How to Get Started with Microsoft Fabric

Microsoft Fabric is Microsoft's unified analytics platform that brings data engineering, warehousing, real-time processing, and Power BI reporting into a single environment. As of April 2026, Copilot and AI capabilities are available on all paid F2-and-above SKUs, removing the previous F64 requirement that blocked smaller teams from using the AI features. This guide walks through setting up Fabric from a free trial to your first published report.

What Microsoft Fabric Actually Is

Microsoft Fabric launched in 2023 as Microsoft's answer to the fragmented analytics stack: separate tools for data ingestion, transformation, storage, querying, and visualization that rarely communicate cleanly with each other. Fabric consolidates those layers into six integrated workloads that share a single storage layer called OneLake.

The six workloads are Data Engineering (Spark-based notebooks and pipelines), Data Factory (no-code data integration), Data Science (ML experiments and model serving), Real-Time Intelligence (stream processing and event analytics), Data Warehouse (T-SQL queries against structured tables), and Power BI (dashboards and reports). Every workload reads from and writes to OneLake, so data you land in one workload is immediately accessible in another without copying or moving files.

For an ops manager or analyst who currently juggles Excel exports, a BI tool, and a shared database, Fabric can replace the entire chain.

Before You Start

You need one of the following: a Microsoft 365 work or school account, or a personal Microsoft account (Outlook, Hotmail). A paid Microsoft Fabric capacity (F2 or higher) is required for Copilot. The 60-day free trial activates an F64 capacity, so Copilot is fully available during the trial period.

You do not need to install any software. Fabric runs entirely in the browser at app.fabric.microsoft.com.

Step 1: Start Your Free Trial

Go to app.fabric.microsoft.com and sign in with your Microsoft account. If your organization has not yet provisioned Fabric, you will see a banner offering a 60-day free trial. Select "Start trial" in the account manager panel in the top-right corner.

The trial activates an F64 capacity assigned to your account. No credit card is required. Trial capacity is sufficient to run every exercise in this guide, including Copilot.

If your organization already has a Fabric capacity, your admin may need to assign it to you in the admin portal at admin.fabric.microsoft.com before you can use it.

Step 2: Create a Workspace

Workspaces are the organizational unit in Fabric, similar to a project folder. Every item you create (a lakehouse, a pipeline, a report) lives inside a workspace.

From the Fabric home page, select "Workspaces" in the left sidebar, then "New workspace." Give it a name, choose your trial capacity (or your assigned capacity), and click Apply.

Assign the workspace to a capacity rather than leaving it on the default shared capacity. Items in shared capacity do not have access to Copilot and have reduced compute limits. Once assigned, the workspace shows a diamond icon in the sidebar confirming capacity is attached.

Step 3: Ingest Your Data

Fabric offers two main paths for getting data in: Lakehouse for file-based and database sources, and Dataflow Gen2 for transformation-heavy pipelines. For most analysts starting out, Lakehouse is the right entry point.

Inside your workspace, select "New" and then "Lakehouse." Name it and click Create. The Lakehouse opens with three sections: Files, Tables, and a SQL analytics endpoint. Drop a CSV or Excel file directly into the Files section by selecting "Upload" from the toolbar, or connect a data source by selecting "Get data" at the top.

For connecting a live source such as SQL Server, SharePoint lists, or Azure Blob Storage, use the Dataflow Gen2 item instead. Select "New," then "Dataflow Gen2." The Power Query editor opens, where you can choose a data source, preview the data, apply transformations, and load the result into a Lakehouse table.

One practical tip: if you are working with CSVs, upload the file to Files, right-click it, and select "Load to Delta table." Fabric converts the file into a queryable table in a few seconds. You do not need to write any schema definitions or transformation code for simple uploads.

Step 4: Analyze With SQL or Copilot

Once data lands in a Lakehouse table, it is immediately queryable via the SQL analytics endpoint. Switch to the "SQL analytics endpoint" view inside the Lakehouse using the toggle at the top-right of the Lakehouse screen. This opens a SQL query editor connected to your tables.

Run a basic query to verify your data loaded correctly:

SELECT TOP 10 * FROM your_table_name;

For aggregations, filtering, and joins, standard T-SQL syntax works directly against your Lakehouse tables.

If your workspace has Copilot enabled (F2+ capacity or trial), a Copilot chat panel appears in the sidebar. Type a plain-English question such as "What are the top five products by revenue this quarter?" and Copilot generates the SQL, executes it, and shows the results. You can review the generated query before running it. This feature was previously restricted to F64 SKUs and is now available to all paid tiers as of April 2026.

Step 5: Build a Report in Power BI

From the workspace, select "New" and then "Report." Choose your Lakehouse dataset as the source. Power BI opens in the same browser tab, connected directly to your tables. There is no need to export data or set up a gateway.

Add a bar chart, line chart, or table by dragging fields from the data panel on the right. For an ops manager who needs a monthly revenue breakdown, drag the date field to the X axis and the revenue field to the Values section. Power BI handles the aggregation automatically.

Copilot in Power BI (available with the same F2+ capacity) can generate report pages from a text prompt. Select the Copilot button in the ribbon, type "Create a summary page showing revenue by region and product category," and Fabric builds the visual layout. You can review and adjust before publishing.

To share the report, select "Publish" and choose your workspace. Colleagues with a Power BI license can view it at app.powerbi.com without needing a Fabric capacity themselves.

What the April 2026 Copilot Expansion Means

Before the April 2026 update, Copilot in Microsoft Fabric required an F64 capacity, which represents 64 Fabric capacity units. The new threshold is F2, the lowest paid tier, which removes that barrier for small and mid-size teams.

In practice, a team of analysts running on a modest Fabric capacity can now use natural-language querying in the SQL endpoint, Copilot-generated report pages in Power BI, and Fabric data agents for automated data summarization. These were all gated to higher-tier customers until this change.

One caveat: Copilot consumption draws from the capacity's compute budget. Teams on F2 should monitor usage with the Fabric capacity metrics app and set alerts in the admin portal to avoid throttling during high-demand periods.

What to Explore Next

Fabric's Data Factory workload handles scheduled data refreshes and orchestration, replacing manual export scripts or simple ETL workflows. The Real-Time Intelligence workload connects to event streams from IoT devices, application logs, or financial feeds and runs continuous queries against them.

For teams that already use Azure Data Lake Storage, Fabric can connect to existing ADLS Gen2 accounts using OneLake shortcuts, making data accessible inside Fabric without moving or copying it.

If your team wants to skip infrastructure setup and go straight to insights from a file upload, VSLZ AI lets you upload a CSV or connect a data source and get instant charts and statistical analysis from a single prompt, with no pipeline configuration needed.

The next logical step after this guide is to build a Data Factory pipeline that refreshes your Lakehouse table on a daily schedule, then connect that table to a live Power BI report. That setup covers the full analytics loop from raw data to an auto-refreshing dashboard.