ServiceNow doubles down with NVIDIA
ServiceNow expanded its partnership with NVIDIA at GTC 2026 to push AI deeper into enterprise workflows, framing the effort around an “Autonomous Workforce” that can automate routine and complex tasks. The move signals that buyers will value vendors who not only embed AI but also offer governance and compliance for scaled deployments, a key selling point for tech services firms (insidermonkey.com).
ServiceNow used its March 16, 2026 presentation at NVIDIA’s GTC conference to show concrete hookups between its new Autonomous Workforce (a set of AI “workers” that can take actions inside business systems) and NVIDIA’s developer and infrastructure tools. (newsroom.servicenow.com) The company said those hookups aren’t just about smarter answers: they add built‑in controls such as audit logs (records of what an AI did), escalation paths (automated ways to hand a task to a human), and centralized oversight so automated workers run inside existing corporate rules. (servicenow.com) On the technical side ServiceNow described integrating its “AI Control Tower” — a central console that monitors, configures and enforces policies for AI agents — with NVIDIA’s Enterprise AI Factory, which is a reference architecture and set of tools for hosting, scaling and certifying models and data pipelines; ServiceNow also showed integrations with NVIDIA’s Agent Toolkit and the AI‑Q blueprint, developer kits and prescriptive designs for building agentic systems. (newsroom.servicenow.com) ServiceNow said deployments will use a mix of open and closed models, calling out its Apriel family (open models that ServiceNow post‑trains on enterprise data to add reasoning and multimodal input) built on NVIDIA’s Nemotron architecture (an open‑model framework with published training recipes), and that those models can run on NVIDIA Blackwell data‑center GPUs (NVIDIA’s current server AI chips). (businesswire.com) (developer.nvidia.com) ServiceNow demonstrated a specific use case: a Level‑1 Service Desk AI Specialist that can read incoming support tickets, search knowledge and historical incidents to find root causes, and then execute remediation steps inside enterprise systems — completing multi‑step workflows rather than only suggesting one action, while the platform enforces governance and logs every step. (marketscreener.com) ServiceNow and NVIDIA previously said Apriel 2.0 (the next Apriel Nemotron model) was engineered to deliver reasoning and multimodal capabilities in a smaller, more efficient footprint and was slated for production rollout in the first quarter of 2026, with model artifacts to be made available for enterprise use. (businesswire.com)
