Business Ontology: A Governed Approach to Business Terminology
Today we’re announcing Business Ontology (public preview soon), a governed framework for teams to define, manage, and approve shared business terms, entities, relationships, and metrics. It’s part of Snowflake Horizon Context, the governed context layer within Horizon Catalog. Business Ontology ensures consistency by providing a single source of approved definitions, enabling AI agents to interpret enterprise data as intended by the business.
Teams can reuse this knowledge across agents and applications, maintaining domain-specific variations for areas like sales, finance, and marketing. This avoids redundant definition for each tool and reduces the risk of incorrect interpretations, potentially leading to poor decision-making.
Horizon Context consolidates knowledge from existing tools used by teams. Its out-of-the-box metadata connectors gather schemas, query logs, dashboard definitions, and usage signals from sources like Tableau, Power BI, dbt, SQL Server, Sigma, Omni, and PostgreSQL. Cross-platform lineage tracks data movement from source databases through transformations to dashboards that display it. This results in a unified view of your data estate, improving understanding of data origins and usage.
Snowflake offers three context layers, each serving a distinct purpose:
- Snowflake Cortex Sense (public preview soon): Automatically infers context based on data estate usage, providing broad coverage initially without manual table and metric definition.
- Business Ontology: Governs business meaning, ensuring teams work with approved definitions and relationships.
- Snowflake Semantic Views: Executes metrics consistently, guaranteeing each metric is calculated identically.
These layers are not sequential; teams can choose the layer addressing their most pressing need and integrate additional layers as they grow. Some may start with glossary needs, while others may prioritize executable metrics. Regardless of starting point, these layers provide a single, consistent, and authoritative business context for all agents, applications, and individuals.
Automate Context with Cortex Sense
Cortex Sense is a natural starting point for many teams because it handles significant background work. It gathers context from everyday activities like querying data, building dashboards, and running pipelines, transforming these activities into business understanding with less time spent on manual table and metric definition.
Cortex Sense leverages metadata and usage patterns already generated by teams, including query history, dashboards, and dbt models. It automatically updates its knowledge based on chosen refresh cadence as the business evolves. Teams define a business use case, connect relevant existing context, and refine it through feedback and evaluations, minimizing manual table modeling. Cortex Sense identifies potential knowledge conflicts and asks teams to choose the source of trust.
When ready to govern critical terms, Cortex Sense can create an initial draft in Business Ontology with proposed definitions, relationships, and data links, eliminating the need to start from scratch.