Excel Get Data From a Semantic Model: Connect and Analyze

How-to basics · Updated November 18, 2026

Yes. In Excel for Microsoft 365 on Windows, connect to a Power BI semantic model through Data > Get Data > From Power Platform > From Power BI, or choose Analyze in Excel in Power BI. Excel creates a connected PivotTable, not a regular table containing a downloaded copy of the model.

What connecting to a semantic model means

A semantic model is the curated data layer behind a Power BI report. It can contain tables and relationships, plus measures, hierarchies, formatting, and security rules that give business data consistent meaning. Microsoft previously called Power BI semantic models datasets, so older instructions and menus may still use the word dataset.

When Excel connects to a Power BI model, it sends queries to that model as you build or refresh a PivotTable. The model calculates the requested totals using its existing definitions. For example, a sales model might include a measure called Net Sales that accounts for discounts. An Excel PivotTable using that measure can report it by region and quarter without rebuilding the calculation in a worksheet formula.

This is different from importing a CSV or database table into a sheet. The connection does not normally place all the model’s source rows in cells. Instead, Excel presents the fields and measures the model allows you to use. That distinction matters when you need to analyze approved business metrics, rather than extract and reshape raw source data.

Two ways to connect from Excel or Power BI

You can start in Excel or begin from the Power BI service. Both routes connect the workbook to a model; the labels and availability can vary with your Microsoft 365 build and organization settings.

  • Start in Excel: Use Data > Get Data > From Power Platform > From Power BI, then choose an available semantic model. Excel opens a connection and provides a PivotTable for analysis.
  • Start in Power BI: In the Power BI service, open the workspace or model options and select Analyze in Excel. The service provides a connection file, often an .odc file. Open it in the supported Excel desktop app, sign in if asked, and save the resulting workbook.

The Excel route is convenient when you already know which model you need. Analyze in Excel is useful when you are browsing a workspace, because it starts from the model’s Power BI location. If neither option appears, the likely issue is access, licensing, or a feature that your administrator has restricted—not a missing worksheet formula.

Connect and build a PivotTable

These steps describe Excel for Microsoft 365 on Windows. The exact picker may look different as Microsoft updates the interface, and an organization may show only models that you are permitted to use.

  1. Open a workbook and select Data > Get Data > From Power Platform > From Power BI.
  2. Sign in with the work or school account that has access to the Power BI content. Choose the workspace or search for the model you need.
  3. Select the model and choose the option to insert or create a PivotTable. If prompted, choose where Excel should place it.
  4. In the PivotTable Fields pane, add dimensions such as Date, Product, or Region to Rows, Columns, or Filters. Add a numeric measure, such as Net Sales, to Values.
  5. Check the result against the model’s intended definitions. Save the workbook so the connection and PivotTable layout are available the next time you open it.

For example, place Product Category in Rows, Fiscal Year in Columns, and Net Sales in Values. The PivotTable asks the model for sales by category and year. It does not need an Excel formula for each category, nor does it necessarily expose every underlying transaction. Use model measures where available: a measure may encode rules, such as excluding cancelled orders, that a simple sum of a source column would miss.

On a Mac, Power BI connection commands and Analyze in Excel support can differ by Excel version and organization configuration. If the Power BI option is absent, create the connection from the Power BI service or use a supported Windows desktop version to establish it. Before relying on a workbook across platforms, confirm that the target Excel app can open and refresh that connection.

What Excel can and cannot do with the model

A connected PivotTable lets you arrange fields, filter results, drill into summaries where the model permits it, and create Excel-based charts or reports. You can use model measures and dimensions without copying their definitions into the workbook. This makes Excel a flexible front end for questions such as “How did margin vary by product and month?” while Power BI remains the place where the shared model is maintained.

The connection does not give you permission to edit the model’s relationships, source queries, or measures. Those changes belong in the model’s authoring environment, such as Power BI Desktop, and require suitable editing access. Nor should you expect the Fields pane to include every source column: model authors can hide fields, restrict detail, or expose only governed measures.

Excel can also use cube functions with an OLAP connection for cell-by-cell reports, but that is a more specialized layout than a PivotTable. Most readers should start with a PivotTable. A connection to a semantic model is also not the same thing as Power Query importing an ordinary table: the model connection is designed to query governed model objects, not to turn the entire model into a freely editable Excel data table.

Refresh, credentials, and workbook sharing

A connected workbook gets results from the model when it queries or refreshes the connection. To update a PivotTable, select a cell in it and use PivotTable Analyze > Refresh, or use Data > Refresh All to refresh workbook connections. Refreshing asks the model for updated results; it does not itself refresh the model’s underlying data. If the model has not been refreshed by its owner, Excel cannot make its source data newer.

People who open a shared workbook may be asked to sign in with their own organizational accounts. Their ability to see data depends on their Power BI permissions and any row-level or object-level security configured for the model. A workbook’s author should not assume that recipients inherit the author’s access simply because they can open the Excel file.

For reliable sharing, include the model’s name and workspace in instructions, and tell recipients to request access through the normal organizational process. Avoid distributing screenshots or exported values as a substitute for access controls. If a workbook is stored in a shared location, test that intended recipients can authenticate and refresh it using their own accounts and supported Excel apps.

Choose the right method for the task

Use a semantic-model connection when you need the organization’s approved calculations, need to explore a model interactively, or want several Excel reports to rely on the same business definitions. A connected PivotTable is especially suitable for summaries that change as users move fields, apply filters, or select different date ranges.

Use another approach when the task is fundamentally different:

  • Need a static snapshot? Ask whether the report or model owner can provide an approved export. A snapshot is convenient for offline review but will not stay connected or update automatically.
  • Need to clean or combine raw source rows? Use an approved source such as a database or dataflow with Power Query, if you have access. Do not assume that connecting to the semantic model exposes its original source tables.
  • Need to change definitions or relationships? Work with the model owner or use the model-authoring tools with the necessary permissions.
  • Need only a chart or dashboard for a broad audience? A Power BI report may be easier to distribute and govern than a workbook that each recipient must refresh.

The practical trade-off is control versus flexibility. The model connection preserves shared definitions and security, while a copied dataset can be easier to manipulate but may become stale or produce conflicting calculations. Choose based on whether you need governed analysis or independent data preparation.

Common connection problems

If the model does not appear in the picker, first confirm that you are signed in with the correct work account and that the model is in a workspace you can access. You may also need Build permission on the semantic model to create your own analysis. Ask the owner or Power BI administrator to check access rather than trying to work around it with someone else’s credentials.

If the connection opens but refresh fails, check that you are online, signed in, and able to open the model in the Power BI service. A changed password, removed permission, unavailable service, or organizational policy can interrupt access. Some workbooks also depend on a particular desktop Excel version or connection configuration, so test the original connection before rebuilding the PivotTable.

If numbers differ from a Power BI visual, compare the filters, date range, aggregation, and measure used in both places. The Excel PivotTable may have a different filter selection, or the visual may use a measure that is not the field you dragged into Values. Check whether the model applies row-level security to your account. A difference does not automatically mean Excel has calculated the model incorrectly.

If the result is blank or a field is missing, check whether the model author hid the field, limited the available detail, or configured a measure that requires particular filters. Contact the owner with the model name, the field or measure, and the steps that produced the issue. This is more useful than reporting only that “Excel is broken.”

Frequently Asked Questions

Can Excel import a Power BI semantic model as a normal table?
Typically, the Excel connection creates a connected PivotTable rather than importing all model rows into worksheet cells. For raw rows, use an approved source or export route and confirm that you have permission.

Do I need Power BI Build permission?
Often, yes, to create your own analysis from a model. You also need access to the workspace or model, and your organization’s license and capacity setup can affect availability. Ask the model owner or administrator to confirm the requirements for your account.

Does Excel refresh the data in the semantic model?
No. Excel refreshes its connection and requests results from the model. The model owner or its configured refresh process is responsible for updating the model’s source data.

Can I use measures created in Power BI?
Yes. Measures exposed by the model can be used in a connected PivotTable, subject to your access and the model’s design. Using a shared measure helps keep Excel results consistent with reports that use the same model.

Why can a colleague see different values in the same workbook?
They may have different filters, use a different account, or be subject to row-level security. Compare the PivotTable filters and ask the model owner to verify each user’s access before concluding that the workbook is inconsistent.

Related

Try it: the Date Difference Calculator: days, months, years and weekdays between two dates, with the DATEDIF formulas.

More in How-to basics

Sources: Function behaviour and menu paths: Microsoft Support documentation for Excel. Method & sources

Spotted an error in this guide? Let us know.