At SailPoint Navigate in Sydney, Chris Cubbage speaks with SailPoint Chief Product Officer Levent Besik about securing the growing number of AI agents in the workplace. Bessick describes three key challenges: the scale of non-human identities, the complexity of agents acting with human credentials and access, and the speed at which risks can unfold.
They discuss why sandboxing alone isn’t enough, the importance of assigning clear ownership to every agent, and a three-stage approach to agent governance: discover, govern, and protect. The conversation also covers real-time authorization, preventing sensitive data uploads, integrating with existing security tools, and setting guardrails around high-impact or irreversible actions.
Chapters
- 00:00 Welcome to SailPoint Navigate Sydney
- 00:56 The scale, complexity, and speed of agentic AI
- 02:52 Why traditional IAM and PAM approaches may fall short
- 05:19 Sandboxing, exposed credentials, and security hygiene
- 06:24 Agent ownership and accountability
- 08:05 Discovering and governing agents
- 11:08 Protecting agents and the enterprise in real time
- 14:42 Integrating with existing identity and security tools
- 15:47 User experience and permission controls
- 17:01 What agents should—and shouldn’t—be allowed to do
- 19:06 Takeaways from SailPoint Navigate in Australia
#AgenticAI #IdentitySecurity #IdentityGovernance #Cybersecurity #SailPoint #Navigate
