NVIDIA Nemotron and Agentic AI

We shared some exciting news in August at SIGGRAPH 2025: Automation Anywhere is teaming up with NVIDIA to bring its new Nemotron reasoning models to our Agentic Process Automation (APA) System and agentic solutions, adding even more intelligence and reliability to our already powerful platform.

If you’ve been following the rise of generative AI (Gen AI) and robotic process automation (RPA), you’ve probably noticed a shift happening. Businesses no longer want automation that just takes repetitive tasks off their plate. They’re asking for something bigger: automation that can reason, adapt, and execute complex, multi-step processes on its own. That’s where Agentic AI comes in.

Why Nemotron?

When we put Nemotron models through their paces inside our Enterprise AI Agent Studio and Automator AI, the results were striking:

Within Enterprise AI Agent Studio, Nemotron matched a leading closed-source reasoning model with an 89.4% success rate, but stood out for efficiency, performance, and enterprise-grade security.

In Automator AI, Nemotron outperformed its closed-source rival, delivering conciseness 55% of the time while also showing improvements in correctness and relevance – all with less than one-quarter of the parameter count.

This makes Nemotron a powerful open-source alternative that enterprises can trust – not just for performance, but also for scalability, cost-effectiveness, and security.

Raising the bar for agentic automation

AI agents empower our customers to transform highly intricate workflows, such as financial reconciliations and supply chain optimizations, into streamlined automations. For these types of use cases, accuracy and resilience aren’t optional. Nemotron’s advanced reasoning capabilities will allow our APA System to:

Handle multi-step logic with precision
Deliver results that are not only correct, but also relevant and concise
Scale across enterprise environments with the speed and reliability of NVIDIA’s accelerated computing platform

As Automation Anywhere CEO Mihir Shukla puts it: "Our customers rely on agentic solutions to automate complex processes, and that demands the highest-performing LLMs for accurate, multi-step reasoning. Nemotron models consistently impressed us with their ability to handle the intricate logic our automation systems require."

Open source, enterprise ready

Another reason Nemotron excites us? It’s open and commercially viable. That means customers can build proprietary solutions confidently without worrying about restrictive licenses. Combined with NVIDIA optimization, we’re making sure enterprises get the freedom, speed, and resilience they need to unlock the next generation of automation.

What’s next

This collaboration is about more than integrating a new model. It’s about empowering every organization to build an autonomous enterprise – one where AI agents don’t just assist, but take ownership of end-to-end processes with confidence.

We’re just getting started. Expect to see Nemotron-powered capabilities rolled out across our platform, fueling the future of work by unleashing human potential through automation.