The release addresses a gap that has frustrated data teams since Fabric became a mainstream data platform: raw data pipelines can deliver large volumes of information to OneLake without the business context, lineage, or relationship metadata that analytical and AI workloads require to produce coherent, consistent answers.
Semarchy positions the workload as an alternative to hand-built semantic models, which carry a maintenance overhead whenever upstream data products change. The integration instead automatically synchronises governed business vocabulary, golden records, hierarchies, and data relationships from Semarchy’s master data management platform into Fabric. Updates propagate to downstream consumers without manual intervention, so the semantic layer stays aligned with current data as organisations evolve their products.
The practical scope covers three connected use cases. Power BI gains a semantic foundation built from certified master data rather than ad-hoc calculations embedded in individual reports, reducing the variation that arises when different analysts define the same metric independently. Copilot responses are grounded in verified business entities, lowering the risk that generative queries return contradictory results depending on which data asset they happen to reference. Fabric Data Agents, which navigate a customer’s data estate to answer natural-language questions, gain the context needed to interpret relationships between business objects accurately.
Organisations in financial services, manufacturing, and retail are listed among Semarchy’s existing customer base. These sectors have historically struggled with conflicting definitions of core entities such as customer, product, or supplier, a problem that multiplies when reporting tools and AI assistants draw on unresolved source data. The workload is designed for enterprises that have already standardised on Fabric but find their analytical and AI outputs still diverge because no single governed source of record underpins them.
The technical mechanism centres on publishing master data directly to OneLake and maintaining a live connection between the master data management layer and Fabric’s storage. Any change to a data product propagates automatically, removing the lag that can otherwise introduce temporary inconsistencies in dashboards or Copilot-generated summaries. Relationships, hierarchies, and lineage travel with the data rather than being reconstructed at query time.
The workload is framed as closing the gap between governance and consumption. The premise is that many enterprises have built capable pipelines into Fabric but have not yet attached the semantic and governance metadata that would let analytical tools, and increasingly AI agents, treat that data as trusted. Semarchy’s approach embeds that context at the source, in the master data layer, rather than asking each consuming team to build and maintain their own interpretation.
No independent performance benchmarks were included in the announcement materials. Semarchy said the Data Products workload is available now for Microsoft Fabric customers.
To stay across the latest in cloud, AI and enterprise tech analysis from Compare the Cloud, subscribe to our weekly newsletter at https://www.comparethecloud.net/newsletter
