McKinsey & Company is utilizing knowledge graphs to provide context for enterprise artificial intelligence applications. According to James Kaplan, a distinguished partner at the firm, these graphs connect disparate enterprise data with business relationships, offering a more intuitive structure than traditional relational databases for handling complex or ambiguous information.
The firm employs a system called EcliptOS, which functions as an AI operating system. It uses a semantic data layer to organize data and relationships, facilitating the connection between corporate strategy and execution through agentic workflows. This approach allows for the creation of a virtual graph that links multiple databases.
AI technology is used to transform unstructured information into structured data and deterministic business rules. This process, which previously required manual labor from data scientists or analysts, now enables organizations to programmatically interrogate messy data and store it within a graph structure to support generative AI.
Business priorities, such as enhancing customer experience or improving productivity, dictate where these AI improvements are applied. The effectiveness of the system depends on the density of interconnections among data nodes, which increases the intelligence available within the knowledge graph for decision-making.
