How do you validate that your side project won't be replaced by AI?

I spent 3 months building a real estate investment simulator — loan modeling, cash flow projections, Stripe integration. Then AI got good enough at finance calculations that users could just ask Claude the same questions and get answers instantly.

It made me realize I never asked the most important question before building: "Does this need to be a product, or can AI do it well enough?"

I wrote up the experience and a simple 4-step "AI Replacement Test" I now do before starting any project:
https://dev.to/toshipon/i-spent-3-months-building-a-saas-then-ai-did-the-same-thing-in-one-prompt-3kbf

Curious — how are you all thinking about AI risk when starting new projects? Anyone found a good framework for this?