Terminology integration
Integrate your specialised terminology directly into AI systems.
So that AI can reproduce your corporate language consistently and accurately worldwide.
How the integration works
Terminology integration can be carried out in two ways
During glossary integration, your specifications for the respective language direction are mapped to one another. Each source term is assigned a preferred target term. Clear rules apply here so provided the data is accurate, the mapping can be carried out automatically. This method is particularly suitable when only parts of the database are to be used as a glossary and when a flexible approach is required to deal with ambiguities or abbreviations.
With database integration, the AI system accesses your terminology directly from the database. The mapping takes place in the background without any manual intervention or separately generated file. This method is particularly effective when the database is constantly being expanded as it always takes the current status into account.
Why terminology integration is essential
AI systems are not aware of your company guidelines. They output technical terms based on word probabilities rather than on whether they are technically correct. Up to 45 percent of corrections to AI translations relate to terminology errors. This results in inconsistent and technically imprecise texts that require a great deal of post-editing.
With terminology integration, your specialist terminology becomes the standard of reference. Whether it’s a glossary or a database, the AI system draws on your specifications and applies them directly – ensuring accurate specialised terminology and a noticeable reduction in the amount of review work required.
Prerequisites for successful integration
Three factors are crucial for seamless terminology integration:
- Unambiguous terminology data with clear guidelines on preferred, permitted or prohibited terms
- A clean database containing only relevant terminology and which does not adversely affect the system
- Technical compatibility: the AI systems used must be capable of integrating glossaries or databases. Our oneSuite offers seamless integration with common machine translation systems and large language models.
When is terminology integration appropriate?
Terminology integration is particularly worthwhile when:
- Your AI or machine translations produce inconsistent results – such as different terms being used for the same concept.
- The amount of post-editing work required for AI translations has so far been very high.
- Your texts contain a great deal of specialist terminology.
- You already have a terminology database, but it has not yet been integrated into AI processes.
- You would like to introduce machine translation or AI-powered tools.
- Your technical terminology is to serve as a knowledge base for chatbots or internal systems.
If you do not yet have any guidelines in place, we can assist you in setting up a new database – from relevant term extraction through to structured maintenance.
Our terminology integration services
We handle the entire terminology integration process, from analysis through to implementation. We review your existing terminology or create a new terminology database – structured, multilingual and unambiguous. We will integrate all your terminology – or terminology that has been prioritised in advance – into the AI system of your choice, either as a glossary or a database.
Following integration, we assess the reliability of the artificial intelligence in adopting your terminology, identify potential areas for optimisation and adjust glossaries and databases accordingly.
FAQs
What is the difference between a glossary and a terminology database?
A glossary is a concise list that maps a preferred target term to each source term. A terminology database contains additional information such as usage rules (preferred, permitted, prohibited), synonyms and contextual examples.
Into which AI systems can terminology be integrated?
We take a technology-neutral approach and adapt your data for use with common machine translation systems (e.g. DeepL, Google Translate) and large language models. We will clarify the exact file formats and interfaces during the initial consultation.
Can I incorporate terminology even if I don’t yet have any guidelines?
Yes. To build your terminology, we extract relevant terms from your texts and create a unique mapping that can be used directly in AI systems.
Is integration a one-off process?
No, terminology work is a dynamic process. New terms are added and old ones are removed. With database integration, the latest version is always available, whereas glossary integration requires regular updates.
Further services in the field of terminology.
Let’s discuss your next project!
Please use our contact form or give us a call:
Make an appointment now:
