Skip to main content
Back

Databricks' Genie Code Now Migrates Entire Tableau and Power BI Dashboards at Once

Databricks moves bulk dashboard migration with Genie Code to general availability, letting teams convert entire Tableau and Power BI folders in one command.

  • Databricks made bulk dashboard migration with Genie Code generally available.
  • One /importBI command processes a whole Unity Catalog folder of reports.
  • Supports .twb, .twbx, .tds, .tdsx, .pbit and .tmdl files from Tableau and Power BI.
  • Agent reuses existing Unity Catalog tables, views and metric views where possible.
  • Tableau imports create local metric views; Power BI imports create Unity Catalog ones.
  • Screenshot-based reconciliation lets Genie Code flag KPI mismatches against original dashboards.

Databricks has made bulk dashboard migration generally available in Genie Code. The /importBI skill now accepts a directory of Tableau or Power BI exports, replacing the previous one-report-per-run workflow. It reads the source files, creates Databricks AI/BI Dashboards, and maps generated assets to a specified Unity Catalog catalog and schema.

Bulk execution reduces the work required to move large report inventories while preserving separate progress, cancellation, and validation controls for each dashboard. The agent can also reuse governed data definitions across reports, limiting the duplicate metrics that often accumulate during a migration.

Point one command at the folder

  1. Export the Tableau or Power BI files.
  2. Upload them to a Unity Catalog volume.
  3. Open a draft AI/BI Dashboard.
  4. Run /importBI with the directory path.
/importBI <path/to/volume/directory>

Using a directory path starts migrations for every supported report in that folder. A natural-language instruction produces the same result:

Migrate every Power BI report in /Volumes/finance/bi/reports to AI/BI Dashboards.

A prompt can also name individual files when only part of a directory should be migrated.

Source Supported files
Tableau .twb, .twbx, .tds, .tdsx
Power BI .pbit, .tmdl

The agent automatically pairs Tableau workbooks with companion data sources. It asks the user to choose only when multiple files could satisfy the same reference.

Genie Code listing Tableau workbooks found in a volume and requesting a target catalog and schema
Genie Code identifies source workbooks before requesting the Unity Catalog destination.

Parallel jobs share governed assets

After the user confirms the target catalog and schema, Genie Code queues the reports and processes them in parallel batches. Each report receives its own background job and planning sequence:

  1. Parse and assess the source workbook.
  2. Identify measures, dimensions, and other semantic definitions.
  3. Reuse or create metric views.
  4. Build the dashboard pages and visualizations.
  5. Validate the generated dashboard.

Before creating an asset, the agent checks for suitable Unity Catalog tables, views, and metric views. Reusing those objects reduces near-duplicate definitions and gives dashboards, Genie agents, and notebooks access to the same governed measures.

Bulk import status card showing queued, running, and completed dashboard migrations
Each report retains its own status and cancellation control.

The status card labels each migration as Queued, Running, or Completed. Users can cancel one report or the entire batch, close the browser tab, and return to the conversation while the background jobs continue.

Metric-view defaults differ

Metric views define governed, reusable calculations over source tables, such as revenue, margin, or customer count. Their default scope depends on the source platform:

Source Default output How to change it
Tableau Dashboard-scoped local metric views Request Unity Catalog metric views in the prompt, or promote the local views later.
Power BI Unity Catalog metric views No additional instruction is required.

A read-only Model Context Protocol connection can give the agent access to the live source report during validation. Databricks recommends this setup for more accurate comparisons than exported files alone can provide. Enabling the connection requires coordination with the Databricks account team.

A completed run still needs proof

Genie Code supports screenshot-based reconciliation after migration. A user can upload an image of the original dashboard with specific filters applied, then ask the agent to compare every KPI and chart total, list mismatches, and show the metric view or SQL query behind each result.

  • Completion status: Completed indicates that the job finished. Each dashboard still requires comparison with its source report.
  • Missing tables: When a referenced source table is unavailable in Databricks, the agent may select the closest available table. Configure data access through source connections before migration and verify every substitution.
  • Batch limits: File-count and file-size limits apply and may change. Check the current limits before planning a large migration.
  • Workspace settings: Partner-powered AI features must be enabled at both the Databricks account and workspace levels.