How we test, rank, and stay honest

These standards govern every review, guide, and video we publish. They are public so you can hold us to them.

The verdict process

• One buyer question per review. Broad "top 10 AI tools" listicles help nobody. We answer a single, specific question a real buyer types into a search bar.
• Evidence synthesis. We gather the strongest published expert reviews and hands-on tests for the category — credited and linked on every review page — and combine them with our own testing criteria.
• Claim verification. Each factual claim is checked against its source context before it can appear in a verdict. Claims that overstate, misattribute, or can't be supported are cut. This step is automated in our pipeline and logged per episode.
• The verdict. One best-overall pick, the best free or budget alternative when one genuinely exists, and the pick's honest limits — always, all three.

Monetization rules

• Affiliate links never decide verdicts. Picks are locked before monetization is attached. Some of our picks have no affiliate program at all; they stay picks.
• Disclosure everywhere. Every page and video description carrying affiliate links says so plainly, next to the links — full policy.
• Sponsorships are labeled placements, never rankings. A sponsor can buy a clearly-marked placement; a sponsor cannot buy a verdict, a ranking position, or the removal of criticism.

Corrections

When a tool changes materially — pricing, licensing, quality — we update or retract the recommendation and note the change on the page. Factual errors are corrected within 48 hours of verification. Report one: hello@theaishortlist.co.

Independence

Vendors do not review, approve, or edit content before publication. Pre-briefings and demo access are welcome; embargoes on criticism are not accepted.