What is Semantic layer?
A semantic layer is a model that sits between raw data and the people or tools asking questions of it. It defines business terms such as revenue or active customer once, together with how each is calculated, so that every query uses the same meaning.
How it works.
Without one, each report or query re-implements the definitions, and the same measure comes out differently in different places. A semantic layer names the measures, the dimensions they can be broken down by and the relationships between tables. Tools then query that model instead of the tables directly.
Why it matters.
For AI-assisted analytics the semantic layer matters more, not less. A language model that writes queries against raw tables has to guess which definition the business uses; querying a governed model removes the guess.
How Corlogic uses it.
Corlogic's semantic layer is StarEngine. It organizes relevant warehouse data into a governed star model of facts, dimensions, relationships and business definitions, and validates the model against the underlying data before an approved version can serve questions.
Keep reading.
Star schema
A star schema organizes analytical data into fact tables that record events and dimension tables that describe what those events involve.
Text-to-SQL
Text-to-SQL converts a question written in ordinary language into a SQL query a database can run, today usually with a large language model.
StarEngine
The governed star model of facts, dimensions and business definitions that Corlogic's answers are compiled from.
Start with a recurring process that is harder to coordinate than it should be.
Talk to us about where context gets lost, decisions stall, or follow-up breaks down across people and systems.
Talk to us