Don't ask AI for your SaaS idea. I learned this the hard way.

I see new founders do this constantly. I did it myself.

Open ChatGPT or Claude. Ask "what's a good SaaS idea I should build?" Get a list. Pick one. Start building.

Three weeks later — five funded competitors already solving the exact problem.

Here's the thing nobody says out loud:

AI gives you ideas that have already been written about hundreds of times. On Indie Hackers. On Twitter. On Reddit. Which means other people are already building them by the time AI suggests them to you.

The idea must come from your own life.

The best ideas come from a moment where you think "why is this so hard, there should be a tool for this." That moment is specific to you. Your workflow, your frustration, your situation. That specificity is your only real advantage as a solo founder with no funding and no audience.

When I stopped asking AI for ideas and started paying attention to my own frustrations — I noticed I kept losing things in my own Notion workspace. Files I couldn't find without re-reading entire pages. That became my idea. Not because AI validated it. Because I felt it myself.

Where AI actually belongs in this process:

Finding the idea → your job, not AI's
Researching competitors → AI helps
Writing landing page copy → AI helps
Building the MVP → AI helps
Writing validation posts → AI helps

The idea must be yours. Everything after — AI can help you execute faster than any solo founder could five years ago.

One question I ask before starting any idea now:

"Did I feel this frustration myself — or did someone else tell me it was a problem?"

If you felt it → worth building.
If you read about it → research first, build never.

Has anyone else made this mistake? What was the idea AI suggested that turned out to already exist?