Guide · Pattern 00 · Last reviewed: September 2026

One workspace

Everything lives in a single workspace, and there is no lifecycle at all.

This is the pattern before any lifecycle exists. It is worth naming so you can tell when you have moved past it. Working this way is fine while you are the only person using the reports. As soon as somebody else relies on them, it is not.

Power BI Desktop, holding The file on your own disk, then Publish, then One workspace, holding Semantic model and Report and The audience, which is you. POWER BI DESKTOPThe file on yourown diskPublishONE WORKSPACESemantic modelReportThe audience, which is youData NovaSave SVG

Who it fits

  • One person building, and nobody else consuming
  • A proof of concept, a spike, a "can Fabric even do this" week
  • Content whose audience is you

How it is set up

There is not much to it, which is the point. One Pro workspace, or one workspace on a trial capacity if you need lakehouses and notebooks. Publish from Desktop, refresh on a schedule, look at it.

The Workspace type panel in Fabric workspace settings, reading Current workspace type: Power BI Pro, with an Edit link beside it.
Workspace settings, Workspace type. Edit is how you put the workspace on a trial capacity, and how you hand it back afterwards.
The Workspace type panel in Fabric workspace settings, reading Current workspace type: Power BI Pro, with an Edit link beside it.

Where the query lives

Power Query sits either inside the Power BI file, or in a Gen1 dataflow if you have moved the transformation upstream. On Pro that is the only dataflow you get, because Gen2 needs a capacity, so it is available on a trial and gone when the trial ends. What else Gen1 costs you is on the Pro only workspace pattern.

Parameters work in both. In a dataflow they are set per dataflow, so each one carries its own.

One habit that pays off later. Put your connection details in Power Query parameters from day one, instead of hard-coding a server and a database into each query. Every pattern above this one uses parameters as the lever that changes between environments, and retrofitting them into a mature model is tedious. Starting with them costs nothing.
// Two parameters, then one source that reads them.
Server = "sql-dev.example.com" meta [IsParameterQuery=true, Type="Text"],
Database = "Sales_Dev"         meta [IsParameterQuery=true, Type="Text"],

let
    Source = Sql.Database(Server, Database),
    Sales  = Source{[Schema="dbo", Item="FactSales"]}[Data]
in
    Sales

The whole of it, drawn. There is no promotion step, because there is nowhere to promote to.

You, in Power BI Desktop, holding Author the model and report and Parameters from day one, then Publish. You are the only audience, so nothing is promoted anywhere, then One workspace, holding Publish and Set the credentials and Refresh on a schedule and Look at it. YOU, IN POWER BI DESKTOPAuthor the modeland reportParameters fromday onePublish. You are the only audience,so nothing is promoted anywhereONE WORKSPACEPublishSet thecredentialsRefresh on ascheduleLook at itData NovaSave SVG

What it costs

One Pro licence. Or nothing, on a 60 day trial capacity, which reports its SKU as FTL64 and behaves as an F64 for those 60 days. Fine for proving something, never for a cutover date.

What breaks

Nothing, until someone else looks at it. Then everything breaks at once: a half-finished change in front of a director, a refresh you broke at four o'clock on a Friday, and no way to answer what it looked like last week.

When this breaks down

Someone who is not you opens the report and makes a decision from it. That is the line. Not team size, not data volume. The moment anyone depends on your reports, the workspace they read from should not be the workspace you build in.

When you cross that line, three Pro workspaces with manual promotion is the next step, and it needs no capacity to stand up.