Better together: Ontotext and Semantic Web Company become Graphwise
Published:
Content Copyright © 2025 Bloor. All Rights Reserved.
Also posted on: Bloor blogs
Longtime partners Ontotext (developer of the GraphDB graph database we’ve covered previously) and Semantic Web Company (developer of the PoolParty Semantic Suite) have come together to form Graphwise, and with it the graph AI platform of the same name that aims to combine graph technology with generative AI.
The Graphwise raison d’être is that it allows you to leverage knowledge graphs as the core part of a semantic layer that sits between your data and your AI processes, providing insights and context derived from your data to said processes via knowledge-graph-enhanced RAG. At the same time, the platform provides substantial graph automation functionality (such as its Taxonomy Advisor, which automates taxonomy creation and enrichment) to help you build your knowledge graphs in the first place. These automation features themselves leverage generative AI and LLMs, making this a virtuous cycle: graph automation drives graph RAG, which drives graph automation, which drives graph RAG, which drives graph automation, and so on.
Notably, Graphwise operates multimodally – which is to say, it handles structured, unstructured, and semi-structured data equally well. Moreover, the knowledge graphs the platform creates can fill multiple roles via different architecture patterns, acting as domain knowledge models, content hubs, and/or data fabrics. This also impacts what kind of data the graph is equipped to serve up during RAG. The use of standard metadata schemas and reuse of existing domain knowledge lower the cost of building knowledge graphs, as well as improving interoperability and easing LLM integration (since the major LLMs are already trained on popular metadata schemas). Entity linking is also available, alongside a customizable inference engine
Additional features include a “content chat” natural language interface driven by generative AI, self-service data/knowledge access and training portal, and RAG cost/performance optimization, the latter delivered by employing only the most appropriate and efficient methods and AI models in each part of the RAG process. It is also worth mentioning that all original GraphDB and PoolParty functionality is retained in Graphwise, and indeed powers much of what we have already covered.
In short, the Graphwise platform is one of the first major efforts we’ve seen to weave together knowledge graph and generative AI technologies, using each to drive the other to new heights. We shall certainly be keeping our eye on it as it develops further.