Nvidia Unveils NIM Agent Blueprints: Accelerate Development of Enterprise AI Applications for Developers

Today, Nvidia unveiled NIM Agent Blueprints, a comprehensive catalog of AI workflows and software resources tailored to expedite the development and deployment of generative AI-powered agents and applications.

Free Download Available

Designed for practical use, these blueprints support key applications such as customer service avatars, drug discovery, and document data extraction. Developers can leverage their datasets to create and deploy customized agents that align with their business functions. Nvidia also intends to expand this resource library with more applications in the near future.

As enterprises across various sectors embrace the potential of generative AI for enhanced productivity and cost savings, Nvidia is collaborating with technology solution providers and global systems integrators to facilitate efficient access and deployment of the blueprints.

Significant Value Potential

According to McKinsey, the implementation of generative AI could generate an annual value between $2.6 trillion and $4.4 trillion across over 60 use cases.

What Do NIM Agent Blueprints Offer?

While many organizations have utilized generative AI for tasks like content generation and summarization, there is a growing demand for advanced applications powered by customized AI agents using proprietary data. This next wave of generative AI holds significant growth potential; however, navigating the multi-step process of building and deploying these agents remains a challenge, causing delays and increased costs for many organizations.

Nvidia’s NIM Agent Blueprints aim to streamline this process by providing essential resources such as sample applications built with Nvidia NIM, NeMo, and partner microservices, along with reference code, customization documentation, and Helm charts for deployment.

By utilizing these pre-trained workflows, developers can significantly accelerate the development of their applications and deploy them seamlessly across various data centers and clouds. The blueprints can be tailored with proprietary data, enabling complex task performance through both information retrieval and agent-based workflows. Additionally, as users interact with these applications, the blueprints facilitate a continuous learning framework that enhances performance over time.

Available Blueprints and Future Plans

Currently, Nvidia offers three foundational blueprints: a digital human for customer service, generative virtual screening for rapid drug discovery, and multimodal PDF data extraction for enterprise retrieval-augmented generation (RAG). The customer service blueprint enables the creation of 3D avatar-based agents using Nvidia ACE, Omniverse RTX, Audio2Face, and Llama 3.1 NIM microservices. The drug discovery and data extraction blueprints incorporate advanced tools such as Nvidia NeMo Retriever, AlphaFold2, MolMIM, DiffDock, and Nvidia BioNemo.

"More NIM Agent Blueprints are under development for applications in customer service, content generation, software engineering, retail shopping advisors, and R&D. We plan to introduce new blueprints monthly," said Justin Boitano, Nvidia's lead for enterprise data center business.

Simplified Access and Deployment

Nvidia is making these blueprints more accessible through partnerships with leading technology solution providers, including Deloitte, Accenture, SoftServe, and World Wide Technology (WWT). These providers will incorporate the blueprints into their offerings, ensuring they are readily available to enterprise clients.

"By integrating NVIDIA’s catalog of workflows into Accenture’s AI Refinery, we can help our clients develop custom AI systems rapidly, transforming their business operations and enhancing customer service to drive stronger outcomes," stated Julie Sweet, chair and CEO of Accenture.

For enterprises aiming for independent deployment of custom blueprints within their data centers or on the cloud, Nvidia offers comprehensive infrastructure support through its global partners. This includes Nvidia-certified systems from Cisco, Dell Technologies, HPE, and Lenovo, as well as Nvidia-accelerated cloud instances from AWS, Google Cloud, Microsoft Azure, and Oracle Cloud Infrastructure.