Nvidia has unveiled Nvidia NIM Agent Blueprints, which it describes as an enterprise-destined catalog of pre-trained, customizable AI workflows designed to simplify the development and deployment of GenAI applications.
The blueprints are designed to provide a comprehensive toolkit for enterprise developers to build AI solutions for various use cases, including customer service, drug discovery in healthcare and retrieval-augmented generation (RAG).
Enterprises can use Nvidia’s AI Enterprise Platform, which includes NIM microservices and the Nvidia NeMo framework, to build data-driven AI systems.
Nvidia offers up blueprints for GenAI applications
Among the initial offerings is the digital human blueprint for customer service, which uses 3D-animated avatars to provide engaging and lifelike customer interactions. The company says this will allow enterprises to offer more personalized and efficient support to customers.
The announcement cites Gartner research predicting that 80% of conversational offerings will embed GenAI by 2025, up from 20% in 2024.
Another key component is the generative virtual screening blueprint for drug discovery, which accelerates the identification of promising drug-like molecules. It uses AI models like AlphaFold2, known for its impact on protein structure pediction, MolMIM, a Nvidia model that generates molecules, and DiffDock, known for its ability to bind small molecules to their protein targets.
A third element, the multimodil PDF data extraction blueprint, accesses vast amounts of enterprise data to allow organizations to build accurate RAG systems.
As part of the announcement, Nvidia confirmed it is partnering with companies including Accenture, Deloitte, Dell and HPE in order to deliver these blueprints to enterprises using AI-ready cloud platforms and on-premises servers backed by Nvidia’s infrastructure.
Nvidia CEO Jensen Huang commented: “Generative AI is advancing at lightspeed… The enterprise AI wave is here.”
More from TechRadar Pro
https://cdn.mos.cms.futurecdn.net/DVffQnnibMWmNpx2Wfb5Se-1200-80.jpg
Source link
