The XTM Agent can now answer questions about the quality of your AI translations. Two new tools let you look at quality scores and at what your reviewers did with those translations. You ask in plain language, the same way you already manage projects and tasks with the Agent.
1. TQI distribution: how your quality scores are spread
See how the TQI scores of your AI translations are spread over any period. Use it to check overall quality, compare languages or projects, and catch a drop early.
Try asking:
- “How did our TQI look in September?”
- “French versus German this quarter.”
- “Any worrying drop in the last two weeks?”
2. Post-edit outcomes: what reviewers did with AI translations
See how many AI translations reviewers accepted without changes, how that depends on the TQI score, and whether a score is a safe cut-off for auto-accepting translations.
Try asking:
- “What was the acceptance rate last month?”
- “Would 92 be a safe cut-off?”
Numbers you can trust
- Any scope: ask about any period, project or language. If you don’t name one, the Agent covers your whole organization.
- Counts included: Percentages will be backed up with exact numbers for edited , accepted with without edit etc.
- Small samples flagged: if a result is based on too few translations to be reliable, the Agent tells you.
How to try it
- Log in to the Transifex Dashboard.
- Click the support bubble in the bottom-right corner.
- Choose XTM Agent and ask your question.
Read the full guide: XTM Agent for Transifex
We’d love your feedback. Tell us below which questions you asked and what you’d like the Agent to answer next. 