How we rate AI tools

Editorial reviews, community averages, and sourced facts are separate signals. We only call a tool editorially tested when its complete test record is published. This page explains our evaluation process.

Who is behind the ratings?

Every published editor rating on One AI Guide is produced by editor and founder Said El Moussaoui and his editorial team. Said trained at The Hague University of Applied Sciences in project management, research, communication and problem solving, and brings that same structured, evidence-first approach to every review. Tools with published editor ratings are tested on real tasks, compared against their closest alternatives, and scored on clear criteria. No outsourced star-farms, no AI-generated reviews presented as human judgment.

What a rating measures

• Hands-on testing: we run real tasks through the tool, not vendor demo scripts
• Value for money: what you actually get per euro or dollar against the best alternatives
• Output quality: is the result good enough to ship, compared side by side with category leaders
• Kept up to date: ratings are revisited on major versions, pricing changes, and reader feedback

The rules we never break

• Commercial relationships and conflicts are disclosed; affiliate links never improve a score
• Sponsored placement is identified where surfaced, as applicable; labels and presentation can vary by format. It stays separate from editorial testing and community averages
• If a tool is mediocre, the review says so
• Scores use a 0-10 scale: 9+ is category-defining, 7-8 a solid pick, 5-6 has real trade-offs
• Corrections are applied to the live review, not hidden