Something interesting happened to me.

I started building Leaflo about a year and a half ago or so. Back then, I thought through the feature set, the rough positioning, did a competitor analysis, and after building about half of the product, I put it aside for different reasons. It happens.

When I came back to it, I started discussing the product with ChatGPT more often. And somehow, over time, I drifted into a completely different positioning — one that moved me away from my original idea.

Originally, I was planning to build a CBT journal that helps a person during therapy, and also before and after it. Because journaling is a common practice in therapy.

But in the end, I drifted into a wellness-style personal journal. And the product itself became a bit split.

On one hand, it had simple personal diary features. On the other hand, it had a library of prompts, which belongs more to a focused, goal-oriented writing format.

At some point, I realized I had moved in the wrong direction. And now I’m spending time and energy trying to get back to the starting point.

The conclusion is simple:

When you discuss your tasks and projects with AI, don’t forget your own ideas and intentions. Most likely, they are more valuable than anything AI will suggest to you.

Don’t switch off your own thinking. Use AI for specific tasks and specific requests, but try not to trust it with anything strategic.

It should not make conclusions for you or pull you away from your direction.

Sounds obvious — but I still fell into this trap.