Building an AI thinking partner — early lessons from user conversations

I’m working on Novarch, an early-stage AI thinking partner focused on clarity, decision-making, and deliberate action — not motivation or hype.

One thing I’m learning quickly: people don’t struggle with ideas, they struggle with holding a thought long enough to reason through it.

I’ve stopped adding features and started watching how users actually talk to the AI. That shift alone has changed my roadmap more than any planning session.

Still very early, still validating, and mostly listening right now. Curious how others here approach early user learning without overfitting too fast.

— Mesum