ServiceNow taps Nvidia to add AI workflows to Now Platform

ServiceNow continues its steady focus on AI for the Now Platform with a partnership with Nvidia announced Tuesday. The two companies will jointly develop generative AI capabilities that allow IT organizations to create more intelligent and customized workflows spanning an enterprise.

Using Nvidia's software, services and infrastructure as the foundation, ServiceNow will deliver customized large language models trained on data to specifically work with ServiceNow's Now Platform. The workflows created address the needs of not just IT professionals, but also customer service teams, employees and developers.

As part of the partnership, ServiceNow will also help Nvidia optimize the performance of its IT operations using the upcoming AI tools. ServiceNow will use Nvidia's data to customize Nvidia's NeMo framework to build foundation models that will run on a hybrid cloud infrastructure anchored by Nvidia's DGX SuperPod AI supercomputer.

ServiceNow has positioned itself well in the market the past few years to take advantage of both the technology assets Nvidia brings to the partnership as well as its own AI development and AI acquisitions, according to one analyst.

Dan Newman

"This [partnership] is all about building smaller foundational models that have specific, narrower functions," said Dan Newman, chief analyst of Futurum Research and CEO of The Futurum Group. "They want models that take less training, use less power and enable specific business processes, and to do it all out of the Now Platform using pre-built foundational [AI]models."

In a press briefing, each company said it is researching a range of different AI use cases all intended to boost productivity across an enterprise. This will include creating intelligent virtual assistants and agents that can more effectively respond to questions and a variety of support requests. This will be accomplished using purpose-built AI chatbots that use large language models and target well-defined IT tasks, spokespeople from each company said.

While generative AI models are effective at gathering data from public domain data sources and are capable of carrying out multiple tasks simultaneously, they don't know how to deal with data residing in an enterprise because have never seen it, said Rama Akkiraju, vice president, AI for IT at Nvidia.

"If I ask a generative AI model about how to connect to a VPN in a company where he or she is a new employee, it wouldn't be able to answer that question accurately," Akkiraju said in a press briefing. "If we are to bring AI to enterprises, we must customize these models. It is the foundational models that can teach them the language of the enterprise and the enterprise-specific skills so they can provide intimate responses with the proper guardrails."

Akkiraju said she sees a significant opportunity in applying these customized models to help service desk agents resolve IT incidents faster. Specifically, routing tickets to the appropriate agent for answers to sidestep delays, a common problem that causes a backlog of issues to quickly build. But these models can also be used to build a database of information that can predict the severity of such cases and automatically give users the answers they need to address the issue themselves.

"We want to identify problems in an IT environment that happened earlier so we can leverage solutions and apply them quickly," she said. "Not just for auto ticket routing and auto resolutions, but [also for] employee questions about benefits, requests to replace a PC that has been stolen or specific problems relating to password resets."

Key to helping users customize their models is Nvidia's NeMo. NeMo contains fine-tuning and knowledge retrieval tools to assist developers in building, customizing and deploying language models that can be applied to specific enterprise use cases. NeMo also features the company's Guardrails software, which allows developers to add safety and security features for AI chatbots.

As Editor At Large with TechTarget's News Group, Ed Scannell is responsible for writing and reporting breaking news, news analysis and features focused on technology issues and trends affecting corporate IT professionals.