You are here:
Agentic Data Transforms in Tableau Next (Beta)
Prepare, clean, and transform your data by using the Tableau agent. You use natural language prompts to generate SQL, preview your changes, and save your work as batch data transforms.
Required Editions
Note: Agentic Data Transforms is a pilot or beta service that is subject to the Beta Services Terms at Agreements - Salesforce.com or a written Unified Pilot Agreement if executed by Customer, and applicable terms in the Product Terms Directory. Use of this pilot or beta service is at the Customer's sole discretion.
| User Permissions Needed | |
|---|---|
| To enable Agentic Data Transforms | Tableau Unmetered Admin or Tableau Next Admin permission set |
| To configure Agentforce for Analytics | System Administrator |
| To use Agentic Data Transforms | Tableau Unmetered Platform Analyst or Tableau Next Platform Analyst |
Prerequisites
- To enable this feature, see Enable Agentic Data Transforms in Tableau Next (Beta). You can’t disable the agentic data prep feature after it’s enabled.
- After enabling the preference, add the Data Transform subagent to your existing Tableau agent, or create an agent and add the Data Transform subagent.
- Enable the feature separately in each personal org.
To start transforming your data, on your DLO, click the agent icon in the top navigation bar to open the agent panel.
Use the agent to clean data and preview changes. For example, you can:
- Filter data by including or excluding specific records.
- Perform inline edits of domain values and change data types.
- Split columns into meaningful fields.
- Make iterative, SQL-based edits to the same output data object.
- View a profile of your cleaned data in a responsive visual environment.
Save your agent-assisted changes as a new batch data transform (BDT). When saving for the first time, the agent asks you to provide a data object name and a transform name. The agent automatically sets defaults for the Definition name, Workspace, Dataspace, Primary key, Key qualifier, and Target object category. After confirming the details, the agent creates the output DLO, assigns it to a dataspace, creates the BDT, adds them to your workspace, and automatically initiates a full run. For DLOs that are already BDT outputs, the agent saves iterative changes by updating the BDT and output DLO using the underlying DCSQL. Then the agent runs the BDT.
You can also manage batch data transforms by having the agent:
- List all the BDTs in your org or workspace.
- Fetch the details of a specific BDT.
- Run a BDT.
- Schedule a BDT.
- Refresh the history of a BDT.
Limitations
- Agentic data prep isn’t available in Data 360 companion orgs.
- After you enable agentic edit capabilities on a DLO, manual edit capabilities aren’t visible to users even if they have the permission for manual edits.
- The agent can only clean and transform data within a single DLO. Instructions that require combining or joining multiple data objects aren’t yet supported.
- The agent edits only the data object that is currently open. To clean up a different data object, go to it first.
