Tableau Agent FAQ
Note: Einstein Copilot has been renamed Tableau Agent, as the platform expands to accommodate more AI agent functionality. Starting in October 2024, you'll see updates to page sections, field names, and other UI text throughout Tableau Prep, Tableau Catalog, and Tableau Cloud Web Authoring. Help content and Trailhead modules are also being updated to reflect these changes.
Find answers to commonly asked questions about using Tableau Agent in Tableau Cloud web authoring. To learn more about Tableau Agent in Tableau Prep or Tableau Catalog, see Create Calculated Fields with Tableau Agent(Link opens in a new window) or Describe projects, workbooks, or data sources(Link opens in a new window).
General Information
What is Tableau Agent?
Tableau Agent is a generative AI feature in Tableau that helps you explore data, create visualizations, and uncover insights using a conversational assistant. It allows you to interact with your data through natural language in a web authoring environment.
Is there a free trial?
Yes. You can sign up for a Tableau Cloud free trial which now includes Tableau Agent features. Simply fill out this form, or contact your Account Executive. You can currently try Tableau Agent in Tableau Cloud web authoring and Tableau Prep on the web as part of a limited trial.
The trial gives you full access to Tableau Agent functionality but there are some restrictions on the overall experience.
You can only trial in a new site, not add a trial of Tableau Agent to your existing production site.
Tableau Agent in Tableau Catalog isn't included in the trial.
Tableau Agent is free to use during the trial but limits may apply to the number of requests you can make per hour.
How do I access Tableau Agent?
You can find Tableau Agent in the web authoring environment of a Tableau Cloud site. Look for the Tableau Agent icon in the toolbar next to ShowMe. The icon is only active when Tableau AI is turned on for web authoring, and you are on a worksheet tab and connected to a data source. You must also be assigned the role of Creator or Explorer.
Why is the Tableau Agent button grayed out?
To use Tableau Agent, the following must be true:
You must be connected to a Tableau Cloud site
You must be on a worksheet
You must be connected to a data source
Tableau AI must be enabled for your site and the Tableau Web Authoring: Builds visualizations or completes tasks for you using a conversational AI check box is selected.
Usage and Capabilities
Can I use Tableau Agent on any type of data?
Tableau Agent is recommended for use on data extract or file-based data for improved performance, but Tableau Agent also works with live connections, or uploaded files with .hyper, .csv, .txt, and .xlsx file types. Cubes are not supported. Also, if you're using data blending, Tableau Agent can only be used with the primary data source.
Only data in English is formally supported. Tableau Agent responses are English only.
What can I ask Tableau Agent to do?
Tableau Agent has two top level capabilities - viz generation and working with calculations.
Tableau Agent can help you build visualizations, choose the best chart types for your data, perform date-based analysis, and manage data filtering and sorting. For example, you can ask it to "show sales data for a specific region" or "compare sales to profit". For more information, see Build and change a viz(Link opens in a new window).
Tableau Agent can also create calculations from natural language. Using the same conversation pane that you would use for generating a viz, you can ask Tableau Agent to create, update, or explain a calculation. For more information, see Create calculations(Link opens in a new window).
Are there limitations to what Tableau Agent can do?
Yes, Tableau Agent has some limitations. For example, it can't pick a data source for you, do data modeling, build dashboards, suggest relevant content or answer data lineage questions, answer consulting questions (such as “How should I analyze my data?”), or create interactivity with elements like filter controls and parameters. It's primarily aimed at helping novice users and those exploring ad-hoc data questions. For more information, see Tableau Agent limitations(Link opens in a new window).
What is the best way to use Tableau Agent effectively?
Start by working with clean data
Make sure field names are descriptive and unique
When you start an Tableau Agent session, if you see similarly named fields in a data source, hide the ones that aren't relevant. Tableau Agent won't ground its answers on hidden fields
Specify the aggregation function for a measure if you know what you want
Make sure fields have the correct field role (measure or dimension) and data type (location, string, boolean, etc.)
Like traditional Tableau, Tableau Agent performs better with extracts. When using live data sources, the time it takes to initiate your Tableau Agent session might be longer and viz load time will be slower
Only ask about analyzing your data. Tableau Agent isn't a general-purpose chat-bot. When it comes time to ask questions, start each request as if Tableau Agent is only aware of the current state of the visualization. Ask questions about what you want to see. For example, “What are my sales over time in California?".
Be specific and descriptive in your requests to help Tableau Agent better understand and respond to your data analysis needs. Break your request into discrete tasks. For example, first request a calculation, then ask for a viz that uses that new field, instead of asking for a viz and a new calculated field at the same time. For more information, see Tips for getting the best results from Tableau Agent(Link opens in a new window).
Can I use Tableau Agent for dashboards or stories?
No, currently Tableau Agent is only available for worksheets.
Technical Details
How does Tableau Agent handle data privacy and security?
Tableau Agent ensures data privacy and security by adhering to a zero data retention policy with third-party LLM providers. It respects all existing permissions and data policies within Tableau Cloud, ensuring only authorized users have access. Additionally, it can also employ pattern-based data masking techniques to protect personally identifiable information (PII) before it is processed by LLMs.
For more information, see Tableau AI and the Einstein Trust Layer(Link opens in a new window) and Select What Data to Mask(Link opens in a new window) in the Salesforce help.
How does Tableau Agent know what is in my data?
Tableau Agent indexes your data source at the start of each session to understand what data is present. Indexing takes in the data type, field name, and samples 1000 unique values for a field. This ensures responses are grounded in your data. Remember that Tableau Agent respects data masking if it is enabled, and might not be able to respond to requests that involve masked data.
If you have fields with high cardinality (many distinct values for a single field), Tableau Agent might not have indexed all the values. If you try to filter on a high cardinality field, Tableau Agent creates a filter for the field but might prompt you to select the values you want to include or exclude.
What if the viz created by Tableau Agent isn’t correct?
As with all AI, it's important to review the results you get when using Tableau Agent. Although Tableau Agent can scan your data source for the field names and data in those fields, it doesn't truly understand the data the way a human can. Always review the results from Tableau Agent. If a viz doesn't meet your expectations, try clarifying your request or asking Tableau Agent to recreate it. You can also interact directly with the viz to adjust the analysis to meet your needs. Think of the viz created by Tableau Agent as a starting point to jump start your analysis.
You can also always provide feedback using the thumbs up and thumbs down icon in the conversation pane.
Does Tableau Agent remember previous interactions?
Tableau Agent is a per-sheet experience, meaning it doesn't retain conversation history across different sheets. It has limited memory based on the information in the conversation pane of the current sheet. Additionally, if you publish your workbook using Publish As, Tableau Agent will not retain the conversation history.
Why was my request cut off?
There's a limit on how long your input can be (500 characters), and how much context from your data Tableau Agent can evaluate.