Last week, Coreon exhibited at Quanos Connect in Nuremberg — and played a key role in one of the conference’s most discussed announcements: the ST4 AI Language Factory, Quanos’s AI lighthouse project for fully automated translation in CCMS environments. Coreon’s Multilingual Knowledge Graph is the component that makes the quality promise credible. While Large Language Models alone produce fluent but inconsistent output, Coreon provides the terminological precision and conceptual structure necessary to bring AI translation below the error rate of the current human-based process. Not just approximately. Demonstrably.
The argument we’ve been making for years ships as a product: technical documentation, translated end-to-end with a human-on-the-loop, but not expensively in-the-loop – at a quality level the translation industry said was impossible without post-editing. The pipeline connects Quanos ST4 directly to the ESTeam Language Factory’s orchestration layer. It draws on a Content Repository, which allows for safe content recycling with variable granularity, as well as Coreon. The result is a complete, automated translation workflow for technical content at enterprise scale. Six Launch Partner slots are open for September 2026 deployment.
Coreon features a Multilingual Knowledge Graph. Every node in the graph is a concept — a meaning that is defined once and is thus language-agnostic — which contains all language expressions. Unlike term databases, the nodes in the graph are connected to each other , expressing relationships such as narrower/broader, is-a or is-part-of. One graph, thousands of concepts, dozens of languages. No duplication, no drift. When a new language is added to the graph, neither the concept nor its relations with other concepts change; only its expression does. This architecture enables the Language Factory to translate consistently at scale: the AI does not decide what a term means across languages – Coreon has already decided, and the AI draws from that.
Beyond translation, Coreon serves as a master knowledge layer for any system that needs to understand what your company means by its own internal vocabulary. This includes applications such as chatbots, search engines, product classification and regulatory documentation. It can be queried via API, MCP, and SPARQL — meaning it can be integrated into any existing content or AI infrastructure without a rip-and-replace. For customers who already use a Coreon graph for terminology governance, the Language Factory integration is not an additional product to manage. It is your existing knowledge, finally doing what it has always been capable of.
The reaction at Quanos Connect was everything you would expect from a genuinely disruptive launch: strong interest on the floor, full sessions, and conversations that continued well beyond the conference, spreading across LinkedIn and other professional communities. This is because the solution challenges assumptions the industry has held for decades. We welcome the debate. Looking ahead, 2026 will be a full year — and not only in terms of translation. The ST4 AI Language Factory is a first expression of a larger opportunity: multilingual knowledge systems as the foundation for a LangOps strategy. Automating translation is the entry point. Operationalizing language across the enterprise — connecting knowledge, content, and AI workflows in every language the business operates in — is the destination. We are looking forward to that conversation.
Upcoming Events
TAUS
Jochen Hummel is juggling a packed schedule these days — after Quanos Connect, he’s now headed for TAUS Massively Multilingual AI Conference in beautiful Rome. The TAUS conference puts European language technology companies in the spotlight. From machine translation and quality assessment to AI-driven coordination and multilingual content management systems – the event celebrates companies whose projects transcend language barriers.
Jochen will present the AI application layer that turns multilingual content into a strategic asset: The Language Factory by ESTeam.