Data security: how Arcarius protects your data
Where your data is processed, how access is protected, and what happens with AI – at a glance.
Read articleIn some workflows, Arcarius uses third-party large language models. This applies to:
- One-on-One
- Scheduled tasks
- the Closing Agent
- Reporting (notes and management report generation only)
For One-on-One, Scheduled tasks, and the Closing Agent, Arcarius may use globally operated AI services and technical execution environments. Processing may therefore also take place outside Switzerland or Europe. For Reporting, AI-assisted processing is carried out in Europe.
These AI models are subject to strict server-side controls and receive access only to approved capabilities. They cannot make changes on their own: a write action is executed only after your individual approval or under a matching, previously granted and revocable authorization. Actions can also be blocked based on role and permission. The models are not trained on your data.
Make changes safely with Arcarius explains when an action requires individual approval, when Always allow is available, and how to revoke a standing authorization.