Enterprise AI in Action: NVIDIA and ServiceNow Deliver Autonomous Desktop Agents
From Generation to Action
Enterprise artificial intelligence has progressed from generating text to reasoning through complex problems. Now, the next frontier is action—AI systems that don't just answer questions but complete tasks across real-world workflows. At ServiceNow Knowledge 2026, NVIDIA founder and CEO Jensen Huang joined ServiceNow CEO Bill McDermott to unveil the next phase of this journey: specialized autonomous agents designed for enterprise environments.
This partnership expands the companies' existing collaboration across the full AI stack. The goal is to deliver agents that are safe, easy to adopt, and powered by NVIDIA's accelerated computing, open models, and domain-specific skills—combined with ServiceNow's workflow context from its Action Fabric and governance from the AI Control Tower.
Project Arc: A Self-Evolving Desktop Agent
ServiceNow introduced Project Arc, a long-running, autonomous desktop agent built for knowledge workers—including 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 takes.
What sets Project Arc apart is its ability to access local file systems, terminals, and applications installed on a machine. It can handle complex, multistep tasks that traditional automation cannot, while still providing the controls enterprises need to deploy AI at scale. The agent is designed to evolve over time, learning from its own actions and improving its performance.
Governance and Security from the Start
Bringing this level of autonomy to enterprises requires control from the very beginning. Project Arc is built on three core requirements: open models and domain-specific skills that can be customized, and security that helps agents act without exposing sensitive data or systems—all running on AI factories that deliver efficient tokenomics.
ServiceNow's AI Control Tower provides a unified governance layer, ensuring every agent action is logged, auditable, and compliant with enterprise policies. The Action Fabric connects the agent to the broader workflow context, so actions are always grounded in business logic.
Open Models and Domain-Specific Skills
To be effective, enterprise AI systems must be adaptable. NVIDIA and ServiceNow are leveraging open models that can be fine-tuned with an organization's own data and domain knowledge. This allows the agents to understand industry-specific terminology, compliance requirements, and operational nuances.
The companies are also developing a library of pre-built agent skills—reusable capabilities for common tasks like ticket resolution, code deployment, and database queries. These skills can be combined and customized, reducing the time to deploy autonomous agents from months to days.
NVIDIA OpenShell: Secure Agent Execution
Project Arc uses NVIDIA OpenShell, an open-source secure runtime for developing and deploying autonomous agents in sandboxed, policy-governed environments. ServiceNow is both building on and contributing to OpenShell to advance a common foundation for enterprise-grade agent execution.
With OpenShell, enterprises can define exactly what an agent can see, which tools it can use, and how each action is contained. Jon Sigler, executive vice president and general manager of AI Platform at ServiceNow, explained: "Project Arc represents the next step in our ongoing collaboration with NVIDIA, bringing autonomous execution to the desktop. By combining OpenShell’s runtime layer with ServiceNow AI Control Tower, and powered by ServiceNow Action Fabric, we’re delivering the governance and security that enterprise AI requires."
What This Means for Enterprise AI
The partnership between NVIDIA and ServiceNow signals a shift from AI as a tool to AI as an autonomous worker. Early agent systems have shown promise, but the enterprise requires trust, control, and scalability. Project Arc, with its sandboxed execution, open models, and deep integration into existing workflows, offers a blueprint for how organizations can safely deploy long-running agents.
As the technology matures, we can expect these agents to handle increasingly complex tasks—managing entire IT operations, orchestrating supply chains, and even assisting in software development. The key enablers will remain: open architectures, robust governance, and hardware optimized for efficient token processing.
For now, Project Arc demonstrates that enterprise AI is ready to act—not just think and talk. With NVIDIA's acceleration and ServiceNow's workflow intelligence, the future of autonomous business processes has arrived.