Sep 20261 min read
We used it.
Here's what
happened.
AI tools reviewed by running them on real work — a published testing protocol, time-stamped results, and honest verdicts including the frequent “don't bother.”
The gap
“Best AI tools for X” is the most saturated content there is — and almost none of it involves using the tools.
The standard article is twelve products assembled from their own marketing pages, ordered by affiliate payout, published by someone who opened none of them. Meanwhile the real questions go unanswered: does it work on your input, not the demo's? What breaks at volume? What does it cost past the free tier? Where does it hallucinate? What happens to your data?
Testing is the differentiator because it's laborious, not because it's clever. That's what makes it defensible — read how we test.
The hard line
- ✗ No untested reviews, ever
- ✗ No rankings ordered by commission
- ✗ No “best AI tools” listicles from marketing copy
- ✗ No AI-generated body prose — a publication that tests AI tools cannot be written by one
- ✗ No affiliate link without the review behind it