Nvidia has announced a partnership with ServiceNow to develop domain-specific generative AI models for enterprises.
ServiceNow is a SaaS provider focused on IT service management tools. The company develops custom large-scale language models trained on data exclusively for his ServiceNow platform, a cloud-based workflow automation platform. ServiceNow uses Nvidia’s NeMo foundation model as a starting point, starting with models for the IT domain and training fine-tuned models on Nvidia GPUs.
Rama Akkiraju, VP of AI/ML for IT at Nvidia, explained at a press conference how generative AI is transforming business. Text generation and summarization, real-time language translation, coding assistance, prompt-based image generation, and drug discovery are promising use cases for generative AI in the enterprise. Akkiraju noted that business leaders are actively building proofs of concept to discover how the capabilities of these technologies apply to their own use cases. According to a recent Accenture report, 98% of his global executives agree that his AI foundational models will play a key role in their organizations’ strategies in the next three to five years.
Because of their tendency to hallucinate inaccurate information, underlying models like GPT-4 trained using public domain data are not necessarily suitable for enterprise use cases, especially those requiring a high level of accuracy. We cannot.
“We know that generative AI models are good at learning from public domain data sources to perform multiple tasks such as text summarization and translation, recording assistance and image generation. The model doesn’t know specifically about the data in the company, it doesn’t look at it,” Akkiraju said.
Domain-specific enterprise data can be fine-tuned to these basic models and customized for specific industries. These customized models can be deployed and hosted, and when used in production environments, companies can continuously apply guardrails to ensure safe and effective use.
“To bring generative AI to the enterprise, these models, or foundational models, can be customized to teach the enterprise language and enterprise-specific skills to provide more in-domain responses for enterprises to have appropriate guardrails. We need to be able to do that,” said Akkiraju.
Akkiraju gives the example of a new employee trying to connect to the company’s VPN. “If you ask a generative AI model how to connect to a VPN as a new employee, it shouldn’t answer based on someone else’s public domain knowledge of the company. It should give you an informed answer and be very specific to the policies specified for connecting to the VPN.”
This is one of several use cases for generative AI capabilities in the IT domain, including help desks, IT ticket summaries, and enterprise search capabilities that help IT professionals find knowledge in their domain. AI can also help detect, predict, and mitigate IT service outages. Automated resolution is also an option, as chatbots and question-and-answer systems allow users to solve problems on their own without waiting in line.
NeMo offers rapid tuning, supervised tweaking, and knowledge retrieval tools. (Source: Nvidia)
As part of this partnership, ServiceNow will use Nvidia data to streamline Nvidia’s IT operations and Nvidia’s capabilities running on hybrid cloud infrastructure, specifically the Nvidia DGX Cloud and its on-premises DGX SuperPOD AI supercomputers. Customize your NeMo foundation model. ServiceNow uses Nvidia’s AI Foundations cloud service along with the Nvidia AI Enterprise software platform, which includes the Nvidia NeMo framework. As part of NeMo Guardrails software, NeMo offers rapid tuning, supervised tweaking, knowledge search tools, as well as safety and security features.
One of the first use cases for this collaboration is IT ticket summaries. Akkiraju estimates that the current helpdesk summary task takes about seven to eight minutes per agent. Generative AI can save valuable time for busy service agents by automating these summaries and adding them to the enterprise’s knowledge base.
The companies also expect to use the technology to improve the employee experience by identifying growth opportunities such as customized learning and development recommendations based on natural language queries.
“IT is the nervous system of every modern enterprise in every industry,” Nvidia founder and CEO Jensen Huang said in a release. “Our collaboration to build hyper-specialized generative AI for the enterprise will improve the capabilities and productivity of his IT professionals around the world who use the ServiceNow platform.”
“As the adoption of generative AI continues to accelerate, organizations are looking to trusted vendors with proven and secure AI capabilities to improve productivity, gain a competitive edge, and keep their data and IP safe. ,” said CJ Desai, President and Chief Operating Officer. of ServiceNow. “Nvidia and ServiceNow are working together to drive new levels of automation to boost productivity and maximize business impact.”
Related
