McKinsey Develops EcliptOS AI Operating System With Knowledge-Graph Infrastructure

McKinsey is developing EcliptOS as an AI operating system that uses knowledge-graph infrastructure to connect enterprise information for generative-AI applications.

The platform is intended to link business strategy with execution through agentic systems. Its engineering work centers on graph-based data pipelines and a semantic data layer that gives data shared definitions and relationships.

McKinsey is hiring engineers to build entity and relationship models, taxonomies, identifiers, semantic tags and graph stores for EcliptOS. Those components help describe what information represents, how entities relate to one another and which rules govern those connections.

A semantic layer can help organize data spread across databases, documents and other enterprise systems. Knowledge graphs connect those elements into a unified view, allowing generative-AI applications to interpret business information in context.

James Kaplan, a distinguished partner at McKinsey, said EcliptOS uses a semantic data layer to organize data and relationships for generative-AI applications.

“The richer the interconnections among nodes, the more intelligence you have in the graph,” Kaplan said. “What we in effect created was a graph of databases.”

Knowledge graphs and ontologies can provide infrastructure for combining structured and unstructured data. Ontologies define the concepts and relationships used to represent information, while the graph connects those relationships across systems.