[ Learn ]
Semantic layers, ontologies and AI on your database
Plain-language guides, templates and tools for the people who have to make an AI agent return correct numbers from a real company database.
Learn
A semantic layer maps business terms like revenue to the tables and joins behind them. An example, and what AI agents need from one.
A context layer is what an AI agent should know about your data before it writes a query. What goes in one, and how it differs from a semantic layer.
A context graph connects your data's meaning to the reasons behind each answer. A database-focused definition, with an example.
Text-to-SQL (NL2SQL) writes valid SQL easily. On real company databases it guesses joins and codes. Why that happens, and what fixes it.
An ontology gives AI a map of your business: what things are and the rules for how they relate. An example, and how to build one from your data.
Ontology, semantic layer or knowledge graph? How the three differ, and which one an AI agent working on your database needs.