ServiceNow and Nvidia have unveiled Project Arc, a new enterprise desktop agent designed to bring autonomous AI capabilities into corporate environments without sacrificing the security and governance controls that IT leaders require. Announced at ServiceNow's Knowledge 2026 conference, Project Arc represents a significant shift in how enterprises might deploy autonomous agents on employee machines, addressing a critical gap between the promise of AI agents like OpenClaw and the practical realities of corporate IT security.

What Makes Project Arc Different From Consumer AI Agents?

The core innovation behind Project Arc lies in its architecture. The agent runs inside Nvidia's OpenShell, a sandbox environment where "everything is denied by default," meaning IT administrators must explicitly grant permissions for any action the agent takes. This contrasts sharply with consumer-facing AI agents that operate with broader access to user systems. ServiceNow's AI Control Tower oversees and manages the agent's activities while grounding decisions in the company's Configuration Management Database (CMDB), a centralized repository of IT assets and configurations.

Employees can think of Project Arc as AI-powered IT support for handling routine, repetitive tasks. The autonomous agent is capable of thinking through problems, writing code, executing tasks, and adapting when things don't go as planned. ServiceNow claims it can handle multi-step work across enterprise tools and systems without requiring prebuilt workflows, a significant advantage over traditional robotic process automation (RPA) tools that often require extensive manual configuration.

"Long-running, autonomous agents are rapidly changing the game for enterprise AI, and delivering them securely at scale requires governance that spans models, software, and AI infrastructure," said Kari Briski, Nvidia's vice president of generative AI for enterprise.

Kari Briski, Vice President of Generative AI for Enterprise, Nvidia

How Does Project Arc Handle Different Types of Tasks?

Project Arc works in tandem with ServiceNow's existing AI specialists, which are prebuilt agents designed for specific business domains like IT and human resources. The division of labor is strategic: cloud-based requests, such as a Zoom access request or an entitlement issue, are handled by these domain-specific agents by default. However, when a task requires direct access to an employee's machine, such as restarting a browser or creating an email cache, Project Arc steps in to handle the work on the desktop itself.

Joe Davis, ServiceNow's executive vice president of AI engineering and delivery, acknowledged that IT-specific desktop agents could be next on the company's roadmap. He noted that many enterprise IT problems involve changing system settings and restarting applications, tasks that a self-evolving desktop agent could learn to handle and then share solutions across users.

"What we're trying to help customers with ultimately is automation," explained Joe Davis, ServiceNow's executive vice president of AI engineering and delivery.

Joe Davis, Executive Vice President of AI Engineering and Delivery, ServiceNow

Steps to Deploy Project Arc in Your Organization

• Assess Current Infrastructure: Evaluate whether your organization's hardware can support the agent running locally or if most tasks will need to run in the cloud; older machines will lean heavily on cloud processing.
• Define Permission Policies: Work with IT administrators to establish what actions the agent is allowed to take on employee machines, leveraging OpenShell's "deny by default" model to grant only necessary permissions.
• Integrate With Existing Systems: Connect Project Arc to your Configuration Management Database and existing ServiceNow AI Control Tower to ensure governance and visibility across all agent activities.
• Plan for Regulated Industries: If your organization operates in financial services or pharmaceuticals, allocate additional time for compliance review and work with ServiceNow specialists to navigate regulatory requirements.

Why Is Compliance the Real Challenge?

While the technology behind Project Arc is sophisticated, the harder challenge is compliance. Davis acknowledged that heavily regulated industries, such as financial services and pharmaceuticals, face a longer road to adoption. ServiceNow sees AI Control Tower as its competitive advantage in this space, providing a cross-platform governance layer that abstracts away the underlying operating system. Nvidia's partnership helps solve the trickier problem of edge connectivity to local devices, ensuring that sensitive work can be handled on-device when necessary.

The emergence of autonomous agents in the so-called "Claw" family, including Nvidia's own NemoClaw, has reset expectations for what these agents can do and how quickly they can do it. However, widespread enterprise adoption of autonomous desktop agents is still in its early stages, and the window for establishing governance standards may be closing faster than expected.

Project Arc is currently available as an early preview, with no public launch timeline announced. The move signals that ServiceNow and Nvidia believe the enterprise market is ready for autonomous desktop agents, provided they come with the security and governance guardrails that IT leaders actually require. As more companies experiment with multiple AI agent platforms, Project Arc positions itself as the option built specifically for corporate environments where compliance and control are non-negotiable.