By default on our own servers in Leiden.
By default open-source models run on our own hardware in Leiden. When a more suitable model appears, we switch without changing your configuration.
AI
By default AI runs on our own hardware in Leiden, with open-source models. Your data then does not go to third parties. If you want to use an external AI provider, that is possible per component, by agreement. AI advises, people decide.
Principles
These agreements apply to every use of AI on Twera.
By default open-source models run on our own hardware in Leiden. When a more suitable model appears, we switch without changing your configuration.
If you want to use an external AI provider, that is possible per component and by agreement, with a data processing agreement. Without that choice your data does not go to third parties.
AI helps with writing, the completeness check and preparing the assessor. A decision is always taken by a person.
With every advice you see which document or passage it is based on. So you can check it yourself.
We test an application of AI on your own examples before it goes live, and again periodically after that.
Every process works fully without AI. You decide per component whether AI takes part.
Use
AI prepares the work, the assessor and the applicant keep the overview.
Regulation
The European AI Act sets requirements for transparency and human oversight. Twera is set up for them.
We record which AI applications run in your environment, for what purpose and with which model. If you choose an external provider, we also record which one and under which processing agreement.
With every advice it stays visible that it comes from AI and what it is based on. The decision rests with a person.
Request an introduction. You speak to the people who build the system.