How DB customizes with glossaries

For global teams like DB, it can be challenging to navigate complex technical terminology, juggle product names, or consistently use company-specific vocabulary. With DeepL’s glossary feature, DB achieves the consistency and customizability its business requires. Consistency in communication across languages reduces costs overall by saving teams time and effort on manual editing.

How does such a massive organization manage its glossaries? DB maintains a central “terminology database” with vocabulary that requires specific (sometimes unintuitive) translations to handle industry-specific terms. The database contains nearly 30,000 entries in up to 16 different languages.

The DB Language Management Department manages the terminology database. They update the corresponding glossaries in their DeepL account every few weeks to ensure that all translations are as current as possible.

“Along with data protection and data security, the ability to adapt the engine to DB’s corporate language via glossaries is a main advantage of DeepL compared to other machine translation systems. [. . .]”

Tom Winter, Data Scientist & Computational Linguist at Deutsche Bahn AG

Glossary best practices

DB has invested a considerable amount of time into maintaining and mastering its glossaries. The company’s main piece of advice to other glossary users? Avoid mistranslations by mapping all unambiguous synonyms to one preferred target term, and only include unambiguous terms.

What’s next for DB?

DB started using DeepL to enable internal communication in January 2022, and has been heavily investing in it across the DB Group. DeepL Translator is used across departments for text and document translation, and is available to all employees as a browser extension.

Moving forward, DB is exploring use cases where AI and LLM technology might be able to assist with railway operations, such as communications for border traffic.