A translation is flagged either because an automatic check found a real problem with it, because the AI was not confident, or because a reviewer flagged it by hand.
Who this is for
Anyone looking at a review queue and wondering why a particular string is in it.
What happens next
Three things put a string in review.
An automatic check failed. A broken placeholder, the wrong plural forms, text over the key's maximum length, or a conflict with a glossary term. These always win: a translation that fails one of these checks is sent to review no matter how confident the AI was. Open the string in the editor and look at Quality findings in the context panel: the specific problem is named there.
The AI was not confident. Each AI translation carries a confidence signal. Translations that pass the automatic checks and come back confident can be marked as not needing review, so your reviewers only see the ones worth their time. Everything else is flagged. Confidence can only add reviews; it can never clear one an automatic check demanded.
Someone flagged it. A reviewer used Flag in the editor.
Automatic marking as not-needing-review is off unless the project has Quality mode on, so on a project without it, everything AI produces arrives flagged, which is the safe default rather than a fault.
Related
Approve or flag a translation
Quality checks
AI translation settings