ServiceNow announces the general availability of Arc, the personal long-running AI agent built for the enterprise, governed and secured by the ServiceNow Super Intelligence Control Tower and enforced with NVIDIA OpenShell
ServiceNow’s recently introduced AI Passport will be integrated with NVIDIA OpenShell and NVIDIA Sentry to help validate credentials that authorize and enforce AI agents across AI stacks
ServiceNow, the Super Intelligence Control Tower for business reinvention, announced significant updates designed to bring governance and security across every layer of AI agent execution as an expansion of its partnership with NVIDIA. Arc, a personal long-running AI agent that can take on everyday work, is now generally available. From day one, Arc is governed and secured by the ServiceNow Super Intelligence Control Tower and enforced with NVIDIA OpenShell, allowing it to work independently while the enterprise stays in control. Additionally, the NVIDIA OpenShell and NVIDIA BlueField data processing unit (DPU) enforcement layers will be able to read ServiceNow’s recently announced AI Passport solution to validate what AI agents, including Arc, have permission to access or execute. Together, the companies are building an end-to-end AI governance architecture on the ServiceNow AI Platform and the NVIDIA OpenShell Reference Architecture, where the Super Intelligence Control Tower governs once, and both software and silicon honor its decisions.
Also Read: CIO Influence Interview with Srinivasan Panchapakesan, Chief Platform Officer at Hexaware
People are turning to personal AI agents to get things done. These AI agents do more than answer questions: they plan work, use apps and systems, write code, and take actions with limited human involvement. This innovation creates significant opportunities for productivity, but also introduces a fundamental challenge of ensuring AI agents only act within their authority. ServiceNow and NVIDIA address this problem by moving enforcement beyond the software layer into the runtime and hardware layers. Arc executes autonomous work, and the Super Intelligence Control Tower defines policies and restricts Arc’s behavior at runtime with NVIDIA OpenShell. AI Passport carries the business rules set by the Super Intelligence Control Tower — including risk tolerance and which external services the AI agent can access — so the NVIDIA enforcement layers can quickly approve or block AI agent actions. This approach helps enterprises confidently hand AI agents real work without losing control, with governance that travels with each AI agent wherever they run.
“ServiceNow and NVIDIA have been building toward the same vision for years. As AI agents take on more autonomous tasks, governance and security have to extend to everything they touch,” said Jon Sigler executive vice president and general manager, AI Platform at ServiceNow. “Arc is the first personal AI agent built for everyday work and governable by IT through the Super Intelligence Control Tower and NVIDIA OpenShell. NVIDIA software and hardware will also read AI Passport, so different technology layers know what an AI agent should be allowed to do. This helps give enterprises the trust needed to scale autonomous AI with confidence.”
“Enterprises are using AI agents to do real work, and that means policy has to be enforced everywhere an agent runs,” said Justin Boitano, vice president, Enterprise, at NVIDIA. “With NVIDIA OpenShell, ServiceNow can extend the governance it defines in the Super Intelligence Control Tower down into the runtime itself — the layer where agents actually take action.”
Arc: personal AI that works on behalf of employees, designed to be governed and secure at every layer
Arc is a personal, long-running AI agent that takes on everyday tasks across ServiceNow workflows, email, Teams, and the other common productivity tools. Arc can run on its own computer inside a secure NVIDIA OpenShell runtime or on an employee’s desktop. It builds memory and develops skills over time, helping to improve performance the longer it works with an employee.
What sets Arc apart is that it’s built for the enterprise. IT defines policies in the Super Intelligence Control Tower and NVIDIA OpenShell can be directed to enforce them at runtime. The result is a personal AI agent that can take action while operating within the guardrails of the business. Arc delivers powerful new capabilities:
- Local, lower-cost inference on NVIDIA RTX Spark: Arc is designed to handle everyday work with local models and turn to frontier models for more demanding tasks, helping to lower inference costs and keep more data on the local device.
- Autonomous task ownership: Assign Arc a goal, not just a sequence of instructions. It can plan how to accomplish it, execute the plan across enterprise systems, check its own work, and adjust along the way, continuing autonomously until the task is complete, whether that’s minutes or days.
- Proactive work identification: Arc proactively surfaces work and action items from an employee’s backlog, inbox, meetings, and conversations, such as assigned tasks, pending approvals, and follow-ups needed.
- AI agents build integrations: When an integration doesn’t exist, Arc can build one on the fly. Arc is designed to run anything it builds inside the same NVIDIA OpenShell boundary to prevent it from reaching further than an organization allows.
- Runtime policy enforcement: Because users can govern actions Arc takes by defining policies in the Super Intelligence Control Tower and enforcing them using NVIDIA OpenShell, they obtain granular control over which systems AI agents can access, APIs they can call, and data they can read.
AI Passport: AI agent validation and revocation authority
AI Passport is the attested record that the Super Intelligence Control Tower is designed to issue to each AI agent, MCP server, and other AI assets it governs. The Super Intelligence Control Tower will carry business context from the ServiceNow AI Platform, such as decision logic and vendor risk, and define which tools and actions each AI agent can use. Enforcement points can use it at the moment of orchestration and operation to confirm what an AI agent is trusted to do. ServiceNow issues AI Passport, and NVIDIA will read and enforce it at the runtime and in purpose-built silicon.
- Action-level control: The Super Intelligence Control Tower defines which tools and actions each AI agent, whether first- or third-party, can use. Enterprises can restrict a single action, such as writing to a production system, while the AI agent continues the rest of its work.
- Software layer enforcement: NVIDIA OpenShell reads the policies distributed by the Super Intelligence Control Tower and is designed to honor passport revocations, stopping AI agent execution at runtime in the sandboxed environment and keeping autonomous activity contained and auditable.
- Hardware layer enforcement: NVIDIA Sentry on BlueField reads the AI agent status through passport with hardware-backed trust.
- Transparent containment in NVIDIA AI Factories: When a passport is revoked in the Super Intelligence Control Tower, NVIDIA BlueField can enforce that revocation at machine speed before the AI agent can proceed. This creates enforcement in silicon designed to prevent overrides from compromised software.
- Sub-second propagation: Revocation enforcement can propagate across both layers within approximately one second.
- Precision containment: A revoked AI agent’s traffic stops while other workloads on the same host continue uninterrupted; restoration requires no infrastructure restarts.
Catch more CIO Insights: The Overlooked Security Risk Behind Enterprise AI Rollouts
[To share your insights with us, please write to psen@itechseries.com ]
