ServiceNow calls for measured response to rogue AI agents

ServiceNow outlined a five-step framework for governing AI agents: discover, observe, govern, secure and measure. VP of AI platform security Bhakti Pitre said a "kill switch" is not just a button, and decisions to pause or revoke an agent's access must weigh risk against business context.

• ServiceNow frames AI governance as five steps: discover, observe, govern, secure, measure
• Containment should scale with risk: pause on failure, pull the plug on limit breaches
• The company uses Veza identity security tech to fix excessive agent permissions
• ServiceNow and Okta are working together to secure agent session tokens and identities