The Pulse
ServiceNow Gives AI Agents Signed Runtime Credentials
ServiceNow announced AI Passport, a cryptographically signed credential that checks what AI agents and MCP servers are authorized to do at runtime. The company says the product and related shadow-AI detection and cost-management tools are e
AI.info Team · Written with AI ·
Only 24% of banks say governance matches their AI use
All banks report using AI, but just 24% say they have governance to match, according to ServiceNow’s 2026 Banking AI Maturity Index. The company announced AI Passport on October 15 as a way to check an agent’s authority while it is working, rather than relying only on records of what it did afterward.
AI Passport is a cryptographically signed credential that AI agents refer to at runtime to verify their authorized scope. The scope can cover what data and tools an agent may use and which actions it may take. ServiceNow says the system also applies to Model Context Protocol (MCP) servers and can check credentials at each step in a workflow involving multiple agents.
“As AI agents proliferate across the enterprise, visibility alone isn’t enough. Enterprises need confidence that every AI agent is operating within its authorized scope,” said Jon Sigler, executive vice president and general manager of AI Platform at ServiceNow. The company positions the credential as a control on what an agent is allowed to do, not merely an audit trail for reviewing its actions later.
ServiceNow ties each passport to enterprise records
The Super Intelligence Control Tower issues each AI agent its own passport, grounded in a company’s live business context. ServiceNow says that context can include its Configuration Management Database, governance, risk and compliance policies, and workflow approvals. For third-party agents, the credential is intended to provide a verifiable permissions record instead of relying on the agent’s own claims about its access.
ServiceNow says the credentials use open specifications, allowing governance products, identity providers and enforcement systems to issue and verify them for their own agents and third-party agents. The company also describes scoped delegation that narrows permissions as authority passes through a workflow, along with precision revocation and compliance audit trails.
If an agent moves outside its authorized scope, ServiceNow says organizations can redirect or restrict it, or reconnect it with the right context. The system also includes kill-switch capabilities for stopping an agent when needed. The announcement describes those responses as enforceable at machine speed, but does not establish how they perform across different customers’ systems.
NVIDIA is building an enforcement implementation
NVIDIA is a design and launch partner for AI Passport. The companies are building an enforcement implementation using the NVIDIA Open Agent Safety Platform, which includes OpenShell runtime software and the NVIDIA Sentry security framework.
“By implementing AI Passport with the NVIDIA Open Agent Safety Platform, ServiceNow customers can continuously verify and enforce what every agent is authorized to do — at machine speed and at runtime,” said Itay Ozery, NVIDIA’s director of Product Marketing for Networking and Cybersecurity. The partnership makes the enforcement layer a key part of the announcement: a credential can state an agent’s limits, while runtime infrastructure checks those limits as the agent operates.
AI Passport is planned for Q4 2026
ServiceNow says AI Passport, shadow-AI detection and AI Cost Management are expected to be available in the fourth quarter of 2026. Its announcement does not describe those three offerings as available on October 15. The company says the broader governance capabilities include discovery of unapproved agents, access mapping, continuous observation of agent traces and behavior, and controls for tracking AI spending against business value.
ServiceNow also cites LTM Limited as a customer with 1,500 AI agents deployed across its organization, each assigned an employee ID and a manager. LTM’s chief information security officer, Chandan Pani, said the company uses ServiceNow’s control tower as a central record for agent inventory, risk and compliance. That example shows the scale of agent deployment behind the governance pitch, but the announcement does not say that LTM is using AI Passport.
The company’s separate Operational Resilience solution for financial services is available now, according to the release. AI Passport and the named discovery and cost-management products remain scheduled for Q4 2026; the practical test will be whether runtime checks and enforcement work consistently across agents and infrastructure from different vendors.