AI Words in XTM Cloud
Source text words processed by the selected AI-based features are referred to as AI Words. If you use an XTM-provided LLM subscription, you must purchase an AI Words allowance for your entire XTM Cloud instance. AI Words are deducted from your allowance for every project Job.
Important
AI Words are based on project metrics, but are calculated differently for each of the AI-based features. For details, see sub-sections below.
AI Words are only deducted for segments that have received a successful LLM response. The exact moment at which the AI Words are deducted differs for each of the AI-based features. For details, see sub-sections below.
If a project is reanalyzed, the AI Words are deducted again for all reanalyzed Jobs.
You can check your current AI Words usage in the Intelligent Automation KPIs Dashboard. The counter is refreshed within a couple of minutes after each AI Words deduction.
AI Words calculation:
Only words from source segments are treated as AI Words.
Words from the translation memory (TM), terminology, project's subject matter, the Style guide field, and context segments (which are also sent to LLMs by the AI Translate & TQI feature) are NOT counted as AI Words.
XTM Cloud Administrators and Project Managers can also specify that certain TM match types are not sent for AI Translate & TQI translation. In that case, AI Words for excluded match types will not be deducted. For details, see the TM settings at these levels:
Global-level settings made by XTM Cloud Administrators. For details, see TM tab section and the appropriate topics in it.
Customer-level settings made by XTM Cloud Project Managers. For details, see Customer Editor – Settings: TM and Terminology.
Project template-level settings made by XTM Cloud Project Managers. Project template editor – Project details: TM and Terminology.
Project-creation level settings made by XTM Cloud Project Managers. Creating a project: TM and Terminology.
For the AI Translate & TQI feature, AI Words are calculated as follows:
number of words in the source segments that are translated by the feature x number of target languages in the project for which the feature has been enabled - number of words in the non-translatable segments - number of words in segments for which an excluded TM match type is available
The source file contains 100 words, of which 10 are non-translatable. The project has 3 target languages. 20 words in one target language are ICE matches, and the project settings exclude ICE matches from the translation scope. XTM Cloud calculates the AI Words as follows:
(100 x 3) - (10 x 3) - 20 = 300 - 30 - 20 = 250
For this project, XTM Cloud will deduct 250 AI Words from your allowance.
The source file contains 100 words, out of which 10 are non-translatable. The project has 3 target languages. 20 words in one target language are ICE matches, but the project settings do NOT exclude ICE matches from the translation scope. XTM Cloud calculates the AI Words as follows:
100 x 3 - (10 x 3) = 270
For this project, XTM Cloud will deduct 270 AI Words from your allowance.
AI Words deduction:
The moment of AI Words deduction for the AI Translate & TQI feature depends on the option selected by your XTM Cloud Administrators for the Send text for translation and scoring (for details, see Managing AI Translate & TQI at global level):
XTM Workbench (default) – the text in a segment is only sent to an LLM when a user selects the segment in XTM Workbench, in Edit mode. The translation is then provided for the selected segment and the two subsequent ones (if they meet the MT/LLM conditions set for the project and are displayed in edit mode). So, the AI Words are then deducted for each of these three segments that have been successfully translated and scored by the feature.
After analysis – the whole text is sent to an LLM during project analysis. When a user enters XTM Workbench, all segments already have translations and scores (if they meet the MT/LLM conditions set for the project and are displayed in edit mode). Depending on the project settings, these translations are either inserted in the target segment, or only displayed in the Matches docked panel. So, the AI Words are deducted during project analysis for all project segments that have been successfully translated and scored by feature.
After analysis and in XTM Workbench – the whole text is sent to an LLM during project analysis. The text can be resent to the LLM on a segment-to-segment basis as in the In XTM Workbench option. So, during project analysis, the AI Words are deducted for all project segments that have been successfully translated and scored by the AI Translate & TQI feature. However, if the LLM fails to translate and score some segments during project analysis, a user can trigger this action in XTM Workbench. To do so, they must select a segment in XTM Workbench in Edit mode. If LLM-translation and -scoring succeed at this manual attempt, the AI Words for this newly translated and scored segment will now be deducted. AI Words will NOT be deducted again for two subsequent segments if they have already been translated and scored during project analysis.
AI Words calculation:
Only words from source segments are treated as AI Words.
Words from the translation memory (TM), terminology, target segment, and context segments (which are also sent to LLMs by the AI Post-Edit feature) are NOT counted as AI Words.
Furthermore, the AI Post-Edit feature automatically excludes target segments that are locked, empty, or contain invalid inline tags (too many, too few, or different inline tags than in the source segment).
For the AI Post-Edit feature, AI Words are calculated as follows:
(number of words in the source files that are to be post-edited by the feature x number of target languages in the project for which the feature has been enabled) - number of words in non-translatable segments in which the target text is identical to the source text - number of words in segments with invalid inline tags - number of words in locked segments* - empty target segments - target segments that contain tracked changes
* Locked segments are excluded from the scope in XTM Cloud 26.2 as of package number 260522_0824. To check your XTM Cloud version, see Checking your XTM Cloud version and package number.
The source file contains 100 words. The project has 2 target languages. Both target language workflows have an automatic step that uses the AI Post-Edit feature. No segments are non-translatable, empty, locked, or contain invalid inline tags. XTM Cloud calculates the AI Words as follows:
100 x 2 = 200
For this project, XTM Cloud will deduct 200 AI Words from your allowance.
The source file contains 100 words, of which 20 words are non-translatable. The project has 3 target languages. All target language workflows have an automatic step that uses the AI Post-Edit feature. In one target language, a Linguist forgot to add inline tags to two segments, containing 10 words each. In the second target language, 20 words are located in locked ICE segments and 30 in locked Leveraged match segments. In the third target language, a segment containing 10 words has not been translated by a Linguist. XTM Cloud calculates the AI Words as follows:
(100 x 3) - (20 x 3) - (10 x 2) - (20 + 30) - 10 = 300 - 60 - 20 - 50 - 10 = 160
For this project, XTM Cloud will deduct 160 AI Words from your allowance.
Important
If a project is rolled back to the workflow step preceding the automatic step that uses the AI Post-Edit feature, and the feature is then run again, the AI Words will be deducted again. This time, they will also be deducted for non-translatable segments. The reason is that these segments have now gained a revision.
AI Words deduction:
AI Words are deducted when the automatic workflow step that uses the AI Post-Edit feature is activated. However, AI Words are only deducted for segments that have received a successful LLM response.