Review Descriptions

The default checks look for:

• AI comments, refusals, or eligibility verdicts,
• not eligible, cannot determine, or insufficient information wording,
• alcohol, tobacco, gambling, or non-work entertainment wording,
• specific claims that are not supported by the entry or evidence,
• empty or placeholder descriptions.

You can turn default checks off and provide your own extra things to find. In that mode, the review only looks for the checks you describe. Sensitive short checks, such as alcohol, are treated as direct description-text matches rather than loose associations. If the description does not actually contain the requested word or a direct obvious variant, it should not be flagged.

Direct category phrases can also work when the description has concrete evidence for that category. For example, AI chatbot subscriptions can match descriptions that mention ChatGPT, OpenAI, Claude, or similar AI tools together with subscription, paid plan, licence, membership, or Plus-style wording.

If default checks and extra checks are both active, findings from your extra checks are shown first and marked with a User badge as well as High, Medium, or Low severity.

For each finding, you can edit the suggested description and save it to the entry, or dismiss the finding for the current review session. Dismissing a finding does not change the entry.