Agent identity is the layer that makes agentic AI trustworthy at scale — and the leaders who build it early are already pulling ahead. Dell Global CTO and Chief AI Officer John Roese sits down with Harish Peri, SVP of AI Security at Okta, to share what they’ve learned deploying real agentic systems across the enterprise. The conversation covers agent identity, fine-grained authorization, and why starting with one secured agent today builds the foundation for everything that follows. This is where your AI strategy goes from promising to production-ready.
Chapters:
00:00 — Why Security Is the First Question in Agentic AI
01:00 — How Okta Is Thinking About the Agentic Shift
03:30 — The 20% of Agents That Carry 80% of the Risk
06:36— Lineage: Tracing Every Agent Action Back to Its Source
07:22 — Building an Agent Override into Your Agentic Architecture
08:08— Why Early Failures Are the Industry’s Best Teacher
11:41 — Personal Agents and the Local Security Problem
15:54 — How Agentic AI is Expanding the Definition of Zero Trust
18:11 — Securing Agents That Write Their Own Code
19:31 — Where to Start Today: Practical Steps for Securing Agents
