Protect Your Agentic AI System: Architecture, Guardrails & DevSecOps Measures Against Hijacking

Secure Agentic AI – from architecture to operation.

This session demonstrates how to secure Agentic AI systems holistically: from trust boundaries and controlled tool usage to protected knowledge sources, monitoring, and incident response. We bridge the gap between risks and concrete protective measures, showing how these can be established as DevSecOps standards for scalable, reproducible security in AI-native systems.

Speaker:

Nico Meisenzahl (COO, white duck)