Review the claim before you cite it
Mark what the public material discloses. The tool returns a risk level, the missing evidence, and a cautious sentence you can use without overstating the result.
- Start by marking what the claim discloses.
Review a claim to generate cautious language.
What the result means
| Tool result | Responsible use |
|---|---|
| Usable with stated limits | The claim discloses most required evidence. Repeat it only with the sample, environment, dates, formula, and limitations. |
| Use only as a publisher-attributed claim | The public material is partly inspectable, but important context is missing. Attribute it and name what is absent. |
| Do not repeat as evidence | Critical evidence is missing. Treat the statement as marketing language, not as a benchmark, proof, or planning assumption. |
Minimum evidence to look for
A responsible AI visibility claim should disclose the exact prompts, denominator, formula, measurement dates, product environment, raw observations, exclusions, and limitations. A lift claim also needs a baseline and comparison that are meaningfully matched.
For the full manual review sequence, use the AI visibility claim review checklist. For rate calculations from your own observations, use the AI citation rate calculator.
What this tool does not establish
This tool reviews disclosure quality. It does not independently reproduce a result, certify that a claim is true, or prove that an AI product will behave the same way in another account, market, date, model, or interface.
Audit the page behind the claim.
The best claim review still starts with the source page: clear entity context, visible evidence, and limitations that travel with the result.
Audit a page free