I designed and launched the CONTROL app, a music app for our indie label, and it cured me of one fantasy fast: shipping an app is not the same thing as having an app idea.

A lot of talk around AI and app building makes the process sound flatter than it is. Prompt well, generate some code, ship something, move on. CONTROL was a good reminder that the distance between concept and release is still full of infrastructure, review cycles, edge cases, and paperwork. The old version of this post named some of that directly: Apple’s developer fee, hosting, API costs, analytics, licensing, privacy disclosures, and App Store review.

That list mattered because CONTROL was not a mockup. It was a live product tied to real music, real users, and a real label.

One detail that changed the feel of the project for me was licensing. Because the app includes original tracks, I had to prepare license agreements before submission. That is not the glamorous part of product building, but it is part of the product. So are privacy disclosures. So are user-flow audits. So is getting something rejected and having to fix it.

AI helped, but in a narrower way than the hype suggests. Tools like Claude and Cursor were useful for scaffolding repetitive work, offering debugging ideas, and helping me think through integrations. They were much less useful when the work became architectural or stateful. They did not understand the product better than I did. They did not save me from having to know what I was building.

That is probably the cleanest way I can put it now: AI was good at acceleration, not substitution.

The more interesting part of CONTROL came after launch. The original post turns there too, toward growth, feedback, and the bigger question behind the app: whether an indie label can use software to build a stronger relationship with listeners than a streaming profile usually allows. That still feels like the right question. What do fans actually want from a label-owned music app? What earns attention beyond novelty? What helps artists, not just engagement metrics?

I do not think CONTROL proved some grand thesis. It did something more useful. It made the real shape of the work visible.

It showed me that product design in this space is not just interface work. It is technical judgment, legal coordination, operational patience, and a willingness to learn from release instead of treating release like the end.

That is the part I would keep.

CONTROL was my first iOS music app, but more than that, it was a lesson in how quickly creative vision becomes systems work once real people are meant to use it.