About AIA Guard
AIA Guard exists to help teams explain — and verify — how an AI agent behaves in one specific, real-world deployment.
Mission
We make risk conversations concrete: which data the agent sees, which tools it can use, where people step in, and which evidence supports the result. Our initial focus is healthcare and finance in Switzerland, with complete information in English, German, French and Italian.
Operating principles
Scope before score. Evidence before claims. Independent review before certification. Critical failures stay visible. Results expire or change when the system changes. These are commitments written into the draft method — not claims that an operational certification body already exists.
Pre-launch status
The scheme, legal-entity details, reviewer roster and service terms are not yet published as operational facts. The public method is a draft, open for scrutiny. We assert no customers, accreditations, certificates, insurer agreements or Swiss-only hosting claims. Follow the method, try the sample dossier — and when you are ready, choose “Discuss your agent”.