Love it or hate it, Agentic AI is arriving in your organization. Our job is to enable its use safely and securely. But where do you start? What exactly is Agentic AI? Join Vis Chirravuri and Lance Spitzner as they walk you through where to start, from both a technical and human perspective. In a single hour, they cover the human and policy side alongside the technical controls because you can't manage agentic AI risk with either one alone.
Unlike a chatbot that just answers questions, an agent takes actions. It calls tools, runs code, spends money, and increasingly hands work off to other agents. That autonomy is exactly where the new risk lives, and it's why traditional application security controls aren't enough on their own.
Key takeaways include:
- What Agentic AI and AI frameworks are and how they differ from the chatbots you already know
- How to develop strategic policies and a risk-based adoption strategy for Agentic AI use
- Technical controls you can implement today, including a managed model gateway (LiteLLM), input/output guardrails, least-privilege tool access via an MCP gateway, execution sandboxing, supply-chain/provenance verification, and runtime governance
Who Should Attend
Anyone interested in or involved in securing AI in their environment, including:
- Chief Information Security Officers
- Network Architects
- DevSecOps Team Leads
- GRC Teams / Policy Teams
- Security Awareness & Culture Officers
- Vulnerability Management Teams
- Security Operations Center Teams
- Cyber Threat Intelligence Teams
Learning Objectives
- Explain what Agentic AI is and its associated risks
- Define what policies apply to Agentic AI and how
- Identify key technical controls that most effectively manage Agentic AI risk in your organization
This session supports concepts from SEC546: Securing Agentic AI and LDR521: Security Culture for Leaders.
