NVIDIA announced its new Agentic AI Blueprints at CES 2025, enabling developers to create custom AI agents that can reason, plan, and take actions autonomously. These blueprints integrate NVIDIA NIM microservices, NVIDIA NeMo, and frameworks from leading providers, marking a significant leap in generative AI. NVIDIA’s CEO, Jensen Huang, emphasized that these tools move beyond basic chatbot interactions, offering solutions for complex, multi-step challenges.
Agentic AI represents the next step in the generative AI journey. According to Huang, enterprise AI agents will play a central role in AI factories, driving unprecedented intelligence and productivity across industries. NVIDIA's blueprints aim to accelerate the deployment of agentic AI applications, helping businesses automate workflows and derive real-time insights from diverse data sources such as videos, PDFs, and images.
Key providers like CrewAI, Daily, LangChain, LlamaIndex, and Weights & Biases collaborated with NVIDIA to develop agentic AI blueprints. These solutions are powered by NVIDIA AI Enterprise software, including NIM microservices and NeMo, and aim to transform industries by offering the following:
CrewAI: Enhances code documentation with NVIDIA NIM microservices, ensuring accessible and well-organized repositories.
Daily: Develops voice agent solutions using NVIDIA Riva automatic speech recognition for real-time conversational AI.
LangChain: Facilitates structured report generation with tools for web searches and detailed topic reports.
LlamaIndex: Creates research assistants for blogging, using AI-driven content generation with source attribution.
Weights & Biases: Integrates debugging and tracking tools for seamless AI development and deployment.
NVIDIA unveiled its own AI blueprints for converting PDFs into podcasts and video summarization, targeting industries with heavy documentation requirements. These tools provide actionable insights while maintaining data security.
For example, the PDF-to-Podcast AI Blueprint leverages NIM microservices to process lengthy PDF files into concise audio summaries. These can be narrated in natural voices, making information easily accessible for professionals and researchers.
NVIDIA introduced Launchables, pre-configured environments that simplify the AI development process. These ready-to-use setups include all necessary components for prototyping and deployment, significantly reducing development overhead.
Launchables can run on data center platforms such as Dell Technologies, Lenovo, and Supermicro, or on cloud services like AWS, Google Cloud, Microsoft Azure, and Oracle Cloud Infrastructure.
NVIDIA and Accenture partnered to introduce AI Refinery for Industry, which features 12 pre-configured agent solutions tailored to various industries. These include:
Public Sector: Recruitment solutions.
Telecommunications: Agent-assist contact centers.
Banking: Legacy modernization tools.
Life Sciences: Clinical trial companions.
Consumer Goods: Revenue growth management systems.
Accenture plans to release over 100 such solutions by year-end, supporting industries like automotive, technology, manufacturing, and healthcare.
NVIDIA’s Agentic AI Blueprints and associated tools enable enterprises to:
Develop customized AI systems that streamline operations.
Access secure and efficient tools for real-time data analysis.
Build scalable AI applications with minimal setup complexities.
Conclusion
NVIDIA’s Agentic AI Blueprints signify a pivotal moment in AI innovation. By providing tools that are adaptable across industries, NVIDIA is enabling businesses to automate complex workflows, enhance productivity, and unlock new growth opportunities.