NVIDIA and ServiceNow Bring Autonomous AI Agents to Enterprise Workflows
Enterprise AI is evolving from generating content and reasoning to taking action. Companies now ask: How should AI act? Early agent systems show promise, but scaling them in enterprise environments requires context, control, and consistency across real workflows. At ServiceNow Knowledge 2026, NVIDIA and ServiceNow announced an expanded partnership to deliver specialized autonomous AI agents that are safe and easy to adopt. These agents combine NVIDIA accelerated computing, open models, domain-specific skills, and secure execution software with ServiceNow's Action Fabric and AI Control Tower for governance. A flagship innovation is Project Arc, a long-running, self-evolving desktop agent for knowledge workers. Below, we answer key questions about this collaboration and its implications for enterprises.
What is the new partnership between NVIDIA and ServiceNow?
NVIDIA and ServiceNow are expanding their collaboration across the full technology stack to deliver specialized autonomous AI agents for enterprises. Announced at ServiceNow Knowledge 2026 by NVIDIA CEO Jensen Huang and ServiceNow CEO Bill McDermott, the partnership focuses on agents that are safe to adopt and easy to deploy. These agents leverage NVIDIA’s accelerated computing, open models (such as those available through NVIDIA NIM), and domain-specific skills tailored for enterprise workflows. On the ServiceNow side, the agents integrate with ServiceNow Action Fabric for workflow context and the AI Control Tower for governance and auditability. The goal is to move beyond simple AI prompts to agents that operate with context, control, and consistency across real business processes, enabling enterprises to deploy AI at scale without compromising security or efficiency.
What is Project Arc and how does it work?
Project Arc is a long-running, self-evolving autonomous desktop agent introduced by ServiceNow, designed for knowledge workers like developers, IT teams, and administrators. Unlike standalone AI agents, Project Arc connects natively to the ServiceNow AI Platform through the Action Fabric, bringing governance, auditability, and workflow intelligence to every action. It can access local file systems, terminals, and applications installed on a machine to complete complex, multistep tasks that traditional automation cannot handle—such as running scripts, managing files, or configuring software across multiple steps. The agent operates with controls that enterprises need for scalable AI deployment, ensuring each action is contained and auditable. Project Arc uses NVIDIA OpenShell, an open source secure runtime, to run in sandboxed, policy-governed environments, allowing enterprises to define what an agent can see and which tools it can use.
How does Project Arc ensure security and governance?
Security and governance are embedded in Project Arc from the start. The agent uses NVIDIA OpenShell, an open source secure runtime that provides a sandboxed, policy-governed environment for developing and deploying autonomous agents. With OpenShell, enterprises can define what an agent can access, which tools it can use, and how each action is contained, preventing exposure of sensitive data or systems. Additionally, Project Arc integrates with ServiceNow AI Control Tower for central governance and auditability, ensuring every action is logged and compliant with enterprise policies. The Action Fabric provides workflow context, so the agent understands the broader business processes it operates within. This combination—secure runtime, governance tools, and workflow intelligence—allows enterprises to deploy autonomous agents with confidence, meeting the three requirements for long-running agents: open models, domain-specific skills, and security.
What are the three requirements for enterprise autonomous agents?
The collaboration between NVIDIA and ServiceNow is based on three essential requirements for long-running, autonomous agents in enterprises. First, open models and domain-specific skills that can be customized to an organization’s unique workflows and data. This allows agents to reason effectively within specific business contexts. Second, security that helps agents act without exposing sensitive data or systems—this is achieved through runtime containment and policy controls. Third, AI factories that deliver efficient tokenomics, meaning accelerated computing infrastructure that optimizes cost and performance for AI workloads. These three pillars ensure that autonomous agents are not only powerful but also safe and economically viable for enterprise-scale deployment. Both companies are building their solutions around these principles, with NVIDIA providing the compute and model foundation and ServiceNow adding the workflow and governance layers.
How does Project Arc differ from standalone AI agents?
Standalone AI agents typically operate in isolation, lacking integration with enterprise systems and governance controls. Project Arc differs by connecting natively to the ServiceNow AI Platform through the Action Fabric, which provides workflow context and ensures every action is governed by the AI Control Tower. This means Project Arc is not a black box; it has auditability, so IT teams can see exactly what the agent did and why. Additionally, Project Arc can access local file systems, terminals, and applications, enabling it to perform complex multistep tasks across different software environments—tasks that standalone agents cannot handle because they lack the necessary permissions and context. The governance-first design allows enterprises to deploy Project Arc at scale with confidence, as it respects policies and security boundaries. This makes Project Arc a practical solution for knowledge workers who need automation that fits into existing enterprise workflows.
Who are the key beneficiaries of Project Arc?
Project Arc is designed for knowledge workers, including developers, IT teams, and administrators. These professionals often spend significant time on repetitive, multistep tasks such as setting up development environments, managing files, running diagnostic scripts, or deploying software updates. Project Arc automates these complex tasks by accessing local file systems, terminals, and applications directly, completing work that traditional automation cannot. For developers, it can streamline code builds and testing. For IT teams, it can automate system configurations and troubleshooting. For administrators, it can handle data processing and report generation. By offloading these tasks to a governed autonomous agent, knowledge workers can focus on higher-value work. The integration with ServiceNow’s platform ensures that all actions are traceable and compliant, making Project Arc a safe and efficient productivity tool for enterprise environments.