Microsoft Fabric bundles ingestion, storage, warehousing and reporting into one analytics platform. Here is what it actually does, why Australian agencies are looking at it, and what it will not fix.
Microsoft Fabric has been in market long enough that most Australian agencies have heard the name and rather fewer can explain what it does. The marketing does not help. Here is a plain description of what Fabric is, what problem it solves, and where it does not help.
What Fabric actually is
Fabric is a single analytics platform that bundles together capabilities Microsoft previously sold separately. Data ingestion, storage, transformation, warehousing, real time analytics, data science and Power BI reporting all sit inside one product with one billing model and one permission model.
The important architectural idea is OneLake. Instead of every tool keeping its own copy of the data, Fabric stores everything once in an open format and points the tools at it. If you have ever watched a data warehouse project spend six months building pipelines to move the same records between four systems, you will understand why that matters.
The parts you will hear named
- OneLake: the shared storage layer that everything else reads from
- Data Factory: pipelines for getting data in from source systems
- Lakehouse and Warehouse: two ways of organising the data once it lands, depending on whether your team thinks in files or in SQL
- Real Time Intelligence: for streaming data such as sensor or telemetry feeds
- Power BI: the reporting layer, now sitting natively on the same storage
Why agencies are looking at it
Three reasons come up repeatedly. The first is consolidation: agencies running a mix of on premises SQL Server, Azure Synapse, standalone Power BI and a few departmental Access databases can collapse a lot of that into one estate. The second is licensing simplicity, because Fabric capacity is purchased as a single pool rather than per service. The third is that Copilot and other AI features are being built into Fabric first, so agencies planning AI work find the data foundation question arrives whether they wanted it or not.
What Fabric does not solve
Fabric does not fix data quality. If your source systems disagree about what a client record is, Fabric will faithfully surface that disagreement at higher speed and in a nicer dashboard. It does not remove the need for data governance, and in a regulated environment it arguably raises the stakes, because making data easier to combine also makes it easier to combine data that should have stayed separate.
It also is not free. Capacity is reserved and billed whether or not you use it, so an agency that provisions generously and then runs one report a week will notice.
A sensible starting point
Pick one reporting problem that is genuinely painful and currently involves a person manually joining spreadsheets. Build that end to end in Fabric, measure the capacity it consumes, and use the result to size the wider programme. That gives you a real number to plan against rather than a vendor estimate, and it gives your team hands on experience before the architecture decisions become expensive to reverse.