AI is useful for prep, not the trust layer

This screenshot is a good example of how far a social prompt can go. The model is told to write like a real person, use posts, comments, likes, user behavior, and demographic context, then keep the reply short and conversational.

That is useful for drafting. It gets much shakier once AI starts deciding what feels natural, what deserves a reply, or how to shape the tone around a person’s identity signals.

For founder growth on X, the split is pretty clear:
• Let AI search and cluster
• Let AI draft and summarize
• Keep the thesis, judgment, and final edit human

That’s the trust layer this playbook is really built around.

Data source: Green Dots Research

Related read: How to Grow a Founder Persona on X: A Practical Founder Growth System