Skip to main content
Prophecy uses knowledge graphs to understand your data environment for AI capabilities (Prophecy Agent and Copilot). A knowledge graph is an index that maps the relationships between data in your fabric. The knowledge graph contains metadata about tables, schemas, columns, and data types—not your actual data. When you interact with AI, Prophecy uses a knowledge graph to add context to your prompts. This context helps the Large Language Model (LLM) generate accurate SQL code that references the correct tables and columns in your warehouse.

​
How knowledge graphs work

When you interact with AI in Prophecy, the system enriches your prompt with context from the knowledge graph before sending it to the Large Language Model (LLM). This enrichment step adds metadata about your project’s datasets, schemas, and other entities, which helps the LLM understand what you’re referring to. The LLM then generates SQL code based on this enhanced context.

​
Knowledge graph generation

Prophecy generates one knowledge graph per fabric. Each knowledge graph indexes metadata from all data connections in the fabric. The indexer crawls your data storage using either your identity or a separately configured identity. You can schedule automatic refreshes or trigger manual indexing to keep the knowledge graph current.
Prophecy uses the knowledge graph of the fabric attached to your project. If you attach to a different fabric, AI features will use that fabric’s knowledge graph.

​
What’s next

Configure the knowledge graph indexer to schedule automatic indexing or set up separate authentication credentials.