A QA check is an automated or semi-automated review that verifies translation consistency, terminology use, numbers, inline tags, placeholders, punctuation, and spelling against project-specific settings. It is typically run within a CAT tool or standalone QA application after translation and editing, before delivery, to catch technical and basic linguistic errors such as mismatched tags, missing placeholders, or changed numerical variables.
QA checks run in the final stages of translation projects, using predefined profiles in CAT tools or QA software. Linguists, project managers, and localization engineers review the results, resolving issues such as untranslated segments, forbidden characters, tag mismatches, and terminology noncompliance before final delivery; the output often feeds into language quality assessment.
A common mistake is relying on QA checks to catch all translation errors. Automated checks miss context errors, false friends, register shifts, and awkward wording, and they can generate many false positives if the profile is too strict; ignoring project-specific settings may also let serious technical defects pass, such as broken variables.
For creative and consumer content the failure mode is flatness, not inaccuracy — and flatness passes an accuracy check. Terms here also carry technical constraints (reading speed, character budgets) that are part of the brief.
Angel Translation locks terms like this into a project terminology base before expert review begins, so the same source term resolves to the same target term across every document in a submission — and stays consistent in the next revision. See the 8-step AI + expert workflow.
Send us the files. Acknowledged in 4 hours, quoted in 24.
Request a Quote →Content Globalization solutions →