Editorial Policy - How I test the tools on this page
Every "best AI tools" list you've ever read was probably written in one sitting, from the tool's own marketing page, by someone who never opened the product.
I know this because I used to half-do it that way too, back when I thought a review just meant summarizing the pricing page in my own words. It's an easy trap. Useless to you, though.
Here's what actually happens before anything gets published on this site.
I use it for real work, not a demo task
No tool gets reviewed off a single test prompt. I run it against my actual inbox, my actual task list, my actual half-finished projects. The messy stuff. Not a clean example built to make the tool look good. Minimum two weeks of daily use before I write a word. Some tools get months, if I end up keeping them around.
I try to break it on a bad day
The real test of a productivity tool isn't how it performs when I'm focused and motivated. It's whether it still works when I'm scattered, behind, and one bad night of sleep away from giving up on the whole system. A tool that only works when I'm already doing fine isn't solving anything.
I score the same five things, every time
- Setup and onboarding. How long from signup to actually useful, and how much of that time was spent fighting the tool instead of using it.
- Does it survive a bad focus day. This is the ADHD-relevant one. It's missing from most reviews, because most reviewers don't need it to be there.
- Output quality. Did it save real time and real thought, or did I spend as long fixing its output as I would have doing the task myself.
- Price versus what you actually get. Not what the pricing page implies you'll get.
- What happens when it breaks. Support, refund policy, whether the company ghosts you the moment you ask a real question.
What doesn't factor in
Nothing else. Whether a company treats me well as a reviewer doesn't move the score. Neither does a nice press kit, or whether they buy ads on this site. If a tool I got for free turns out to be mediocre, the review says mediocre. I've had exactly one company pull a partnership over that so far. Good sign, not a bad one.
Who's doing the testing
Right now, it's just me, Theo Keene. No team, no quiet ghostwriter running the actual tests while a byline gets slapped on top. If that changes, this page changes with it, and I'll tell you exactly who's testing what.
When a review goes stale, I say so
AI tools change fast enough that a review from eight months ago can be actively wrong. I re-test anything load-bearing on this site at least once a quarter. Every article shows the date it was last checked. If I haven't gotten to something and you catch it out of date, tell me (email: theokeene@icloud.com).
That's the whole process. Not complicated. Just slower than writing from a press release, which is why so few sites bother doing it.