A concept music app that gets you to the right sound in two taps, and learns your work from the way you listen.
Spotify, Apple Music, YouTube Music: they're all built around one goal: keep you in the app as long as possible. More browsing, more playlists, more recommendations, more scrolling.
Nord starts from the opposite end. After two weeks of daily use, it knows you better than any playlist ever could: lo-fi at 9am, something instrumental after lunch, ambient when you drop into deep work. Open it, get the right sound, get back to your task. If you can do that in two taps, every day, the app has done its job.
The product thinking is front-loaded. Define the user, the behavior, and the win condition, then ask for an MVP scope.
I want to build a music app like Spotify, but designed for focus instead of discovery. It's for users who just want to play music fast without exploring content. It should learn which songs they like from previous listens and behavior — for example, music to accompany work. Simple and intuitive enough that when they're working, they open the app and play suggested music. If they open it every day and finish their task within two taps, this app is a win. Suggest an MVP: what pages do I need to design, and help me with the content on each page.
Not one perfect prompt hundreds of them. We treated AI like a creative partner: throwing references, pushing back on outputs, redirecting when it went generic. The build phase wasn't about typing the right thing once. It was about thinking out loud, iterating fast, and knowing when to stop.
First result → two rounds of judgment → final. The AI generates; the designer decides what's working and what to cut.
This concept was fully created with AI assistance. But AI is only as powerful as the direction behind it: the real value lives in the product thinking, the behavioral insight, and the prompts that shape the outcome.
The hard part of Nord wasn't the UI: a tool can generate a music screen in seconds. The hard part was deciding what the app should not do.
No infinite scroll. No endless recommendations. No discovery feed. That restraint came from product thinking, not from the AI, and it's the part of the job that doesn't get automated away.
What if Coretax was designed around humans, not just rules? — Concept