Global AI Post-Edit settings
Business use case
The AI Post-Edit feature helps you reduce post-editing time and cost while ensuring that the final output meets your required quality standards. A key benefit is its flexibility: an automatic workflow step enables Project Managers to use Large Language Models (LLMs) to enhance various outputs, including human translation (HT), machine translation (MT), and LLM-generated translation. It can be used in either fully automated (LLM-only) or hybrid workflows (HT/MT + LLM).
The AI Post-Edit feature leverages XTM Cloud’s resources. It sends the Translation Memory (TM) for reference and has access to your terminology and project-specific Style Guide. Because of this, it can adjust the content's tone and voice automatically, while ensuring that the high-quality output adheres to company terminology and brand voice.
Note
To use this feature, you must purchase an AI Words allowance for your entire XTM Cloud instance. AI Words are source words sent to LLMs. Words from the target segment, translation memory (TM), terminology, and context segments are NOT counted as AI Words. For details about AI Words: calculation and deduction for projects, see AI Words in XTM Cloud. For pricing information, contact your account manager.
Important
Clients running the new Enterprise or Business subscription (for details, see Licensing) receive a free monthly allowance of 100,000 or 50,000 AI Words per XTM Cloud instance, accordingly. These words can be used for one or more of these AI-driven features: AI Translate & TQI (used either with or without the TQI component), AI Post-Edit, and TQI. For details, see AI-driven Features tab.
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.
If you are using your own LLM keys, the size of the source text, terminology, TM records, and context segments can influence the LLM usage costs. For details, refer to your LLM subscription rules.
The AI Post-Edit feature automatically improves the quality and accuracy of translations created by a Linguist or generated by an MT (Machine Translation) engine, or an LLM (Large Language Model). For every eligible segment, the feature sends useful details to an LLM for analysis. Depending on their availability, these details can be:
the source and target segment contents.
the source and target segments' context (the previous and the following segment).
the three highest translation memory (TM) matches that were added or updated most recently.
terminology.
Important
To grant the AI Post-Edit feature access to your terminology, ensure that you select the Match terms in analysis setting in the global terminology settings. For details, see Terminology engines. Alternatively, Project Managers can enable this setting at customer level (for details, see Customer Editor – Settings).
the automatic step settings (for details, see Workflow step settings).
Important
At least one of these automatic step settings must be made, for the AI Post-Edit feature to work:
Check terminology (recognized for the segment).
Basic QA check.
Parameters:
Formality.
Gender neutrality.
Follow locale conventions.
For details, see Workflow step settings.
the Style Guide that has been created at the global level by an XTM Cloud Administrator (for details, see Style Guide tab and then selected for a project by a Project Manager. For details, see Creating a project: AI Translations.
Once it has analyzed all available resources, the LLM is set to correct and adjust target segments in XTM Workbench to your specific requirements. The previous content of the target segment is saved in the segment history. Content entered by the AI Post-Edit feature is also displayed in the History docked panel. For details, see History docked panel).
While adjusting the target segment, the feature can:
perform a basic QA check, if this option has been selected for the automatic step. For details, see Workflow step settings.
ensure that the terminology recognized for the segment has been used correctly, if this option has been selected for the automatic step. For details, see Workflow step settings.
adjust the formality level and gender neutrality to the settings made for the automatic step. For details, see Workflow step settings.
adjust the time, date, number, measurement, currencies, and address formatting to the target language rules, if this option has been selected for the automatic step. For details, see Workflow step settings.
Some segments are automatically excluded from the AI Post-Edit feature scope and are not sent to an LLM. These segments are:
empty target segments.
locked target segments.
non-translatable segments in which the target text is identical to the source text.
target segments that contain tracked changes. For details, see Tracked changes feature.
target segments that contain invalid inline tags.
Note
If required, changes made to the target segment by the AI Post-Edit feature can be reverted. To do so, users with the Project Manager role should use the workflow rollback feature. For details, see Rollback to a previous step in the workflow.
To display the AI Post-Edit section, as a user with the XTM Cloud Administrator role, select Configuration > Settings > AI Control Center. Then, in the AI Control Center tab screen, select the Features sub-tab. AI Post-Edit is one of the sections displayed in the Features tab screen.
In the AI Post-Edit section, you see:
the Get Started button (if you are running the new Enterprise or Business subscription. For details, see Licensing). It is displayed if the AI Post-Edit feature has not been enabled for your XTM Cloud instance. Select this button to automatically generate an email message to our team, with which you can request feature activation. Initially, some values in the message you generate are inserted by default, but you can adjust them and add other details to suit your requirements.
the Contact Sales button (if you are running Team or any other subscription type. For details, see Licensing). It is displayed if the AI Post-Edit feature has not been purchased for your XTM Cloud instance. Select this button to automatically generate an email message to our Sales team, with which you can request feature activation. Initially, some values in the message you generate are inserted by default, but you can adjust them and add other details to suit your requirements.
the Activated
label. It indicates that the feature is currently active at global level.Important
Once you have activated the AI Post-Edit feature at global level, create one or more automatic steps that leverage this function (for details, see Workflow step settings). Then, you can create one or more workflow definitions that contain such automatic steps (for details, see Creating workflow definitions). Project Managers will be able to use these workflow steps or entire workflow definitions for their projects.
the Manage button. Select it to:
activate AI Post-Edit at global level. For details, see Activating AI Post-Edit at global level.
manage the AI Post-Edit settings at global level. For details, see Managing AI Post-Edit at global level.
deactivate AI Post-Edit at global level. For details, see Deactivating AI Post-Edit at global level.