ServiceNow is accelerating enterprise AI with a brand new reasoning mannequin in-built partnership with NVIDIA — enabling AI brokers that reply in actual time, deal with complicated workflows and scale features like IT, HR and customer support groups worldwide.
Unveiled right this moment at ServiceNow’s Data 2025 — the place NVIDIA CEO and founder Jensen Huang joined ServiceNow chairman and CEO Invoice McDermott throughout his keynote handle — Apriel Nemotron 15B is compact, cost-efficient and tuned for motion. It’s designed to drive the subsequent step ahead in enterprise giant language fashions (LLMs).
Apriel Nemotron 15B was developed with NVIDIA NeMo, the open NVIDIA Llama Nemotron Submit-Coaching Dataset and ServiceNow domain-specific knowledge, and was skilled on NVIDIA DGX Cloud working on Amazon Internet Companies (AWS).
The information follows the April launch of the NVIDIA Llama Nemotron Extremely mannequin, which harnesses the NVIDIA open dataset that ServiceNow used to construct its Apriel Nemotron 15B mannequin. Extremely is among the many strongest open-source fashions at reasoning, together with scientific reasoning, coding, superior math and different agentic AI duties.
Smaller Mannequin, Larger Influence
Apriel Nemotron 15B is engineered for reasoning — drawing inferences, weighing targets and navigating guidelines in actual time. It’s smaller than a few of the newest general-purpose LLMs that may run to greater than a trillion parameters, which implies it delivers quicker responses and decrease inference prices, whereas nonetheless packing enterprise-grade intelligence.
The mannequin’s post-training passed off on NVIDIA DGX Cloud hosted on AWS, tapping high-performance infrastructure to speed up growth. The end result? An AI mannequin that’s optimized not only for accuracy, however for pace, effectivity and scalability — key components for powering AI brokers that may assist hundreds of concurrent enterprise workflows.
A Closed Loop for Steady Studying
Past the mannequin itself, ServiceNow and NVIDIA are introducing a brand new knowledge flywheel structure — integrating ServiceNow’s Workflow Information Cloth with NVIDIA NeMo microservices, together with NeMo Customizer and NeMo Evaluator.
This setup allows a closed-loop course of that refines and improves AI efficiency by utilizing workflow knowledge to personalize responses and enhance accuracy over time. Guardrails guarantee clients are in command of how their knowledge is utilized in a safe and compliant method.
From Complexity to Readability
In a keynote demo, ServiceNow is displaying how these agentic fashions have been deployed in actual enterprise eventualities, together with with AstraZeneca, the place AI brokers will assist staff resolve points and make selections with higher pace and precision — giving 90,000 hours again to staff.
“The Apriel Nemotron 15B mannequin — developed by two of essentially the most superior enterprise AI corporations — options purpose-built reasoning to energy the subsequent technology of clever AI brokers,” mentioned Jon Sigler, govt vp of Platform and AI at ServiceNow. “This achieves what generic fashions can’t, combining real-time enterprise knowledge, workflow context and superior reasoning to assist AI brokers drive actual productiveness.”
“Along with ServiceNow, we’ve constructed an environment friendly, enterprise-ready mannequin to gas a brand new class of clever AI brokers that may purpose to spice up workforce productiveness,” added Kari Briski, vp of generative AI software program at NVIDIA. “Through the use of the NVIDIA Llama Nemotron Submit-Coaching Dataset and ServiceNow domain-specific knowledge, Apriel Nemotron 15B delivers superior reasoning capabilities in a smaller measurement, making it quicker, extra correct and cost-effective to run.”
Scaling the AI Agent Period
The collaboration marks a shift in enterprise AI technique. Enterprises are shifting from static fashions to clever techniques that evolve. It additionally marks one other milestone within the partnership between ServiceNow and NVIDIA, pushing agentic AI ahead throughout industries.
For companies, this implies quicker decision instances, higher productiveness and extra responsive digital experiences. For expertise leaders, it’s a mannequin that matches right this moment’s efficiency and price necessities — and might scale as wants develop.
Availability
ServiceNow AI Brokers, powered by Apriel Nemotron 15B, are anticipated to roll out following Data 2025. The mannequin will assist ServiceNow’s Now LLM providers and can turn out to be a key engine behind the corporate’s agentic AI choices.
Be taught extra concerning the launch and the way NVIDIA and ServiceNow are shaping the way forward for enterprise AI at Data 2025.
