The Best AI Investment I Ever Made Was a Business That Didn't Work

A few years ago I spent $99 on a course and convinced myself I was about to build a business.

I wasn't wrong about the building part.

The $99 course and the dream of passive income

The promise was simple: use AI to build a Shopify store that makes money while you sleep. At the time, ChatGPT, DALL·E, Midjourney, and Make.com were just becoming accessible to everyday people. Not just developers. The course showed how to generate artwork with AI, place those designs on print-on-demand products, and automatically publish everything to Shopify.

I already had a small shop called CopperBoom with about twenty products. I remember thinking: if this works, I can actually build something real.

So I went all in.

I built automations that generated artwork, created product listings, and published them without me touching anything. I expanded from canvas prints into phone cases, apparel, home décor, pretty much anything print-on-demand would support. I was learning as I built, which is always my favorite way to learn. The store grew.

5,000+

Products in the CopperBoom store AI-generated. Automatically listed. Almost entirely unbought. It turns out that filling a store with thousands of products doesn't automatically create customers. Who knew.

CopperBoom never became the passive-income machine I imagined. The maintenance outweighed the revenue. The automation worked fine. The business model didn't. Eventually I stopped chasing the fully automated store dream.

But something else was happening the whole time I was failing at it.

What I was actually learning

Nobody tells you this, but building a complicated thing that doesn't work still teaches you how to build complicated things.

While I was busy trying to make CopperBoom profitable, I was quietly picking up a whole set of skills I didn't have words for yet:

How to think in workflows instead of individual tasks. How to automate with Make.com, connecting tools, building logic, handling errors. How to write prompts that actually do what you want. How to connect AI tools together to solve a problem end-to-end. How to design a system instead of doing the same thing manually over and over.

None of that showed up in the revenue numbers. But it showed up in everything I built after.

As new AI capabilities kept emerging, I kept experimenting inside CopperBoom. I used AI to design original t-shirts instead of just generating abstract art. I tried AI-generated product photography. I wrote product descriptions, marketing copy, and social content with AI. Each experiment added another layer of fluency I wouldn't have found any other way.

At some point CopperBoom stopped being a business and started being a laboratory.

A laboratory where every failed experiment taught me something useful.

The reframe I didn't see coming

Where it actually showed up

I didn't realize it while it was happening. But the months I spent building and breaking CopperBoom became the foundation for almost everything I do now.

Khloe and the executive dashboards built on the same workflow thinking I learned trying to automate product listings.

The migration automation work is six reusable workflows connecting AI to live systems. Same pattern, completely different problem.

Coach M and Hannah came from learning to design systems around how people actually operate, not how you assume they do.

The workshops I'm building exist because I know how to teach this stuff. I know how to teach it because I learned it by breaking things, not by reading about it.

The consulting work is credible in part because of a failed t-shirt store. I never would have predicted that.

Almost none of it has anything to do with selling t-shirts.

What CopperBoom is now

The store still exists. It just serves a different purpose than the one I built it for.

The thousands of AI-generated products are mostly gone. What's there now are designs that actually mean something to me. Things I'd wear. Things I'd give to someone I care about. Things that make me smile when I look at them. It became a creative outlet, and honestly I like it better this way.

The business I imagined never happened. The skills I built while chasing it are behind almost everything I do now.

What I keep thinking about

I spent $99 on a course and got something I never expected. Not a business. A completely different way of thinking about problems. The store didn't work. The thinking did. I'm not sure I could have gotten there any other way.

If you have an experiment you've been putting off because you're not sure it'll work, I get it. I really do. But some of the most useful things I've ever built started exactly that way.

I'm still figuring out what CopperBoom wants to be. I already know what it made me.

Want to start experimenting?

AI Without the Overwhelm

This is one of the workshops I am planning, designed for exactly the kind of curious experimentation that eventually becomes fluency. You don't need a business idea or a technical background. You just need a problem worth solving and a willingness to break a few things along the way. If this story felt familiar, come build something with me.