Every brand has words that shouldn’t be translated — product names, taglines, trademarked terms — and words that need one specific, consistent translation across every page. Without a managed glossary, AI translation can inadvertently translate a brand name literally in one article and leave it untranslated in another, creating an inconsistent experience across a site.
What Belongs in a Translation Glossary
- Product and company names that should stay in their original form across every language.
- Taglines and slogans that were deliberately crafted and shouldn’t be auto-translated word-for-word.
- Industry-specific technical terms that need one approved, consistent translation.
- Feature names and UI labels that should match across marketing copy, documentation, and the product itself.
How a Glossary Improves AI Translation Output
Supplying a glossary to the AI translation process constrains its output to use approved terminology consistently, rather than letting the model choose fresh phrasing for the same term on every page. This matters most for terms with multiple valid translations, where consistency across a site is more important than any single translation being marginally more natural.
Building and Maintaining a Glossary
| Step | Purpose |
|---|---|
| Audit existing content for brand-specific and technical terms | Identifies what needs glossary protection before scaling translation |
| Define approved translations per target language | Removes ambiguity for terms with multiple valid translations |
| Update the glossary as new products or features launch | Keeps terminology consistent as the brand and product evolve |
Frequently Asked Questions
Do all brand names need to stay untranslated?
Most do, though some brands intentionally localize their name or tagline for specific markets — the glossary should reflect whichever decision the brand has made per language.
How often should a glossary be updated?
Whenever new products, features or significant terminology are introduced, ideally as part of the product launch process rather than after translation inconsistencies are discovered.
Can a glossary cover multiple languages at once?
Yes, a well-structured glossary maps each term to its approved translation per target language, so it can guide AI translation consistently across every locale a site supports.
Conclusion
A maintained glossary is what keeps AI-translated content feeling like one coherent brand voice rather than a collection of independently translated pages. The upfront effort of defining approved terminology pays off in consistency across every language a site scales into.