Trust3 AI Launches AgentDOS for Enterprise AI Oversight

Trust3 AI has unveiled AgentDOS, an enterprise control plane offering real-time visibility into AI agent activity, data access, and token consumption across platforms like Databricks Agent Bricks and Microsoft Copilot Studio. The tool addresses growing risks from autonomous agents acting beyond their intended scope while helping organizations meet regulatory requirements under the EU AI Act and NIST AI Risk Management Framework. Unlike developer-focused observability tools, AgentDOS targets security, compliance, and governance teams, providing policy-driven token monitoring, automated agent discovery, and full decision traceability. In a real-world case, a healthcare provider using AgentDOS identified agents accessing regulated patient data without authorization and caught a token overrun that would have exhausted monthly budgets in 11 days.