WisdomInterface

Governing AI agents: from blueprint to production

AI agents do not create an entirely new access problem. They make an existing one much harder: multiple identities, broad standing privilege, fragmented policy enforcement and weak attribution – all moving at machine speed.

In this session, Neha Duggal, Chief Product Officer at P0 Security, uses the Blueprint Alliance’s Governing Agentic Execution reference architecture as a practical starting point. She breaks down the six zero-trust principles behind the Blueprint and focuses on the two capabilities it says remain aspirational with today’s tooling: task-scoped access and end-to-end delegation.

You’ll learn how to:

  • Preserve the originator, acting agent and delegated context across the full action chain
  • Evaluate policy at the moment an agent attempts an action, not only at sign-in
  • Grant only the privilege required for the approved task, then remove it when the task ends
  • Enforce approved access in the target system instead of stopping at the gateway or tool layer
  • Record the policy decision, approval, privilege granted, action and outcome as one traceable chain
  • Start with one high-risk agentic workflow and expand governance from there

We’ll also follow a real action chain end-to-end so you can see where identity, policy and enforcement need to travel together in practice.

This session is for security, identity, cloud and AI platform leaders who are moving beyond agent inventory and asking the next question: now that we know the agents exist, what are they actually allowed to touch?

SUBSCRIBE

    Subscribe for more insights



    By completing and submitting this form, you understand and agree to WisdomInterface processing your acquired contact information as described in our privacy policy.

    No spam, we promise. You can update your email preference or unsubscribe at any time and we'll never share your details without your permission.

      Subscribe for more insights




      By completing and submitting this form, you understand and agree to WisdomInterface processing your acquired contact information as described in our privacy policy.

      No spam, we promise. You can update your email preference or unsubscribe at any time and we'll never share your details without your permission.

      This also includes the use of email tracking technologies, such as tracking pixels, to measure email opens and engagement, as described in our privacy policy.