NVIDIA Unveils Enterprise AI Factory and AI Blueprints for Intelligent Digital Agents

Talk to Me: NVIDIA and Partners Boost People Skills and Business Smarts for AI Agents

NVIDIA Enterprise AI Factory validated design and latest NVIDIA AI Blueprints help businesses add intelligent AI teammates that can speak, research and learn to their daily operations. Call it the ultimate proving ground. Collaborating with teammates in the modern workplace requires fast, fluid thinking. Providing insights quickly, while juggling webcams and office messaging channels, is a startlingly good test, and enterprise AI is about to pass it -- just in time to provide assistance to busy knowledge workers. To support enterprises in boosting productivity with AI teammates, NVIDIA today introduced a new NVIDIA Enterprise AI Factory validated design at COMPUTEX. IT teams deploying and scaling AI agents can use the design to build accelerated infrastructure and easily integrate with platforms and tools from NVIDIA software partners. NVIDIA also unveiled new NVIDIA AI Blueprints to aid developers building smart AI teammates. Using the new blueprints, developers can enhance employee productivity through adaptive avatars that understand natural communication and have direct access to enterprise data. Enterprises can use NVIDIA's latest AI Blueprints to create agents that align with their business objectives. Using the Tokkio NVIDIA AI Blueprint, developers can create interactive digital humans that can respond to emotional and contextual cues, while the AI-Q blueprint enables queries of many data sources to infuse AI agents with the company's knowledge and gives them intelligent reasoning capabilities. Building these intelligent AI agents is a full-stack challenge. These blueprints are designed to run on NVIDIA's accelerated computing infrastructure -- including data centers built with the universal NVIDIA RTX PRO 6000 Server Edition GPU, which is part of NVIDIA's vision for AI factories as complete systems for creating and putting AI to work. The Tokkio blueprint simplifies building interactive AI agent avatars for more natural and humanlike interactions. These AI agents are designed for intelligence. They integrate with foundational blueprints including the AI-Q NVIDIA Blueprint, part of the NVIDIA AI Data Platform, which uses retrieval-augmented generation and NVIDIA NeMo Retriever microservices to access enterprise data. Customers around the world are already using these AI agent solutions. At the COACH Play store on Cat Street in Harajuku, Tokyo, imma provides an interactive in-store experience and gives personalized styling advice through natural, real-time conversation. Marking COACH's debut in digital humans and AI-driven retail, the initiative merges cutting-edge technology with fashion to create an immersive and engaging customer journey. Developed by Aww Inc. and powered by NVIDIA ACE, the underlying technology that makes up the Tokkio blueprint, imma delivers lifelike interactions and tailored style suggestions. The experience allows for dynamic, unscripted conversations designed to connect with visitors on a personal level, highlighting COACH's core values of courage and self-expression. "Through this groundbreaking innovation in the fashion retail space, customers can now engage in real-time, free-flowing conversations with our iconic virtual human, imma -- an AI-powered stylist -- right inside the store in the heart of Harajuku," said Yumi An King, executive director of Aww Inc. "It's been inspiring to see visitors enjoy personalized styling advice and build a sense of connection through natural conversation. We're excited to bring this vision to life with NVIDIA and continue redefining what's possible at the intersection of AI and fashion." Watch how Aww Inc. is leveraging the latest Tokkio NVIDIA AI Blueprint in its AI-powered virtual human stylist, imma, to connect with shoppers through natural conversation and provide personalized styling advice. Royal Bank of Canada developed Jessica, an AI agent avatar that assists employees in handling reports of fraud. With Jessica's help, bank employees can access the most up-to-date information so they can handle fraud reports faster and more accurately, enhancing client service. Ubitus and the Mackay Memorial Hospital, located in Taipei, are teaming up to make hospital visits easier and friendlier with the help of AI-powered digital humans. These lifelike avatars are created using advanced 8K facial scanning and brought to life by Ubitus' AI model integrated with NVIDIA ACE technologies, including NVIDIA Audio2Face 3D for expressions and NVIDIA Riva for speech. Deployed on interactive touchscreens, these digital humans offer hospital navigation, health education and registration support -- reducing the burden on frontline staff. They also provide emotional support in pediatric care, aimed at reducing anxiety during wait times. Ubitus and the Mackay Memorial Hospital are making hospital visits easier and friendlier with the help of NVIDIA AI-powered digital humans. Cincinnati Children's Hospital is exploring the potential of digital avatar technology to enhance the pediatric patient experience. As part of its ongoing innovation efforts, the hospital is evaluating platforms such as NVIDIA's Digital Human Blueprint to inform the early design of "Care Companions" -- interactive, friendly avatars that could help young patients better understand their healthcare journey. "Children can have a lot of questions about their experiences in the hospital, and often respond more to a friendly avatar, like stylized humanoids, animals or robots, that speaks at their level of understanding," said Dr. Ryan Moore, chief of emerging technologies at Cincinnati Children's Hospital. "Through our Care Companions built with NVIDIA AI, gamified learning, voice interaction and familiar digital experiences, Cincinnati Children's Hospital aims to improve understanding, reduce anxiety and support lifelong health for young patients." This early-stage exploration is part of the hospital's broader initiative to evaluate new and emerging technologies that could one day enhance child-centered care. AI agents are one of the many workloads driving enterprises to reimagine their data centers as AI factories built for modern applications. Using the new NVIDIA Enterprise AI Factory validated design, enterprises can build data centers that provide universal acceleration for agentic AI, as well as design, engineering and business operations. The Enterprise AI Factory validated design features support for software tools and platforms from NVIDIA partners, making it easier to build and run generative and agent-based AI applications. Developers deploying AI agents on their AI factory infrastructure can tap into partner platforms such as Dataiku, DataRobot, Dynatrace and JFrog to build, orchestrate, operationalize and scale AI workflows. The validated design supports frameworks from CrewAI, as well as vector databases from DataStax and Elastic, to help agents store, search and retrieve data. With tools from partners including Arize AI, Galileo, SuperAnnotate, Unstructured and Weights & Biases, developers can conduct data labeling, synthetic data generation, model evaluation and experiment tracking. Orchestration and deployment partners including Canonical, Nutanix and Red Hat support seamless scaling and management of AI agent workloads across complex enterprise environments. Enterprises can secure their AI factories with software from safety and security partners including ActiveFence, CrowdStrike, Fiddler, Securiti and Trend Micro. The NVIDIA Enterprise AI Factory validated design and latest AI Blueprints empower businesses to build smart, adaptable AI agents that enhance productivity, foster collaboration and keep pace with the demands of the modern workplace.

NVIDIA is Betting Big on AI Agents for Enterprises | AIM

The newly released Tokkio and AI-Q AI Blueprints help developers build intelligent digital avatars that converse naturally, integrate with enterprise data, and adapt to emotional or contextual shifts. At Computex 2025, NVIDIA unveiled a suite of technologies designed to accelerate AI and integrate it into the workplace. The new Enterprise AI Factory validated design and an expanded collection of AI Blueprints aim to assist companies in deploying AI-powered digital coworkers capable of reasoning, speaking, and adapting. The company mentions that the concept of an AI teammate is evolving, and businesses will benefit from it. Whether assisting with fraud detection, customer support, or health education, AI agents are being trained not only on data but also on empathy, context, and memory. For enterprises juggling generative AI, data integration, and infrastructure scale, NVIDIA's new releases seem to pitch full-stack solutions. But this time, the focus is more specific: transforming AI agents into something meaningful. NVIDIA also announced plans to invest in an AI supercomputer in Taiwan. The newly released Tokkio and AI-Q AI Blueprints are designed to help developers build intelligent digital avatars that converse naturally, integrate with enterprise data, and adapt to emotional or contextual shifts. One example comes from the COACH store in Tokyo's Harajuku district, where shoppers now chat with "imma," a virtual stylist powered by Tokkio and NVIDIA ACE. Instead of playing pre-recorded scripts, it handles unscripted, real-time dialogue. For a brand built on personal style, the ability to personalise recommendations live is more than a gimmick. It sounds like a redefinition of customer service. Meanwhile, financial institutions like the Royal Bank of Canada are deploying agents like "Jessica" -- an internal-facing AI that helps employees handle fraud cases. Jessica retrieves the latest updates and documentation on fraud trends, enabling faster and more informed decisions. At hospitals in Taipei (Taiwan) and Cincinnati (US), AI agents serve pediatric patients and visitors. From providing directions to explaining treatments in child-friendly ways, these avatars reduce cognitive overload, both for patients and overworked staff. This may not be surprising, considering many users go to ChatGPT as their doctor for many things. These aren't isolated experiments. It shows the initial forms of agentic AI, in which software is interactive, cooperative, and predictive. With video now accounting for over half of all global data traffic, yet less than 1% of it analysed, NVIDIA also launched a new blueprint to help organisations tap into this underused resource. The AI Blueprint for Video Search and Summarisation (VSS), built on NVIDIA's Metropolis platform, enables enterprises to develop vision AI agents capable of understanding and summarising vast volumes of video content. These agents, powered by vision-language models (VLMs) and large language models (LLMs), can perform real-time analysis, generate summaries, and even narrate complex events. From monitoring manufacturing floors to managing urban infrastructure, the VSS blueprint has already been used by companies like Pegatron and smart city operators in Kaohsiung, Taiwan. Results include a 67% reduction in defect rates and up to 80% faster incident response times. With features like RAG, speech transcription, and GPU-accelerated processing, the VSS blueprint represents a leap in productivity for video-heavy industries. Even the NHL is getting in on the action -- using AI agents to tag, summarise and retrieve game footage in real time. Building these agents is a full-stack endeavour behind the scenes. NVIDIA's Enterprise AI Factory validated design offers a blueprint for deploying agentic AI on modern, accelerated infrastructure. It is tailored to run on RTX PRO 6000 Server Edition GPUs, NeMo Retriever, microservices, and partner integrations -- essentially a plug-and-play back end for conversational, knowledge-aware agents. The AI Data Platform reference design, meanwhile, brings compute closer to storage. That's critical for AI agents that must search millions of internal documents and respond in real time. Companies like IBM, VAST Data and NetApp are embedding GPUs and NeMo Retriever into their storage solutions, enabling AI to reason directly at the data layer. This infrastructure is already being manufactured at scale. Taiwanese Original Design Manufacturers (ODMs) like Foxconn, Supermicro, ASUS, and Wistron are producing AI Data Platform-compatible systems that include the full NVIDIA stack -- from NVIDIA's BlueField DPUs to Spectrum-X Ethernet. In short, NVIDIA's ecosystem now spans blueprints, silicon, software, and deployment partners. With orchestration tools from Red Hat, Nutanix, and Canonical, as well as data tools from platforms like Galileo and DataRobot, enterprises can easily use AI agents with NVIDIA's tech.