Semantic Layer
A governed layer that defines business metrics, dimensions and joins once so BI tools, applications and AI agents query data consistently.
When marketing, sales and finance calculate revenue or pipeline differently, meetings turn into reconciliation; a semantic layer gives one governed definition and offers AI assistants a safer interface than raw tables.
A data team defines net new ARR once in the dbt Semantic Layer, with filters for contract type and a monthly time grain; the BI dashboard, a spreadsheet connector and an AI assistant all query that metric and return matching figures.
It only standardizes definitions teams agree on; disputes about what a metric means remain an organizational task. Adoption requires tools to query through it, and complex metrics or caching rules add latency and maintenance work.
Models describe entities, dimensions, measures and metrics on top of tables in a Data Warehouse, usually in YAML or code under version control. A query engine such as MetricFlow or Cube compiles metric requests into SQL with the right joins and aggregations and exposes them through SQL, REST or GraphQL APIs. Access rules and caching or pre-aggregations are applied centrally, which supports Data Governance and consistent Business Intelligence (BI).
AI Governance, Revenue Operations
dbt Developer Hub — dbt Semantic Layer — https://docs.getdbt.com/docs/use-dbt-semantic-layer/dbt-sl; Cube — Introduction to Cube — https://docs.cube.dev/docs/introduction