Use the workflow to enable the Tableau Next Model Context Protocol (MCP) Server
Required Editions
| View supported editions. |
| User Permissions Needed | |
|---|---|
| To add MCP servers: | Tableau Next Admin |
To use Tableau Next MCP Server, your environment should meet these requirements:
- Feature Enablement: Tableau Agent with the Data Analysis subagent enabled in your Salesforce org, so it can use the query_semantic_data tool.
See Also
Step 1: Configure MCP on Salesforce Settings
Watch this 3-minute video to see how to enable the Tableau Next MCP Server. You can also follow the written steps.
Before you connect an AI agent, follow the instructions in the Salesforce Hosted MCP Servers developers site.
Step 2: Configure the MCP Client
Watch this 3-minute video or follow the written steps to connect Tableau Next MCP to Claude Desktop.
Add the MCP server name and URL and your External Client, such as Claude, Cursor, or ChatGPT. Refer to : Connecting an MCP Client for specific server names and URLs for the client you are configuring.
After connecting the server, verify the integration by asking a discovery question in your AI client. Use a prompt that targets your metadata to ensure the server can communicate with your Salesforce org.
Sample Prompt:
- "What semantic models are available in my Tableau Next environment?"
Expected Result:
- The AI agent lists your available semantic models using the metadata discovery tools provided by the server.
Sample prompt:
- Ask about your data from the AI client, or use Tableau Agent with the Data Analysis subagent in the Salesforce UI.
Excepted result:
- You should get the same answer to that question from both Tableau Agent with the Data Analysis subagent and the AI client.