ServiceNow urges nuanced approach to rogue AI agents

ServiceNow is pushing enterprises to treat AI agent containment as more than a simple kill switch. Bhakti Pitre, VP of AI platform security, argues that responses must weigh business context alongside risk. A poorly performing agent may need only a pause, while one manipulated into offering unauthorized 100% discounts warrants full shutdown. ServiceNow maps agents to dependent business processes and uses Veza's identity tools to trim excessive permissions. The company is also partnering with Okta to secure agent session tokens and identities, calling for broader industry collaboration on AI governance.