ServiceNow Nvidia Launch Governed AI Agent Services

The ServiceNow Nvidia partnership has expanded with a new set of agentic AI offerings designed to bring governance and security to every layer of AI agent execution, the companies announced on October 21. The centerpiece is Arc, a personal long-running AI agent built for everyday work that remains governable by enterprise IT.

According to ServiceNow, the ServiceNow Nvidia expansion addresses a problem that has shadowed the agentic AI boom: agents that plan work, use applications, write code, and take actions with limited human involvement also raise the question of whether they stay within their authority. The new offerings aim to let agents work independently while keeping the enterprise in control.

Arc: a personal agent under IT governance

Arc is described by ServiceNow as a personal long-running AI agent that can handle everyday work on behalf of employees. What sets it apart from consumer chatbots is the governance layer around it: Arc is governed by the ServiceNow Super Intelligence Control Tower and enforced with Nvidia OpenShell, so IT teams can set boundaries on what the agent may access and execute.

ServiceNow says the combination lets the agent work independently while the enterprise stays in control. Jon Sigler, executive vice president and general manager of the AI Platform at ServiceNow, called Arc the first personal AI agent built for everyday work and governable by IT through the Control Tower and Nvidia OpenShell.

The pitch lands at a moment when enterprises are moving from pilot projects to production deployments of agents. According to ServiceNow, the challenge is no longer whether agents can do useful work, but whether organizations can trust them at scale across systems that hold sensitive data and execute real transactions.

AI Passport, OpenShell, and BlueField enforcement

Beyond Arc, the ServiceNow Nvidia announcement extends enforcement down the technology stack. Nvidia OpenShell and BlueField data processing unit enforcement layers will be able to read ServiceNow's recently announced AI Passport tool to validate what AI agents, including Arc, are permitted to access or execute.

The idea is that different technology layers should share the same understanding of an agent's permissions. As Sigler put it, Nvidia software and hardware will also read AI Passport, so different layers know what an AI agent should be allowed to do, giving enterprises the trust needed to scale autonomous AI with confidence.

That layered approach mirrors a broader shift in the agent ecosystem. Recent weeks have seen a wave of governance-focused launches, from governed API access for agents to new agent identity and access controls, as vendors race to answer the question enterprises keep asking: who watches the agent?

Why governance is now the selling point

The ServiceNow Nvidia announcement arrives amid rising scrutiny of autonomous agents. Governments and regulators have begun treating agent incidents as a distinct category, with the UK committing funding for agentic AI incident response and US security agencies publishing guidance on detecting and shutting down misbehaving autonomous systems.

Consumer trust is also uneven. Research by PYMNTS Intelligence has shown that while 56 percent of consumers will permit an agent to search and compare products, just 35 percent will allow an agent to access saved payment credentials. Most consumers say they would let agents make purchases if the right guardrails are in place, according to the report, which suggests governance is not just a compliance story but a precondition for adoption.

ServiceNow and Nvidia are betting that the vendors who win enterprise agent deployments will be the ones who can prove control, not just capability. As Sigler said, the two companies 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.

What to watch next

The key test for the ServiceNow Nvidia offering, Arc and the Control Tower, will be real deployments: whether IT teams find the governance controls practical rather than performative, and whether agents governed this tightly can still move fast enough to be useful. Enterprises evaluating the offering will also want to see how AI Passport interoperates with the identity and access tools they already run.

For the wider agent ecosystem, the announcement is another sign that the competitive axis is shifting. The first wave of agent products competed on what agents could do. The next wave is competing on whether anyone can prove what they did, and stop them when they go too far. The ServiceNow Nvidia team is positioning itself squarely on that second front.

Sources: PYMNTS reported the ServiceNow Nvidia announcement on October 21, ServiceNow and Nvidia, and related agent governance coverage on genznewz and October declared AI Agent Governance Month provide further context on the governance trend.