https://drive.google.com/file/d/1j-vI_R2fV3S0_L8dia8_fWjOCHNIDcom/view?usp=share_link

Co-presenting with Ora Lassila of the Amazon Neptune:

Enterprise organizations possess vast technical documentation repositories containing invaluable institutional knowledge trapped in unstructured formats. While Semantic Web standards offer mature knowledge representation foundations, the challenge remains extracting and connecting knowledge from documents never designed for machine interpretation.
This presentation introduces a practical pipeline combining RDF transformation techniques with LLM capabilities to build "context graphs" - the semantic layer capturing not just what happened, but why decisions were made.

We demonstrate embedding lightweight knowledge graphs within technical documents, enabling dual pathways: SPARQL-based reasoning for structured queries and LLM-powered insight extraction for pattern recognition across document corpora. This hybrid approach transforms decades of accumulated documentation into queryable organizational intelligence.

Drawing from real-world implementations, we present concrete examples of extracting entities, relationships, and decision context from technical documentation. Attendees will learn a repeatable framework applicable across industries - from engineering specifications to regulatory documentation to operational procedures.

The result: institutional knowledge that compounds over time rather than disappearing with departing experts.