A recent update to SharePoint Premium Document Translate now allows organisations to create custom glossaries in a file so specific terminology is always translated a certain way or even not at all. Many organisations will have specific acronyms, industry specific terms, brands, trademarks, employee names where they want to override how they would normally be translated . A glossary file allows you to specify rules in a file – which can then be used when translating documents.
The custom glossary feature is beneficial for translating documents that are already in the target language (e.g., English) but contain a lot of industry-specific jargon. It is also useful for documents that need to be translated into another language while allowing further control, such as ensuring specific words are translated a certain way or avoiding the translation of trademarks, brands, and certain terms. Additionally, you can have different glossary files depending on the target language for the translation.
I have been a big fan of SharePoint Premium Document translation and seen great uptake as the translations are very accurate and also only take minutes compared to waiting for a human to do it. So it’s great to see this next evolution of document translation using glossaries. My other blogs on SharePoint Premium Document Translation are below:
Glossary files (link to Microsoft Docs) can be in various formats, such as CSV, TSV, or TAB files, or the more complex XLF file, which is a commonly used file type for translation data between different localisation tools and systems. The terms and phrases within the glossary are case-sensitive.
Road Testing Document Translation Using a Custom Glossary File
I will now show you how you can build a glossary file and show how it can work for translations with a document. I will show how the glossary can be used in native language (English) and be used to translate to different language i.e. French.
My Glossary File
Here is how my glossary file appears in a CSV file – note no column headers and just two columns Source Language and Destination Language.
The Document Translation service uses the Source Language column to match exact words or phrases. A key point is that it is case-sensitive and only matches words with the exact capitalisation i.e. will match on Bike but not bike. Once these are found, it replaces them with the matching value in the Target Language column.
I have also included a table of my glossary file below, where I have added a third column titled Objective. This column is not used in the actual glossary file but serves to provide my narrative explaining the purpose of each row/rule.
| Source Language (Exact Case) | Target Language (Exact Case) | Objective (if glossary file used) |
|---|---|---|
| Bike | vélo | Show if the word Bike is used it will always be changed to vélo (French word for bike) |
| Bottle | Bidon | Show if the word Bottle is used it will always be changed to Bidon (French word for water bottle) |
| Echelon Adventures | Echelon Adventures | Show that if the phrase Echelon Adventures (name of the cycling company) is used it will always remain as Echelon Adventures and wont ever be translated into native language. |
| Mountain | Mountain | Show that if the phrase Mountain is used it will always remain as Mountain and wont ever be translated into native language. |
| Fox | UppercaseFox | Show the the Source Language is case sensitive so the word Fox with a capital F results in a different translation “UppercaseFox”. Than lower case fox (see below) |
| fox | lowercasefox | Lower case fox has a different translation “lowercasefox” to uppercase Fox (see above) |
| UK | United Kingdom | Shows that if the acronym UK is used it translates to United Kingdom |
| BAU | Business As Usual | Shows that if the acronym BAU is used it translates to Business As Usual. If the word is formatted any different i.e. B-A-U it will not be identified and translated. |
| KPI | Key Performance Indicators | Shows that if the acronym KPI is used it translates to Key Performance Indicators |
| ROI | Return On Investment | Shows that if the acronym ROI is used it translates to Return On Investment |
| CEO | Chief Executive Officer | Shows that if the acronym CEO is used it translates to Chief Executive Officer |
| John Rabbit | John Rabbit | Wanted to show that often people’s names can have words in them that would be translated and to avoid this – the name can be added to the glossary so the name remains in the original format |
| rabbit | lapin | Show if the word rabbit (lower-case) is used it will be changed to lapin (french for rabbit) but when “John Rabbit” is used it will not be translated. |
Once I was satisfied with the glossary file, I uploaded the CSV to the library along with the file intended for translation.
The glossary file can also be added to another library within the same site collection, and a link can be used. However, the key criteria is that the glossary file must be in the same site collection as the source documents; otherwise, it will need to be moved.
The Test File Used
I then used Copilot to create me a word document with all of my target words in and made it around 150 words. I then highlighted all the words to be translated in red text so we can easily identify the target words.
Use The Document Translation Service in SharePoint
Navigate to a SharePoint document library where the document is that needs to be translated. For further information on getting started with SharePoint Premium document translation refer to my earlier blog or the Microsoft learn docs.
The document to be translated needs selecting in the library and once selected go to Translate – which is accessed either via the document menu or document library menu button (see below).
A Translate documents pop out then appears, asking you to enter a language to translate the document into along with a new optional section to select a glossary file.
Specify the Language to translate the document into (only one language can be chosen if using a glossary) and then chose the glossary file to use. Once finished press Translate.
A confirmation message is then displayed which confirms the document is being translated and the translated copy will be saved to this library.
The translated document is then added to the library after a few minutes – it is also appended with the language code of the translation i.e. _en for English.
I will now show where I have translated the document using the glossary file in English and French.
Translated using custom glossary English -> French
I will now display the translated document (French to English) on the right, alongside the original document on the left for comparison. Click on the image to expand it for better visibility. I wanted to show that a glossary can be used to translate documents into a different language and there are rules in the glossary i.e. if a particular word or phrase is listed it is translated into a specific word or phrase.
Below is my summary of the translated document:
- UK has been translated to United Kingdom – correct!
- The CEO John Rabbit has not been renamed to the french equivalent to Rabbit lapin – correct!
- CEO has been translated to Chief Executive Officer – correct!
- The companies trademarked name Echelon Adventures has not been translated – correct!
- Bike has been changed to velo – correct!
- Mountain has not been translated – correct!
- Bottle has been translated to Bidon – correct!
- Fox (with a capital F) has been translated as UppercaseFox – correct!
- But fox (all in lowercase) has been translated as lowercasefox – correct!
- The acronyms KPI & ROI has been translated into their full meanings – correct!
- Translation of KPIs (not listed in glossary file) has been performed EN->FR to now just KPI but is then not translated like KPI should be – point to note!
Translated using custom glossary English -> English
I will now display the translated document (English to English) on the right, alongside the original document on the left for comparison. Click on the image to expand it for better visibility. I wanted to show Document Translation could be used to translate documents just for translating industry specific jargon etc and not just translating documents into a foreign language.
Below is my summary of the translated document:
- UK has been translated to United Kingdom – correct!
- The CEO John Rabbit has not been renamed to the french equivalent to Rabbit lapin – correct!
- CEO has been translated to Chief Executive Officer – correct!
- The companies trademarked name Echelon Adventures has not been translated – correct!
- Bike has been changed to velo – correct!
- Mountain has not been translated – correct!
- Bottle has been translated to Bidon – correct!
- Fox (with a capital F) has been translated as UppercaseFox – correct!
- But fox (all in lowercase) has been translated as lowercasefox – correct!
- The acronyms KPI & ROI has been translated into their full meanings – correct!
Summary
Document Translation using custom glossaries is a great addition to SharePoint Premium. One that I feel will be very useful to many companies especially multi-national companies. Ensuring their names, trademarks, models, acronyms etc get translated the correct way, The AI in Azure AI Translator Service which SharePoint Premium Document Translation uses under the hood for translation is good but it does translate literally so might not know the nuances of language in your company/region etc.
When I first tested SharePoint Premium Document Translation in February 2024, I was very impressed with its ability to translate documents into 133 languages. The translations are completed within minutes, and all formatting of the documents was retained. I also had native speakers review the translated documents, and they were very impressed with the quality, estimating it to be around 95%+ accurate. It is much quicker than waiting for a human to translate the document and also not having to send the document out of the tenant to an external organisation etc. It’s also super cheap in comparison!
The only improvement I’d like to see in the glossary file is the addition of a third column to the CSV named “IsCaseSensitive,” with values TRUE or FALSE. If FALSE, the translation will match any combination of capitalisation for a word/phrase (e.g., “Bike” will match “bike,” “Bike,” or “BIKE”) instead of having to list every capitalisation combination in a different row. Currently all the matches for the glossary file need to be in the exact same case as they appear in the document.
I’d encourage you to try SharePoint Premium Document Translation today in your organisation and allow users to translate documents direct from document libraries and now use the glossary service. Check my other blogs for how to automate the translation of documents using PowerShell or Power Automate. On the automation story for glossaries – check back soon for another blog where I will explain how documents can be translated programmatically using glossaries with PowerShell!
Thanks for reading – I hope your organisation really benefits from Document Translation with glossaries – do let me know how you are using it and whether you have any questions?
