Skip to content

Configure reports⚓︎

After enabling Portal AI in your environment, you need to activate it on each report individually. This involves enabling the AI toggle and synchronizing the report's semantic model schema.

Prerequisites⚓︎

Before configuring a report for Portal AI:

  • Portal AI must be enabled in your environment settings.
  • The report must already be saved in PowerBI Portal, so you cannot enable AI on a new, unsaved report.

Supported Report Types⚓︎

Portal AI works with the following report types:

Report Type Supported
Report Page Yes
Visual Yes
Dashboard Tile Yes
Paginated Report Yes
Dashboard (full) No
Manual No

If the report type is not supported, the AI toggle will be disabled and a message will be shown.

Step 1: Open the Report Editor⚓︎

  1. In the PowerBI Portal navigation menu, go to Reports.
  2. Click on the report you want to configure, or select Edit from its actions menu.

Step 2: Enable Portal AI for the Report⚓︎

  1. In the report editor, expand the Advanced Details panel on the right-hand side.

  2. Select the Portal AI tab (the fourth tab, visible only when Portal AI is enabled for the environment).

  3. Toggle Enable Portal AI to on.

Note

If the toggle is disabled, check that:

  • The report has been saved at least once. A message "Save the report first to enable Portal AI" is shown for unsaved reports.
  • The report type is supported. A message "Portal AI is not available for this report type" is shown for unsupported types.

The toggle and the schema discovery section only become available after saving the report for the first time.

Step 3: Synchronize the Semantic Model Schema⚓︎

When you enable Portal AI for the first time on a report, the system automatically resolves the Semantic Model ID from Power BI. You'll briefly see a "Resolving Semantic Model ID from Power BI..." message.

Once the Semantic Model ID is resolved, the Semantic Model Schema section appears:

This section displays:

  • Semantic Model ID: The unique identifier of the Power BI semantic model linked to this report.
  • Schema Status: The current state of schema synchronization:
    • Not discovered: The schema has not been fetched yet.
    • Discovering...: Schema discovery is in progress.
    • Synchronized: The schema is up to date and ready for use.
    • Failed: Schema discovery encountered an error (an error message will be shown).
  • Last update: When the schema was last successfully synchronized.
  • Get/Update Schema button: Click to start or refresh the schema discovery process.

Running Schema Discovery⚓︎

  1. Click the Get/Update Schema button.
  2. The status will change to Discovering... while the system reads the data structure from Power BI.
  3. Wait for the status to change to Synchronized.

Info

Schema discovery typically takes between 10 seconds and 2 minutes depending on the size and complexity of the semantic model. The page automatically checks for status updates while discovery is in progress.

  1. Once the status shows Synchronized, save the report by clicking Save.

When to Refresh the Schema⚓︎

You should refresh the schema whenever significant changes are made to the Power BI semantic model, such as:

  • Adding or removing tables or columns.
  • Adding, modifying, or removing measures.
  • Changing relationships between tables.
  • Renaming tables or columns.

Portal AI uses the schema to understand your data structure. If the schema is outdated, the AI may not be aware of new fields or may reference columns that no longer exist.

Troubleshooting⚓︎

Schema discovery failed⚓︎

If schema discovery fails, check the error message displayed below the status. Common causes include:

  • Power BI permissions not configured: Ensure all four settings are enabled in the admin portal in Power BI as described in Enabling Portal AI.
  • Permissions not yet propagated: After changing settings in the admin portal in Power BI, wait up to 15 minutes and try again.
  • Service principal not added to workspace: The PowerBI Portal service principal (or its security group) must have access to the Power BI workspace containing the report.

The semantic model is too large to be used with Portal AI⚓︎

Portal AI validates the semantic model schema size before it can be used. If this error appears, the model contains more metadata than Portal AI can process. Consider hiding unnecessary tables, columns, and measures in Power BI Desktop, then refresh the schema again.

Portal AI tab is not visible⚓︎

The Portal AI tab only appears in the report editor when Portal AI is enabled at the environment level. Go to the Portal AI tab in Settings to enable it first.

Reports with Row-Level Security (RLS)⚓︎

Portal AI supports reports that use Row-Level Security (RLS). When a user asks a question, Portal AI respects the RLS rules configured for that user's report access in PowerBI Portal. This means answers are limited to the same data the user can see when viewing the report normally.

RLS Configuration⚓︎

Portal AI uses the RLS username and roles configured in PowerBI Portal for each user or group. Keep these values aligned with the RLS configuration in Power BI so Portal AI can apply the correct data restrictions.

Requirements⚓︎

  • Premium or Embedded capacity is required: Reports with RLS require the report's workspace to be assigned to a Premium or Embedded capacity. If no capacity is available, Portal AI will show an error message: "This report uses Row-Level Security which requires Premium capacity."
  • RLS roles must match: The role names configured in PowerBI Portal must match the role names defined in the Power BI semantic model (case-sensitive). If there is a mismatch, Portal AI will show an error instead of returning an answer.

Tip

If you encounter errors related to RLS roles, verify that the role names in PowerBI Portal match exactly with those defined in Power BI Desktop (or the Tabular Model). Role names are case-sensitive.