ServiceNow launches AI Passport to govern enterprise AI agents

ServiceNow has unveiled a new AI Passport that attaches a cryptographically signed credential to each AI agent. The passport records the agent’s identity, the tools and data it may use, and the owner who authorized it. ServiceNow’s AI Control Tower issues the passport and checks it at runtime, stopping an agent that tries to act outside its approved scope. The technology is built on open specifications and is being integrated with NVIDIA’s Open Agent Safety Platform, which can enforce limits at the infrastructure layer and shut an agent down instantly if needed.

Runtime enforcement

Beyond passport issuance, ServiceNow’s platform adds several governance layers. It can detect shadow AI agents, map access to actual entitlements, and keep a full audit trail of every decision. The company also offers cost governance that tracks spending by user, department, and runtime, and ties consumption to business value. For financial institutions, an Operational Resilience add‑on combines governance, risk and security in one place, giving regulators the audit evidence they need and allowing quick containment of rogue agents.

In parallel, ServiceNow is expanding employee‑centric AI tools. The updated EmployeeWorks harness lets staff build task‑specific assistants and routines without writing code, using drag‑and‑drop interfaces. ToolSmith can create integrations from natural‑language instructions, drawing on a library of more than 75,000 APIs. A Conversations API lets organizations embed these assistants in internal apps and messaging services. These features are available in controlled availability and use the same AI credit model as other ServiceNow products.