How we spot a bought review.
Every "best AI tools" list is for sale. Review pools are cheaper. A few hundred bucks buys a launch-week glow and 40 five-star strangers. When we built the database we found patterns in the review data we weren't looking for. The same tells, tool after tool. Five of them below.
1 · the launch burst
Twenty reviews in 48 hours on a tool nobody's heard of. Real users trickle in. Bought ones arrive by the busload. Check the dates. A cluster around launch day or a funding announcement is the first flag.
2 · the adjective pile
Fake reviews praise everything and describe nothing. Game-changer. Seamless. Intuitive. Real users name the job: invoices, follow-ups, the $47 overage they didn't see coming. No job named, no weight given.
3 · the missing failure
Six months into any tool, real users are annoyed about something. Always. A pool with zero gripes is a pool somebody paid for.
4 · the one-tool resume
Click the reviewer. If their whole history is one glowing review, that's not a user. That's an invoice. It cuts the other way too. Our own filters caught a single account behind a whole stack of reviews. One voice pretending to be a crowd is not a crowd, and it got filtered out of the pool.
5 · the copy-paste chorus
Same phrasing across a dozen accounts, sometimes the same typo. Nobody organically writes "streamlines my workflow" twelve times in one week.
Believing a bought pool costs you twice. First the subscription. Then the weeks you spend making a bad tool work because 400 strangers swore it does.
When the flags stack up, the tool lands in BOUGHT on the news desk. BOUGHT is the negative space of this whole site: I score tools that work, that shelf names the ones buying their way in. Receipts included, every time.