The layer that tells your query engine what the data means
Your warehouse knows the column is called “cust_stat_cd”. It does not know that finance and sales disagree on what counts as active. TopQuadrant holds the meaning, the metric definitions, and the entitlements in one governed layer that BI tools, applications, and agents all resolve against.
Trusted to govern the data behind the world's most regulated decisions.
Business logic leaks, and the semantic layer is where it should live
Definitions end up scattered across dbt models, dashboard formulas, stored procedures, and analyst notebooks. Each copy drifts. A governed semantic layer puts the definition in one place, connects it to the physical data through mappings you can inspect, and serves it to every consumer through the same interface.
More accurate AI answers when models reason over governed context rather than raw schemas.
Cheaper to operate than assembling and re-assembling context per application.
One definition per metric, resolved at query time rather than copied into each tool.
Governed metrics & definitions
Define a metric once, in language the business agreed on
Every metric carries its business definition, its owner, its calculation, and its mapping to physical columns. When the definition changes, it changes in one place and every consumer picks it up.
- Definition and calculation together. The plain-language meaning and the logic live on the same governed asset.
- Named owner per metric. Disagreements resolve to a person, not a thread.
- Version history. See what the metric meant last quarter, and what changed since.
Mapping to physical data
Connect meaning to the tables it actually lives in
Mappings from governed concepts to physical schemas are explicit, inspectable assets. When a source system changes a column, you can see which concepts and metrics are affected instead of finding out from a broken dashboard.
- Explicit mappings. Concept to table to column, held as data you can query.
- Virtualised access. Query the graph and resolve to Snowflake, Databricks, or Postgres without moving the data.
- Schema change alerts. A source change surfaces every affected concept, metric, and consumer.
Lineage & explainability
Every number arrives with its origin
Trace any value from the report back through the metric definition, the mappings, and the transformations, to the system it came from. This is what makes an answer defensible in a review, an audit, or a model risk committee.
- End-to-end lineage. Source, model, metric, and report connected as one graph.
- Provenance on the answer. A returned value carries what it means and where it came from.
- Impact before change. See which reports depend on a definition before you edit it.
Entitlements at query time
Enforce access on the meaning, wherever it is queried
Permissions expressed against governed concepts apply consistently across every consumer, including agents. A rule about salary data holds whether the request arrives from a dashboard, a notebook, or a natural-language question.
- Concept-level rules. Entitlements attach to the governed asset, so they follow it everywhere.
- Evaluated per request. Access is resolved at query time against the requester’s role.
- Consistent for agents. An agent sees exactly what its caller is entitled to see, and no more.