💵 Make money building apps with AI (honestly)

AI builders collapsed the cost of making software — which means "I build simple tools for small businesses" is now a viable service. We run a version of this business ourselves, so here's the honest playbook: where the money actually is, how to price, and the traps.

Where the real demand is (hint: not app-store dreams)

Forget the next Uber. Small operations pay real money for boring, specific tools: job quoting calculators, scheduling boards, inspection checklists that generate PDFs, lead trackers (like this one), customer portals. The pattern: a spreadsheet that's outgrown itself, a paper process embarrassing someone, or a $300/month platform where they use two features.

Step 1: Prove it on someone you know

Your first client is in your phone: a contractor, a shop owner, a league organizer drowning in a spreadsheet. Build their tool free or cheap in exchange for: honest use, feedback, a testimonial, and permission to show the work. One real deployed tool with a happy user outranks any portfolio site.

Step 2: Price the outcome, not your hours

Typical reality for simple business tools: hundreds to a few thousand dollars per project, plus modest recurring for hosting and tweaks. Not startup riches — but 3-4 projects a month is a real income, and recurring support compounds.

Step 3: Productize the second sale

Your second quoting calculator takes a third of the time — same bones, new branding and logic. That margin is the business. Keep every build as a reusable base, and niche down: "I build tools for [plumbers / gyms / ag operations]" beats "I build anything" in referrals, pricing, and speed.

The traps (learn from our scars)

• Scope creep: "one more little thing" is how fixed-price projects go underwater. Written scope, priced change requests, kindly enforced.
• The maintenance cliff: every deployed tool is a small forever-promise. Charge for support or set an honest end date — never imply free-forever.
• Overbuilding: clients buy outcomes. The 4-screen version that ships this week beats the 12-screen version that ships someday.