EU AI ACT · OPERATIONAL EVIDENCE
Keep the evidence your AI operations need.
The EU AI Act asks organisations that use AI systems to follow instructions for use, assign human oversight, monitor operation, keep logs and act on incidents; some systems also carry transparency duties. RADAR helps you record and evidence that work for instrumentable generative-AI applications and agents. It does not certify compliance, decide your legal role, or replace legal advice.
Record a reasoned screening of each registered system against prohibited practices. The conclusion remains yours.
Record legal role, risk classification and obligation applicability with rationale and reviewer approval. These are your evidence records, not binding legal classifications.
Name the people responsible for each system, with authority, training evidence and escalation path. Findings that need judgement go to a review queue; decisions and rationale are recorded.
Keep instructions for use and approved operating constraints next to monitoring results. Events your application sends are kept in a tamper-evident audit trail, with retention set per system.
Record applicability and exceptions, and evidence that required disclosures and content marking exist: text, location, owner, approval and version. RADAR does not insert notices into your application.
Manage findings, incidents and corrective actions with owners, due dates and notification decisions. RADAR does not decide whether an incident is reportable and does not submit reports.
How RADAR helps
- AI-system register with applicability records
- Activity via HTTP API or Python SDK
- Human review queues and recorded decisions
- Incidents and corrective actions
- Article 50 transparency evidence
- Evidence reports in JSON, HTML and PDF
Need something specific?
Tell us about your requirements and infrastructure. We will get back to you.
Talk to us