NVIDIA and ServiceNow Extend AI Governance From Desktops to Data Centres

Why It Matters

By embedding robust governance across the full AI stack, enterprises can safely scale autonomous agents, reducing risk and accelerating productivity gains.

Key Takeaways

• Project Arc enables autonomous desktop AI tasks under strict governance.
• NVIDIA OpenShell provides sandboxing and policy control for AI agents.
• AI Control Tower now monitors workloads in NVIDIA’s Enterprise AI Factory.
• NOWAI‑Bench offers open benchmarks for multi‑step enterprise AI workflows.
• Early preview available; integration with AI Factory is generally available.

Pulse Analysis

Enterprises have been quick to adopt generative AI, but the speed at which autonomous agents can act across systems has raised governance concerns. ServiceNow’s partnership with NVIDIA tackles this gap by extending policy‑driven oversight from the user’s desktop to the underlying data‑center infrastructure. By leveraging NVIDIA’s Enterprise AI Factory—a validated design for AI workloads—the collaboration creates a unified control plane that can audit model usage, enforce compliance, and surface cost metrics in real time. This end‑to‑end visibility is a prerequisite for scaling AI responsibly in regulated sectors such as finance and healthcare.

At the heart of the announcement is Project Arc, a desktop‑resident AI agent built on ServiceNow’s Action Fabric. Unlike traditional bots that follow static scripts, Arc can generate code, invoke APIs and adjust its behavior based on contextual cues, all while operating inside NVIDIA OpenShell’s sandbox. The AI Control Tower continuously logs file accesses, command executions and API calls, providing an auditable trail that IT and security teams can query. This combination of autonomous capability and granular monitoring promises to reduce manual effort in IT service management, HR onboarding, and other repetitive enterprise processes.

The release of NOWAI‑Bench, an open‑source benchmarking suite, adds a layer of industry accountability. By standardizing performance and safety tests for multi‑step workflows, the suite enables vendors and buyers to compare agents on objective criteria rather than marketing hype. As more organizations look to embed AI agents in critical operations, the ability to benchmark, govern and cost‑track these models will become a competitive differentiator. Analysts expect the ServiceNow‑NVIDIA framework to accelerate enterprise AI spend, driving a wave of vendor integrations that prioritize security and compliance alongside innovation.