We're delighted to be exhibiting at SEMANTiCS 2026 – the 22nd International Conference on Semantic Systems – taking place September 15–17 in Ghent, Belgium. Come and find us at our booth to learn how metaphactory helps enterprises build trusted, governed, AI-ready knowledge architectures.

Opportunities to connect with us

At our booth

Visit us on the exhibition floor to see metaphactory in action. We'll be showcasing how enterprises use our Knowledge Graph Platform to power Enterprise Information Architecture and an AI-ready Semantic Layer — preventing hallucinations, enabling semantic data discovery, and grounding agentic AI in trusted, governed knowledge. Bring your use case and let's talk.

At our tutorial

Enterprise Semantic Layer for SOP Automation and Agentic AI Thomas Kaminski, Senior Knowledge Graph Solutions Engineer, Digital Science / metaphacts

Enterprises rely on Standard Operating Procedures to govern critical operational decisions, yet these procedures are typically maintained as static documents requiring manual interpretation. This tutorial introduces an Enterprise Semantic Layer approach for transforming SOPs into agent-ready semantic assets. Participants will learn how to extract operational business rules from SOP documents, normalize them into reusable rule statements, tag them with Business Objects, and connect those Business Objects to source systems, datasets, and data products. The session walks through the full workflow — SOP ingestion, business rule extraction, Business Object identification, RDF materialization, source-system mapping, and agentic retrieval through a semantic layer — using a realistic billing/refund scenario.

This tutorial is intended for practitioners and researchers interested in knowledge graphs, semantic governance, enterprise AI, ontology engineering, and trustworthy multi-agent automation.

At our conference session

Engineering the Enterprise Semantic Layer - From Business Meaning to Machine-Executable Data Architecture

Thomas Kaminski, Senior Knowledge Graph Solutions Engineer, Digital Science / metaphacts
Business language and technical data architecture still live in separate worlds — and that gap is quietly undermining enterprise AI initiatives. Data catalogs and APIs have improved access to data, but they rarely capture what that data means or how it connects to the business concepts that govern it.

In 15 minutes, Thomas presents a concise engineering blueprint for an Enterprise Semantic Layer: an architectural approach that connects Business Object ontologies, physical data assets, and technical metadata through explicit semantic models and schema mappings. Drawing on real enterprise implementations, he shows how RDF-based schema representations and business object modeling give both humans and AI agents a shared, machine-interpretable view of the enterprise data landscape — transforming Enterprise Information Architecture from static documentation into an operational layer that enables semantic data discovery, automated lineage, and explainable AI at enterprise scale.

What we'll be showcasing

metaphactory – Knowledge Graph Platform for Enterprise Information Architecture

metaphactory is the platform enterprises use to build, govern, and activate knowledge graphs at scale. At SEMANTiCS 2026, our team will be demonstrating how metaphactory delivers:

• Enterprise Information Architecture: Create a semantic digital twin of your organization — connecting data sources, business objects, and technical metadata into a single, governed knowledge layer.
• AI-ready Semantic Layer: Ground your AI agents and LLMs in trusted enterprise context. Prevent hallucinations by providing machine-interpretable business knowledge rather than raw, siloed data.
• Scalable governance & regulatory readiness: Support production-ready knowledge graphs with versioning, auditability, editorial workflows, and compliance-ready traceability.
• Trustworthy & explainable AI insights: Enable reasoning systems that can show their work — connecting conclusions back to governed, auditable semantic models.