Alation has launched Semantic Model Mastering in its Data Products Marketplace, a feature designed to help companies catalogue and govern semantic models across multiple data platforms.
The product lets organisations bring semantic models from different systems into a central layer, manage them as data products, and synchronise them back to source platforms. It is available immediately through YAML upload, with broader support for the Snowflake connector and synchronisation functions due later.
Semantic models define the business terms, dimensions, and metrics used by analytics tools and artificial intelligence systems. As companies adopt multiple cloud data and business intelligence platforms, those definitions can become fragmented across separate semantic layers.
That problem has grown as AI tools rely more heavily on underlying data definitions. When semantic models are inconsistent or outdated across platforms, organisations risk conflicting outputs and weak governance over the business context used by analysts and AI agents.
Platform-specific governance tools do not solve that issue for businesses operating across several environments. Alation's approach is to create a central mastered version of a semantic model, with ownership, approval workflows, version control, and quality standards managed in one place.
Cross-platform push
Under the new feature, semantic views from Snowflake can be ingested through an expanded connector, while definitions from Power BI and Tableau can also be brought into the marketplace. Customers can also create a data product for Databricks by uploading a metric view YAML file, while synchronisation back to Databricks is planned for a later stage.
Each imported model becomes a governed data product aligned with the Open Semantic Interchange standard, an emerging format for sharing semantic models between systems. Alation is a launch partner with Snowflake for the standard.
The product is positioned as a way to extend governance beyond the limits of a single platform. Instead of relying on separate controls inside each vendor environment, data stewards and data product owners can approve, enrich, and manage definitions centrally before distributing them back into operational systems.
The model echoes master data management architecture, where companies maintain one approved record for entities such as customers or products and distribute it to multiple applications. Alation is applying that idea to semantic definitions used in analytics and AI.
AI context
The launch reflects growing concern that data governance programmes do not yet fully cover the semantic layer used by AI services. Businesses may have controls around raw data and reporting assets, but the business logic that connects data to common terms and metrics often remains scattered.
By centralising those definitions, Alation argues that organisations can reduce manual reconciliation work and provide a clearer source of truth for systems that consume semantic models. Several enterprise customers in financial services, technology, and manufacturing had separately asked for a way to manage semantic models centrally and send governed definitions back to their data platforms.
The feature also adds business context to semantic models beyond the technical metadata held in a single source platform. That context can then move with the model when it is synchronised back to the systems where users and AI applications interact with data.
In practice, analysts or AI tools can continue working inside familiar platforms while relying on definitions that have been reviewed and governed centrally. Users do not need to work directly inside Alation's software for those governed models to be used downstream.
Marketplace extension
Semantic Model Mastering extends Alation's existing Data Products Marketplace, which serves as a governance layer for approved data assets. With the addition of semantic models, the marketplace now covers a broader set of information used to define and interpret enterprise data.
The move also underlines how software suppliers in the data management market are adapting their products to meet AI governance demands. While much of the industry focus has centred on model governance and data quality, semantic consistency is emerging as another area of concern as organisations deploy AI across multiple business systems.
Alation said the aim is to replace fragmented management of semantic definitions with a governed process that spans several platforms and maintains one approved version of key business terms and metrics.
The company said: "When a definition changes, it changes in one place and propagates to consuming systems through a controlled sync."