Glean to expand enterprise choices for generative AI with NVIDIA

With the AI landscape changing by the day, enterprises need the flexibility to choose the right language models for their own applications. However, every organization is different, with its own unique circumstances and constraints that call for specific large language models (LLMs). That is why Glean has been committed to working with whatever LLM our customers prefer. 

As the next step along that journey, Glean is excited to enable customers to use NVIDIA NIM to power Glean Assistant’s generative AI capabilities. Customers will be able to use NIM to easily adopt a wide range of language models. 

Deploying AI quickly with NIM

NIM microservices are included in the NVIDIA AI Enterprise software platform and speed up generative AI deployment in enterprises. Supporting a wide range of AI models, including NVIDIA AI foundation and custom models, NIM microservices provide seamless, scalable AI inferencing, on-premises or in the cloud, leveraging industry-standard APIs.

With NIM, our customers can choose from a wide variety of language models to self-host on their own infrastructure. Customers can use the NVIDIA API Catalog to experiment with the latest GPU-accelerated LLMs the day they are released, and easily integrate them with Glean Assistant. NIM makes it easy to deliver optimized  LLM inference at scale, no matter where the model is hosted.

“Volumes of enterprise data is stored across communications and collaboration platforms, and accessing this data can unlock the value of generative AI for businesses in every industry,” said Justin Boitano, vice president of Enterprise Products at NVIDIA. “Pairing Glean’s retrieval system with NVIDIA NIM inference microservices gives customers a flexible RAG application development platform that protects enterprise IP and prompts by keeping them on their network.”

An enterprise-ready solution

With the combination of Glean’s retrieval system and NIM in a single RAG pipeline, enterprise customers can build their own assistants that can process a wide range of documents from tools such as Confluence, SharePoint, Google Drive, and OneDrive, along with employee collaboration tools like Slack and Microsoft Teams. Glean’s retrieval system retrieves relevant data from multiple sources while honoring the customer’s existing document access control policies. A self-hosted NIM microservice can then take the retrieved data to generate and serve high-quality responses while running optimized LLM inference in the backend. This combined solution can handle thousands of simultaneous queries – all at enterprise scale.

We are excited to collaborate with NVIDIA on progressing the capabilities of generative AI for enterprise environments. Join me on a panel at NVIDIA GTC on Tuesday, March 19, at 10 a.m. to continue this discussion, and visit us on the show floor at booth G108 for a demo to learn more about how Glean and NVIDIA NIM can help your organization empower employees to do their best work with generative AI.