The proposed integration seeks to eliminate the need for secondary, fragmented control systems by embedding security directly into the NeMo workflow. By leveraging its background in open-source standards like Apache Ranger and Apache Atlas, the company intends to provide visibility into previously untracked 'Shadow AI' and enforce access policies tied to specific agent purposes rather than static roles.

Neeraj Sabharwal, co-founder of Trust3 AI, notes that the platform is designed to function beneath agents to ensure security teams can monitor decisions, audit data lineage, and verify compliance with frameworks like the EU AI Act and the NIST AI Risk Management Framework. This approach complements existing tools like NeMo Guardrails, providing a vendor-neutral layer that captures prompts, tool calls, and retrieval processes to ensure auditability without impeding development velocity.