Get a primer on how knowledge graphs enhance AI and KM with Factor

Knowledge graphs can revolutionize knowledge management by structuring and linking information to reveal hidden relationships and insights based on context.

Semantic networks map how concepts, documents, and data connect, while AI can be leveraged for search, discovery, and content recommendations. Together, they enable organizations to move beyond static repositories to dynamic, intelligent systems that deliver the right knowledge to the right people at the right time.

Bram Wessel, founder/principal at Factor and Bob Kasenchak, lead information architect at Factor joined KMWorld for a webinar, The Power of Context: Using AI and Knowledge Graphs to Enhance KM, to discuss semantics, knowledge graphs, ontologies, and more.

Graphs allow for complex, multidimensional connections, making them especially valuable for nuanced analytics, metadata alignment, and AI, Wessel and Kasenchak said.

Knowledge graphs (KGs) are what computer scientists call ontologies. They link concepts, entities, and external data sources. KGs can combine taxonomies and domains to model enterprise KM concepts, classification, metadata sources. This type of graph is great for RAG and other AI-enhancing applications, they explained.

Content graphs combine concepts with documents/content and documents and metadata. Documents are represented (not stored) in the graph. Using this type of graph solution offers powerful analytics and queries. It can be used for content management, KM, metadata management, and more.

People graphs connect concepts to people. They include information about employees, customers, and more, and can connect expertise, authoring of documents, sales, interests, locations, roles, and teams. This type of graph is useful for sales, HR, or marketing.

Context graphs add contextual information to existing data, including when, where something happened; timespans; provenance; and confidence. These graphs are powerful when combined with (or extension of) KGs. Information is embedded in edges (relationships).

Decision graphs are another emerging graph type, they noted. It’s used for agentic AI reasoning, process modeling, or decision modeling. This graph records and replicates how decisions are made within workflows.

According to Wessel and Kasenchak, KGs provide structured, machine-readable foundations for AI applications to:

• Reduce hallucinations and improve accuracy of natural language queries
• Add semantics to define critical enterprise concepts and help conceptually disambiguate text strings
• Allow explainability and increase trust by citing sources for answers
• Enable GraphRAG and similar processes to improve information retrieval
• Afford centralized and integrated data sources for queries and analytics

For the full webinar, featuring a more in-depth discussion, Q&A, and more, you can view an archived version of the webinar here.