A Dead QR Code, a $10,000 Customer and the Software I Built With Claude Code

What happened?

A friend of Rob's owns a local business where the average customer is worth about $10,000. He printed and mailed thousands of postcards, then asked Rob a simple question: "Can you see if this QR code works?" Rob scanned it. Nothing. The code was dead.

They couldn't fix it. Not because QR codes can't be updated, but because whoever made the code controlled where it pointed, and there was no way to get access. Thousands of postcards and thousands of dollars in printing and postage, all tied to one broken link.

Why are broken QR codes a hidden revenue leak?

The worst part isn't that the code breaks. It's that owners usually don't know. With no scan tracking, they assume nobody scanned it and write the campaign off as a dud. In reality, people did pull out their phones, hit a dead page and walked away.

Rob calls it the infrastructure trap. Most QR code services control the redirect. If the company shuts down, changes its prices or you lose your login, your printed marketing is held hostage. That's a risk for restaurants, real estate agents, contractors, insurance agencies and anyone who prints marketing.

How did Rob build the fix?

1. He started with the problem, not the code. A specific, expensive problem made every decision easier.
2. He used Claude Code to plan and build the foundation. He laid out how the redirects, scan tracking and billing should work, and Claude Code generated the core of the app.
3. He used Lovable to finish the screens and launch it.

The result is a real platform, not a toy: user dashboards, editable QR codes, scan tracking, an admin back end, Stripe billing and secure user permissions.

How do editable QR codes protect your print budget?

Say you print 10,000 flyers. At launch, the code points to your summer promo. Mid-year, you log in and point the same code to your fall promo. Next year, you launch a new website and point it there. The printed code never changes, so you never have to reprint just because your plans did.

What's the bigger lesson about AI?

Most people use tools like Claude Code to build demo apps and clones. That's fun, but the real opportunity is finding painful, expensive business problems and building software to fix them. As Rob puts it: AI didn't invent the opportunity, it accelerated the solution.

A person still has to decide what to build, test it and handle the odd cases. AI is the very fast helper, not the decision-maker.

What's your $10,000 problem?

Look for where money quietly leaks out of your business: calls nobody answers (an AI receptionist fixes that), a website with no clear next step (see this redesign before and after), slow training (like the onboarding app Rob built) or numbers scattered across tools (like his YouTube analytics dashboard). Building the fix is what our custom AI tools service does.