At NVIDIA GTC, Materna is introducing Sinfonie, a new platform for sophisticated digital humans. The solution combines generative AI, voice interaction, and interactive 3D avatars to create digital expert assistants that support employees in knowledge work, internal processes, and complex issues. All components can be operated entirely on the customer’s premises without data ever leaving their own infrastructure. Sinfonie emerged from the experience gained during a customer project for a North Rhine-Westphalia state administration in Germany. Materna has further developed the approach created there, using open-source and NVIDIA technologies, into a transferable platform for professional use in government and business.
Digital humans are considered the next evolutionary step beyond traditional chatbots and AI assistants. They communicate through speech, facial expressions, and gestures, thereby enabling a more natural interaction between humans and AI. At the same time, knowledge, data, and decision-making processes remain entirely under the control of the respective organization.
“With Sinfonie, we’re taking the next step from text-based AI assistants to digital expert companions. People communicate intuitively through speech and personal interaction. That’s exactly where Digital Humans come in, creating low-threshold access to knowledge, processes, and AI-powered support,” says Michael Hagedorn, CEO of Materna. Every Digital Human created with Sinfonie has its own voice and individual identity. Combined with facial expressions, gestures, and natural language interaction, this creates a digital counterpart that conveys information in a way that is understandable, trustworthy, and intuitive.
From Practice, For Practice
The idea for Sinfonie originated as part of a digital human project for a state administration in Germany. The customer’s goal was to support employees with questions about internal processes, regulations, and knowledge bases while simultaneously meeting the highest standards for data protection, information security, and digital sovereignty.
While the original project for the customer focused primarily on supporting employees with knowledge and service inquiries, Sinfonie expands this approach into a platform for digital humans. The solution combines knowledge management, voice dialogue, AI-powered assistance, and visual interaction into a unified user experience. This creates a digital expert assistant that conveys information in a way that is intuitive for humans. Its applications range from knowledge management and onboarding to internal service processes and support for complex specialized procedures across all industries.
NVIDIA Technologies Enable a New Generation of Digital Humans
Sinfonie was specifically developed for on-premises operation. All components—from speech recognition and AI models to avatar animation—run locally and require minimal infrastructure. For animation, engines developed by Materna are available, based either on open-source technologies or NVIDIA technologies such as NVIDIA Audio2Face. This makes the platform particularly suitable for government agencies, public IT service providers, and regulated industries that want to use generative AI without relinquishing control over their data.
Based on this, Sinfonie can understand spoken language, process information from knowledge bases and specialized systems, and convey answers via an interactive 3D avatar using natural language, facial expressions, and gestures. The platform combines modern generative AI with the requirements of secure and controlled IT environments.
“Data sovereignty and modern AI experiences are not mutually exclusive. With Sinfonie, speech understanding, knowledge retrieval, and the avatar run entirely within the customer’s infrastructure. For many of our customers, this is a prerequisite for using generative AI productively at all,” says Thomas Feld, Vice President of Data Management & AI at Materna.
Sovereign AI with Human Interaction
Sinfonie was developed specifically for organizations and companies that want to leverage modern AI capabilities without relinquishing control over their data. The platform can be operated in controlled environments and integrates existing knowledge sources, role- and permission-based models, and business processes. With this solution, Materna specifically targets organizations that want to leverage the added value of generative AI while simultaneously meeting stringent requirements for compliance, data protection, and digital sovereignty.
Digital Humans as the Next Stage in AI Development
With Sinfonie, Materna is actively shaping one of the most important trends of the next generation of AI: collaboration between humans and digital assistants through natural communication. While the first wave of generative AI was primarily characterized by chatbots, digital humans are creating a new level of interaction. They combine voice dialogue, subject matter expertise, and visual presence to form a trustworthy digital counterpart.
For Materna, Digital Humans therefore serve as the bridge between powerful AI systems and the people who use them in their daily work. With Sinfonie, Materna is demonstrating at NVIDIA GTC how this vision can already be implemented today in confident, productive use cases for government agencies, public IT service providers, and regulated industries.
Images