I ran 50 AI posts against 50 human ones. Here's what actually happened.

I ran 50 AI posts against 50 human ones. Here's what actually happened.

Month 1: one $12 domain, niche was indoor gardening. I split it cleanly — 50 posts written by AI with light editing, 50 written by hand. Same niche, same internal linking, fair fight. I wanted the real answer, not the hype or the hate.

Month 5: both sets ranked. Honestly, the AI posts ranked FINE — Google didn't smite them. Traffic split roughly even, 11,000/mo total. The AI-vs-human ranking panic was overblown.

The turn: month 8, conversions told the real story. The human posts converted at 3.1%; the AI posts at 0.9%. Same traffic, triple the revenue from the human half. The AI posts ranked but didn't persuade — no real recommendation, no "I actually used this."

The setback: I'd assumed the experiment was about rankings. It was about trust. Visitors read the AI posts, sensed the hollowness, and left without buying.

Month 12: I rewrote the 20 best-trafficked AI posts by hand, injecting real opinions and photos. Their conversion tripled. Total revenue $2,900/mo.

The takeaway: AI content can RANK but it struggles to CONVERT, because affiliate sales need a believable human recommendation. Use AI to find what gets traffic, then put a real person on the pages that need to sell.