7 Breakthroughs in Enterprise AI: How NVIDIA and ServiceNow Are Redefining Autonomous Agents

Enterprise AI has moved from generating text and reasoning through problems to taking decisive action. But for businesses, the challenge isn't just building smarter agents—it's deploying them safely, securely, and at scale within complex workflows. At ServiceNow Knowledge 2026, NVIDIA founder Jensen Huang and ServiceNow CEO Bill McDermott unveiled a deepened partnership that promises to deliver exactly that. Their collaboration spans the full AI stack, from accelerated computing and open models to governance and secure execution. At the heart of their announcement is Project Arc, a groundbreaking autonomous desktop agent designed for knowledge workers. But Project Arc is just one piece of a larger puzzle. Here are seven key breakthroughs that enterprises need to know about this next phase of AI autonomy.

1. A Full-Stack Partnership for Enterprise-Grade AI

NVIDIA and ServiceNow are expanding their collaboration beyond isolated tools. Their new approach integrates NVIDIA's accelerated computing, open models, and secure runtime software with ServiceNow's Action Fabric and AI Control Tower. This full-stack synergy ensures that autonomous agents can operate with enterprise-level context, control, and consistency. Instead of fragmented solutions, businesses get a cohesive system where every action an agent takes is governed and auditable. The partnership leverages NVIDIA's expertise in high-performance AI infrastructure and ServiceNow's deep understanding of enterprise workflows, creating agents that are both powerful and safe. By combining domain-specific skills with secure agent execution, this collaboration sets a new standard for how AI can act within real business environments.

2. Project Arc: The Autonomous Desktop Agent That Works While You Sleep

Project Arc is a long-running, self-evolving desktop agent built for knowledge workers like developers, IT teams, and administrators. Unlike earlier AI tools that require constant prompting, Project Arc connects natively to the ServiceNow AI Platform via Action Fabric, bringing governance and workflow intelligence to every action. It can access local file systems, terminals, and installed applications to complete complex, multistep tasks that traditional automation cannot handle. Imagine an agent that manages your software updates, troubleshoots system errors, or processes data migrations without human supervision—all while maintaining enterprise-grade controls. Project Arc represents a leap from simple task automation to true autonomous execution, but with the guardrails enterprises need to deploy AI at scale.

3. Three Pillars for Autonomous Agents: Open Models, Skills, and Security

To make long-running agents viable, the companies identified three non-negotiable requirements. First, open models that can be customized and fine-tuned for specific industries or tasks—no black boxes. Second, domain-specific skills that allow agents to perform specialized functions, such as IT operations or customer service, out of the box. Third, security that ensures agents act without exposing sensitive data or systems. All of this runs on AI factories that deliver efficient tokenomics, meaning enterprises can scale AI actions without spiraling costs. These pillars form the foundation for any company looking to deploy autonomous agents reliably in production environments, not just in experimental labs.

4. OpenShell: An Open Source Runtime for Secure Agent Execution

Security is paramount when agents can access local machines and sensitive data. That's where NVIDIA OpenShell comes in—an open source secure runtime for developing and deploying autonomous agents in sandboxed, policy-governed environments. ServiceNow is building on and contributing to OpenShell to advance a common foundation for enterprise-grade agent execution. With OpenShell, organizations can define exactly what an agent can see, which tools it can use, and how each action is contained. This prevents agents from roaming freely across systems and introduces policy enforcement at the execution level. By open-sourcing this technology, NVIDIA and ServiceNow are inviting the broader developer community to collaborate on safe agent infrastructure, accelerating adoption while lowering risks.

5. AI Control Tower: Centralized Governance for Decentralized Agents

As enterprises deploy dozens or even thousands of agents, managing them becomes a nightmare without a central control system. ServiceNow's AI Control Tower provides exactly that—a unified dashboard to monitor, audit, and govern all autonomous agents in an organization. Combined with Action Fabric, which connects agent actions to real workflow contexts, the Control Tower ensures that every agent decision is transparent and aligned with business rules. This is critical for compliance, risk management, and accountability. Rather than treating agents as isolated tools, the Control Tower makes them part of an orchestrated ecosystem, where human oversight remains integrated. It empowers IT leaders to trust AI actions because they can see exactly what happened, why, and who (or what) initiated it.

6. Action Fabric: Bridging Agent Actions with Real Workflows

An agent that acts in isolation is useless in an enterprise. ServiceNow's Action Fabric is the middleware that connects autonomous agents to actual business workflows, systems, and data sources. When Project Arc performs a task—say, updating a user's permissions or deploying a software patch—Action Fabric ensures that action is logged, integrated with the appropriate IT service management process, and triggerable for follow-ups. This creates a closed-loop system where agents contribute to, rather than disrupt, established processes. Action Fabric also enables agents to pull context from the enterprise, making their decisions smarter and more relevant. It's the difference between a bot that does a single task and an agent that becomes a seamless part of your operational fabric.

7. The Future of Work: From Prompts to Proactive Agents

The collaboration between NVIDIA and ServiceNow signals a shift from reactive AI (chatbots that answer questions) to proactive AI (agents that autonomously solve problems). Early agent systems showed promise, but they lacked the governance, security, and workflow integration needed for enterprise adoption. Now, with Project Arc, OpenShell, AI Control Tower, and Action Fabric, companies can deploy agents that run for long periods, learn from their environments, and act without constant human hand-holding. This doesn't eliminate jobs—it augments them, freeing knowledge workers from repetitive, multistep tasks so they can focus on higher-value work. The next decade of enterprise AI is about action, and NVIDIA and ServiceNow are building the infrastructure to make it happen safely and at scale.

In conclusion, the partnership between NVIDIA and ServiceNow marks a pivotal moment in enterprise AI. By combining open models, secure execution, centralized governance, and deep workflow integration, they are laying the groundwork for a new era of autonomous agents. Project Arc is not just a product—it's a blueprint for how AI can act responsibly within complex business ecosystems. For enterprises considering autonomous AI, these seven breakthroughs offer a clear roadmap: start with secure foundations, leverage open technologies, and always maintain human control. The future of work is proactive, and these tools ensure that future is both powerful and safe.