Accenture and NVIDIA partner to accelerate AI adoption in enterprises.

October 4, 2024

Accenture and NVIDIA announced a large-scale partnership to accelerate the adoption of AI in enterprises. In this partnership, Accenture will establish the "NVIDIA Business Group" to support the development and implementation of AI agents using NVIDIA technology, with over 30,000 experts.

Forming a specialized team of 30,000 people and deploying AI solutions in 120 countries.
This partnership aims to build the foundation for next-generation AI capabilities, leveraging Accenture's AI platform and NVIDIA technologies such as NVIDIA Inference Microservices and NeMo. This enables simplification of AI model deployment and scaling, construction of large-scale AI models, fine-tuning, and deployment.

Lan Guan, Accenture's Chief AI Officer, stated that the generative AI projects generated $3 billion in revenue in the previous fiscal year. However, the company estimates that only about 10% of businesses have fully scaled up generative AI.

At the core of this initiative is the "AI Refinery" platform developed by Accenture. The AI Refinery utilizes the entire NVIDIA AI stack, including NVIDIA AI Foundry, NVIDIA AI Enterprise, and NVIDIA Omniverse. This platform includes model training and tuning recipes for companies to build models using their own agents. It also provides access to multiple foundational models and integrates external and internal data.
Julie Sweet, the CEO of Accenture, stated that the AI Refinery creates opportunities for companies to rethink processes and operations, discover new ways of working, and scale AI solutions across the entire organization.

AI agents transform business processes of companies.

Accenture and NVIDIA are focusing on a new AI system called "agentic AI," which autonomously acts based on human intentions, creates new workflows, and rebuilds entire processes. This refers to AI systems that act autonomously based on human intentions, creating new workflows and rebuilding entire processes.
Accenture defines various categories of AI agents such as Research Agents and Orchestrator Agents. For example, Research Agents investigate topics without human intervention and create useful summaries for decision-making. Orchestrator Agents act as project managers managing multiple agentic workflow capabilities.
Justin Boitano, Vice President of Enterprise AI Software at NVIDIA, stated that by using NVIDIA's infrastructure, the execution speed of models can be improved by 2-5 times. Additionally, he mentions that new data generated during AI applications interacting with people can provide feedback for model improvement through continuous learning cycles.