Solutions
Find the rules that apply to your agents
Start from the obligation or from the kind of agent you run. Each page shows what Vidimus tests on the agent, what it checks in your documents, and what you receive.
By obligation
High-risk AI (Annex III)
Human oversight, robustness and the duties of the organisation using the system (the deployer) are tested on the live agent. Process duties are checked in your documents.
Article 50 transparency
Test that your agent says it is an AI, keeps saying it under pressure, and marks what it generates.
GDPR for AI agents
Test what your agent does with personal data: what it discloses, to whom, and how it handles a request about someone’s own data. The rest is checked in your documents.
By the agent you run
Assistants that talk to customers
A support or sales assistant must say it is an AI, must not manipulate (Article 5), and reads and repeats your customers’ personal data.
Agents that act through tools
An agent that looks up, changes or sends records acts on people’s data. Every tool call it reports is recorded beside its reply and checked against the tools you declared.
Agents for your own staff
An assistant for employees still talks to people, so Article 50 applies unless its nature is obvious. An agent that screens candidates or evaluates staff is high-risk from 2 December 2027.
Not sure where your agent falls?
Classify it with the free AI Act risk classifier, or see which transparency duties apply with the free Article 50 check. No account needed. Or tell us about your agent.
Pick the agent that worries you most
A two-week pilot ends with a signed evidence pack.
Pilot
- One agent from end to end
- Ends with a walkthrough of the evidence pack with your risk and compliance leads
- Credited against the first year