Quality Control
A defined workflow to ensure consistency, accuracy, and reviewable MTPE output.
Why Quality Control Matters
In MTPE and AI-assisted translation, output quality depends not only on language skill,
but on how translation, revision, and verification are structured.
Quality control is therefore a process, not a claim.
Quality Control Principles
-
Defined scope for each service level
-
Clear separation between translation and revision
-
Terminology consistency across the document
-
Reviewable and auditable output
Workflow Overview
Standard MTPE
-
Review of source content and MT output quality
-
Post-editing according to defined MTPE level
-
Terminology consistency check
-
Final review before delivery
Enhanced MTPE / AI-assisted Translation
-
AI-assisted or MT-based draft
-
Human-controlled linguistic revision
-
Additional domain-specific terminology checks
-
Final consistency and readability review
Terminology Handling
Terminology consistency is treated as a core quality factor, especially for technical, academic, and regulatory content.
Depending on project scope, terminology may be:
-
Verified against client-provided references
-
Checked for internal document consistency
-
Reviewed for domain-appropriate usage
Reviewable Output
All deliverables are produced through controlled workflows rather than raw machine output.
This ensures that:
-
Linguistic changes are intentional
-
Decisions can be reviewed
-
Output quality remains stable across similar projects
What Quality Control Is Not
-
Not stylistic rewriting beyond the agreed scope
-
Not creative or marketing copy optimization
-
Not a substitute for subject-matter expert validation
Relationship to Service Levels
The depth and scope of quality control depend on the selected service category
(MTPE – Standard, MTPE – Enhanced, or AI-assisted Human Translation).
Quality control is applied consistently within each defined level,
rather than adjusted ad hoc.