Okta: risk chiefs press CISOs on control over AI agents

Okta's APJ head Dan Mountstephen said chief risk officers and CFOs increasingly question whether CISOs and CIOs have a handle on AI agents. Okta polling found only 13% of Australian organisations can shut down a rogue agent across their environment, and 40% have no visibility of their agents at all.

• Only 13% of Australian organisations can kill a rogue agent environment-wide
• 40% of companies have no visibility into their AI agents
• Okta pitches runtime authorisation to terminate sessions or whole agents
• Shadow AI Agent Discovery targets agents set up on employee laptops