The Agent Is an Identity: Why AI Agents Need Least Privilege, Limited Access, and Runtime Checks
Date / Time: Thursday, August 20th, 2026 06:25 PM – 8:00 PM
Speaker: Venkata Patelkhana, of Dell Technologies
Description: AI agents are starting to do more than answer questions. They can search systems, choose tools, update records, trigger workflows, and make changes in business applications.
Once an agent can take action, it should be treated like a system identity. It needs its own identity, a clear owner, limited permissions, and rules that control what it can do.
This talk uses a simple employee-badge analogy to explain the idea. An agent is like an employee or contractor, and its identity is the badge. Least privilege means the badge opens only the doors needed for the job. Bounded access means it works only in the right building, on the right floor, for the right task, and for a limited time.
The session also explains why every action should be checked at runtime, why the target system must still protect its own resources, why access should be short-lived, and why every agent action needs a clear audit trail.
Attendees will leave with a practical way to think about AI agent security: treat the agent as an identity, limit its authority, and check every action before it reaches an enterprise system.
Biography: Venkata Phani Patelkhana is an IEEE Senior Member , with more than 19 years of experience in enterprise Application Programming Interface (API) platforms, distributed systems, API security, and agentic AI architectures.
He currently holds the title of Distinguished Engineer at Dell Techonologies, where he focuses on AI control planes, MCP gateways, agent identity, runtime authorization, and helping enterprises transition safely from API-first to agent-first ecosystems. He holds patents related to decentralized API gateway architectures and regularly writes and speaks about practical approaches to securing and governing enterprise API and AI ecosystems.
Venkata holds a BS degree in Electronic and Communications Engineering from the Jawaharlal Nehru Technological University, and has completed the Post Graduate Program in Artificial Intelligence and Machine Learning: Business Applications, AI, and Machinge Learning at the McCombs School of Business, UT, Austin.
Time: 6:30 PM
6:25 to 6:35 — Open for participants to network
6:35 to 6:40 — IEEE Life Members and Consultants Network business, followed by speaker introduction
6:40 to 8:00 — Presentation and Q&A
Location: Zoom meeting
Reservations: Please RSVP using the vTools link https://events.vtools.ieee.org/m/571370. Registered attendees will be emailed connection instructions in the 24-hour period prior to the start of the meeting. Look for an email from IEEE eNotice, from jayasudha.yedalla@gmail.com, uday.kanike@gmail.com,or from me.
You do not need to be an IEEE Member to register.
For more information about CTCN go to https://r5.ieee.org/ctx-cn/