Unity Catalog Pages: a governed home for your business knowledge in Genie Ontology
Databricks introduces Unity Catalog Pages to centralize and govern business definitions for AI agents, ensuring consistent and accurate interpretations of key terms across enterprises.
Databricks has launched Unity Catalog Pages as a governed layer within Genie Ontology, enabling organizations to define and standardize critical business terms such as 'active customer' or 'completed trip' in a single, authoritative location. This addresses inconsistencies where the same term may be interpreted differently across teams, reducing reliance on scattered knowledge sources like Slack or spreadsheets. By providing a centralized hub, Pages ensure AI agents interpret queries accurately rather than making assumptions based on inference alone.
Unity Catalog Pages integrate with Genie Ontology’s existing context layers, including metric views, domains, and certification, to form a cohesive picture of an organization’s business logic. Each Page includes structured fields like owner, synonyms, and description, alongside rich content such as links, images, and references to Unity Catalog assets. Pages are organized within Discover under relevant domains, aligning business context directly with the data estate it describes for easier navigation and verification.
Genie Code, Databricks’ AI assistant, can automate the creation of Pages by extracting business knowledge from sources like Confluence, Slack, or Google Docs, significantly reducing the manual effort required. Users can bulk import concepts by pointing Genie Code to key documents, which then generates atomic Pages for each term. This feature accelerates the process of documenting organizational terminology, turning weeks of manual work into minutes of automated curation.
Unity Catalog Pages are now available in Beta as part of Unity Catalog semantics, offering enterprises a collaborative platform to curate and govern definitions that AI agents rely on. Each Page cites its sources and related assets, ensuring transparency and traceability for users reviewing AI-generated answers. The integration with Genie One and other agents provides consistent, trusted context, bridging business logic with the underlying data infrastructure.