Skip to scenario
Translation on demand

Translate the task with the labels the user sees

A German-speaking administrator opens a new topic before a reviewed translation is available. On-demand translation can help only if product terms and interface labels remain recognizable. A termbase and a clear review state give users and translators a stronger starting point.

See the comparison
The task in this example

Show this administration topic in German.

Business goal
Build brand loyalty
Business metric
Retention

The task should survive the language change

The same administrator uses the same German interface. The comparison changes whether the translation has approved product terms and UI-label mappings.

Product terms drift

The translation model has no approved termbase.

Forecasting appears under several translations that do not match the German interface.

Source topic
Current English administration task
UI labels
Uncontrolled translations
Review state
Unclear
The administrator has to infer the command

Terms match the interface

The workflow supplies approved German terms and preserves translation status.

The topic uses the labels in the German interface and is identified as awaiting translator review.

Source topic
Same current English task
UI labels
Approved German mappings
Review state
On-demand translation; review pending
The administrator can assess and follow the task

Check task parity before connecting language access to retention

Translation availability is an opening. The near-term test is whether administrators complete the same task with comparable success; renewal requires a longer observation period and other account context.

  1. Preserve terms

    Use approved product names and UI labels.

    • Approved terms used
  2. Review language quality

    Record corrections and their causes.

    • Translator edits
  3. Test the task

    Compare completion for equivalent tasks and users.

    • Task parity
    • Non-English tickets
Business metric

Retention

Follow retention for comparable accounts with documented language needs; treat task parity as an earlier diagnostic measure.

What to measure and how to interpret it

Use matched tasks and similar experience levels for German and English users. Record source revision, termbase version, translation model, and review status.

Approved terms used
Share of checked product terms and UI labels matching the approved German termbase.
Translator edits
Substantive corrections per reviewed topic, distinguishing terminology, meaning, and style.
Task parity
Difference in observed task completion between comparable German and English user groups.
Non-English tickets
Language-related help requests for the tested tasks, separated from unrelated product problems.

Check the business result

Assess language-related account feedback and renewal patterns without attributing every change to translation. Include translator review and corrections in content cost. A pending translation may be unsuitable for content whose approved process requires review before use.

Link terms to the interface and source revision

A short self-contained topic is easier to translate and maintain. The workflow must keep its source identity and review state visible when delivering a machine-produced version.

Administration / German / On-demand
Translation package

Source and terminology context

Translate the current topic using approved German product terms and UI labels.

Source
Topic ID and current English revision
Language
German
Term constraints
Approved termbase and UI-label mappings
Delivery state
On-demand; translator review pending
Maintenance
Retranslate or review when the source or termbase changes

Intelligent content

Maintain self-contained topics and source revisions so translation can target the relevant task and detect later changes.

Taxonomy

Use a termbase with approved translations and interface mappings; preserve product names and other terms that should remain unchanged.

A practical starting point

Pilot one frequently used administration task. Have German-speaking users attempt it and translators review the output before broadening availability.

The model translates remaining prose. Localization owners approve terms and determine which content can be delivered before review.

See how this scenario works

Decide what may be shown before review

Agree with localization and content owners which topic types may use on-demand delivery, how pending review is identified, and where users can report a confusing translation. The policy should follow the content’s use and consequences.

Lief
Lief
Amber
Amber
Dan
Dan
Explore the work with Intuitive Stack

Which terms make translated tasks harder to follow?

Intuitive Stack can help structure source topics and terminology so your localization team can test on-demand translation against the interface and user task.

  • Approved terms and UI labels
  • Source and review-state tracking
  • Task parity across languages