Metaphacts
About the organisation
Metaphacts is a knowledge graph platform vendor specialising in semantic data management, hybrid AI and enterprise information architecture. The company’s platform, metaphactory, supports the construction, governance, exploration and visualisation of large RDF-based knowledge graphs across industry, life sciences and cultural heritage. Metaphacts sits squarely in the symbolic AI tradition, with a long-standing commitment to W3C standards, ontologies and federated semantic infrastructure, and is an active voice in the international knowledge graph community through events such as the Knowledge Graph Conference, SEMANTiCS and Connected Data London. Larry Swanson, who attended the first GRAPHIA / LUMEN Innovation Prototyping Lab (IPL) in Zagreb and moderated the Day 2 panel on hybrid AI and the future of semantic research and industry, contributes to that community-facing role. The company’s recent work on Enterprise Information Architecture, drawing on the data-centric enterprise tradition associated with Dave McComb, treats business entities as first-class semantic objects in their own right. That perspective on knowledge representation aligns naturally with the design principles guiding the GRAPHIA federated SSH knowledge graph and the project’s commitment to interoperability across heterogeneous research domains.
What challenges can the GRAPHIA project help solve?
Two threads from the IPL panel sit at the heart of the alignment between metaphacts and GRAPHIA. The first is hybrid AI: how to combine LLMs with knowledge graphs and ontologies so that AI systems answer questions in ways that can be checked, audited and corrected. Metaphacts has been developing platform-level patterns for that integration; GRAPHIA brings a federated SSH knowledge graph with explicit attention to ontology design, provenance and competency-question methodology.
The second thread is trust. The panel converged on transparency, provenance, integrity and evidence-based policy as the foundations of trustworthy semantic infrastructure, with parallels drawn to JUST data principles of judicious, unbiased, safe and transparent data practice.
SSH is a domain where these properties are unusually difficult and unusually consequential, since cultural and contextual nuance resist superficial aggregation. GRAPHIA offers metaphacts a substantive testbed for hybrid AI patterns under conditions that will surface real challenges, alongside a community of partners willing to engage with the harder questions about cross-domain interoperability and the representation of disciplinary plurality.
What is the proposed use case?
A potential use case has emerged from the panel discussions: an alignment between metaphacts’ platform capabilities and the emerging GRAPHIA SSH knowledge graph, framed around hybrid AI integration and enterprise reach. On the platform side, metaphactory’s strengths in graph construction, governance, ontology management, and visual exploration of federated RDF resources map onto challenges raised in the GRAPHIA workshops, including ontology alignment, scalable querying across heterogeneous endpoints, and the design of usable interfaces over scholarly knowledge graphs. On the audience side, metaphacts works with industrial and life-science clients whose questions increasingly touch SSH territory, from supply-chain ethics to cultural-context modelling for autonomous systems and AI governance. There is also a thematic alignment with the Data Product Passport concept that has been gaining traction in the wider knowledge graph community, addressing provenance and access at the level of individual triples. None of this constitutes an implementation commitment. The IPL and subsequent conversations have established mutual interest and a shared vocabulary; a more concrete scoping exercise would be the appropriate next step.
What type of SSH data or content are involved?
In this use case, metaphacts contributes platform expertise, with the relevant SSH content flowing from GRAPHIA’s federated corpus and its partners. Of particular interest are the structured outputs of the SSH knowledge graph as it is built and federated: ontologies and competency questions, citation-aware research metadata, full-text-derived entities for people, places, organisations and events, and cross-disciplinary linkages across sociology, cultural studies, ethics, history and adjacent humanities. Metaphacts brings to this material a long history of working with cultural heritage and scholarly knowledge graphs, and the kinds of provenance and shape-validation patterns (ShEx and SHACL, VoID summaries, named graphs) that are already part of GRAPHIA’s emerging stack. There is potential for metaphacts to advise on or contribute reference patterns for SSH-specific concerns such as multilingual ontology alignment, conflicting interpretive claims, and the representation of disciplinary plurality without flattening it into a single dominant ontology. Any data flow remains conditional on further discussion between metaphacts and the relevant GRAPHIA team.
What technical aspect would be used?
The technical interest is principally on the platform and AI side. Three areas stand out. First, hybrid AI architecture: the integration of LLMs with the SSH knowledge graph, including retrieval-augmented generation grounded in RDF, agent-based query construction such as the text-to-SPARQL pattern demonstrated by Odoma, and the broader question of how natural language interfaces and structured graph queries can coexist in a single research workflow. Metaphactory has been evolving in this direction, and the IPL panel surfaced strong mutual interest.
Second, knowledge graph platform infrastructure: visualisation, governance, ontology editing, and federated SPARQL endpoints over heterogeneous backends, all areas where metaphacts has substantial production experience. Third, trust and provenance mechanisms, including triple-level access control patterns adjacent to the Data Product Passport concept, named-graph strategies, and ShEx-based shape validation, all of which surfaced repeatedly during GRAPHIA workshop sessions.
The most realistic near-term contribution from metaphacts would be conceptual and architectural input. Prototype-level engagement remains contingent on further conversations between the two organisations.