Fivetran outlines an open, AI ready data strategy for Microsoft Fabric
Fivetran and dbt Labs have outlined an open data strategy for organizations building analytics and AI systems on Microsoft Fabric. The approach combines Azure storage, portable dbt business logic and governed context for AI agents.
- Fivetran Managed Data Lake writes data from 700+ sources to Azure Data Lake Storage.
- Open Apache Iceberg tables can be discovered through the OneLake Catalog.
- The dbt MCP Server exposes governed models, metrics, lineage and tests to AI agents.
- dbt Job in Microsoft Fabric is generally available.
- Gerflor is modernizing its data platform across 40 subsidiaries and more than 400 applications.
An open foundation for Fabric workloads
In a November 9, 2026 blog post about the Microsoft Fabric Community Event in Barcelona, Fivetran described five developments it says can help teams build on Fabric while keeping their data and future technology choices open.
Fivetran Managed Data Lake writes data from 700+ sources directly into Azure Data Lake Storage as open Apache Iceberg tables. Delta Lake metadata is available as an option. Through integration with the OneLake Catalog, those tables can be discovered and governed inside Fabric.
Fivetran says Fabric, other Azure services and any engine that reads Iceberg or Delta can query the same copy. This approach is intended to avoid re ingesting data and creating duplicate pipelines as organizations add teams, engines and workloads.
Portable business logic and AI context
Fivetran also highlighted Microsoft’s dbt adapters for Fabric Warehouse and Fabric Lakehouse. Existing dbt projects can target Fabric without a rewrite, allowing teams to define transformations, metrics and tests once in SQL and dbt and use them across compute environments.
The dbt MCP Server is now generally available, according to Fivetran. It exposes dbt models, metrics, lineage and tests to AI agents through the Model Context Protocol, an open standard. The company says this allows agents, including copilots built on Microsoft’s AI stack, to use the same governed definitions as analysts instead of interpreting raw tables independently.
At FabCon Barcelona, Microsoft also announced that dbt Job in Microsoft Fabric is generally available. The release supports building, scheduling and running dbt transformations inside Fabric. It adds broader enterprise support for dbt at scale, dbt logs in OneLake, more Fabric adapters including Lakehouse, and variable library support in deployment pipelines.
Gerflor puts the model into practice
French flooring manufacturer Gerflor shared its Fabric modernization work at the event. The company is replacing a 15 year old on premises data platform while supporting 40 subsidiaries and more than 400 applications. It adds an average of 4 new subsidiaries each year.
Fivetran says Gerflor’s fragmented data and inconsistent business definitions made information difficult to reuse and trust. Month end consolidation required 10 employees and roughly 5 days of manual preparation. The move to Microsoft SaaS also created demand for consistent commercial reporting across CRM and Business Central for more than 1,700 users.
Gerflor is building its cloud data foundation on Azure Data Lake Storage Gen2 and Microsoft Fabric. Fivetran handles change data capture and data movement, while dbt standardizes transformations and business logic. Gerflor plans to expand its Fivetran connectors, scale Fabric capacity, standardize more transformations with dbt and support AI services.
Fivetran presents this combination as Open Data Infrastructure: one copy of the data, trusted definitions and the freedom to choose engines and tools as analytics and AI requirements grow.
