Multilingual Quality
The Multilingual Quality module records scoped evidence for language coverage, translated or bilingual outputs, terminology usage, parity checkpoints, and unresolved language gaps.
Use it when a tenant needs to prove that a governed workflow supports the selected languages and that translated, bilingual, or terminology-sensitive outputs were reviewed under the tenant’s operating model.
What It Owns
Multilingual Quality owns:
- language quality profiles with required language keys and optional terminology refs;
- language evidence records for UI catalog parity, AI output language, translation review, terminology usage, bilingual reports, OCR language confidence, and training-material language;
- translation review decisions;
- language parity checkpoints for bilingual, terminology, RTL layout, report completeness, AI-output, and training-material parity evidence;
- language gaps for unsupported languages, provider setup gaps, missing terminology, RTL layout gaps, translation review gaps, and bilingual parity gaps;
- audit export fields, compliance-posture signals, activity entries, and work item descriptors for parity checkpoint and language gap review.
Do not paste full source text, full translated text, raw OCR output, or confidential document content into language quality snapshots. Use source refs, language keys, terminology refs, review decisions, status labels, and scoped evidence summaries.
When To Configure It
Configure Multilingual Quality when a tenant expects bilingual or multilingual evidence in governed workflows, including AI-assisted document treatment, audit packages, training material, or customer-facing reports.
Use configured when the profile, required languages, terminology refs where needed, evidence records, and review path exist. Use not-configured when a tenant has not defined language quality expectations. Use provider-not-configured when a translation, OCR, or AI provider is required but unavailable. Use evidence-required when the capability can be supported but proof has not been recorded.
Operating Model
Start with a profile for the workflow or content family. Record the required language keys, default source language, and terminology set/version refs where terminology is governed.
Record evidence when language-sensitive work is produced or reviewed. Record parity checkpoints when a bilingual, terminology-sensitive, RTL, AI-output, report, or training-material parity check is complete and needs a durable status. Record gaps when the product or tenant setup cannot satisfy a required language expectation. Assign parity or gap review through the shared work-management descriptors rather than creating a local task truth.
For no-AI deployments, use the same profile, evidence, review, and gap records for manual translation review or deterministic language checks.
Audit And Posture
Audit packages and posture views should use the module’s descriptors for profile key, evidence kind, evidence status, review decision, parity checkpoint kind/status, and gap status. These descriptors prove the language quality state without exporting raw translation content.
Multilingual Quality can reference Terminology, but Terminology remains the source of truth for approved terms and versions.