looks Like in Practice

Pick the persona that matches where you are right now. The playbook bends slightly for each, but the underlying engine is the same.

The Mistake Almost Everyone Makes

They pick one path and treat the other as a someday-maybe. SaaS-only founders run out of runway. Agency-only founders get stuck on the treadmill and watch the moat dry up under them.

The dual play is the only structure I've seen consistently work for founders coming in with no money and no track record. Agency for cash. SaaS for upside. Each one solves the weakness of the other.

And the meta-point: the infrastructure that makes this operationally work — revenue share contracts, automated tracking, payouts, AI matching to clients — didn't exist a few years ago. It does now. If you're starting today, you have an unfair advantage previous founders didn't have. Use it.

What to Do This Week

Two years from now, the founders who hedged with both engines will still be standing. The pure-agency founders will be racing to the bottom on AI implementation rates. The pure-SaaS founders without runway will already have pivoted to something else.