You are here:
About Conversational Analytics in Tableau Agent
Tableau Agent with the Data Analysis subagent is a deeply integrated, intelligent layer for conversational analytics, built to meet business users where they work, across Tableau Next, Salesforce, and Slack. Ask Tableau Agent freeform questions about your data, and get answers in the form of rich text and visualizations, grounded in your organization's semantics and business logic.
Required Editions
| View supported editions. |
When you have access to a Tableau Agent that has been set up for data analysis, you can explore data through natural conversation in the Agentforce panel. Ask questions about your data to get answers in the form of insights and relevant visualizations. For transparency, responses from the agent include the sources and reasoning used for analysis.
When Agentforce isn't enabled, the metric details page uses the default insight exploration experience without the use of generative AI.
Conversational analytics are enabled in Tableau Agent through the Analytics and Visualization template in Agentforce. This template includes the pre-built Data Analysis subagent with the default Analyze Data action to craft rich-text, natural language responses that directly answer your questions about data.
When Tableau Agent is enabled in Slack, you can use it to ask questions about shared Tableau Next metrics and dashboard data in that context. For more information, see Ask Questions About Shared Metrics and Dashboards in Slack Using Tableua Agent.
Tableau Agent requires that any semantic model it queries is enabled for use in Tableau Next. A Tableau Next user who has Semantic Model editing permission must enable the semantic model in the Analytics Agent Readiness dialog box. For more information, see Enable Analytics Agent Readiness in Data 360 help.
Trust is built into the experience by inviting thumbs-up or thumbs-down ratings and open feedback. These responses are used to improve the conversational experience. For more information, see Trust and Agentforce.
.Capabilities of Tableau Agent with Data Analysis
Conversational analytics in Tableau Agent provide advanced reasoning and a broad range of deep analysis.
- Answer consistency and accuracy. Agents produce reliable and reasonable responses across a variety of question types. In addition the agent enforces adherence to business context and guidance.
- Advanced reasoning and narrative generation. The agent provides informative narratives and rationale behind its conclusions to help you understand the reasoning behind an answer. Responses include useful observations and high-level insights in addition to answers to questions. As the agent thinks about your question, it provides transparent updates about its reasoning process.
- Deep context and memory for multi-turn conversations. Refine questions, refer to previous questions and answers in a session, or ask natural follow-up questions without having to remind the agent of the context. Conversations with the agent are accurate and effective because the agent remembers you asked and manages context in back-and-forth conversation.
- Guided alternatives. When the agent can't answer a question directly, it suggests alternative questions or directions to support further exploration.
- Metadata awareness and discoverability. Users can ask the agent "What questions can you answer?" or "How is the ACV metric defined?" to better understand what data is available and how to query it. Users can ask the agent what data objects, fields, metrics, dimensions, and semantic definitions are available in the data.
- Support for multiple visualization types. In addition to bar charts, line charts, and table charts, the agent supports donut charts, scatter plots, and heatmaps. Users can ask to change an existing visualization type for quick exploration of the same data through different perspectives without needing to re-ask a question.
- Support for a variety of question types. The agent seamlessly handles advanced queries, including period-over-period, trend analysis, top drivers, and distribution and statistical correlation. For a full list of the types of questions you can ask Tableau Agent with Data Analysis, see Supported Questions and Surfaces for Tableau Agent.
How Tableau Agent Queries Data
Tableau Agent uses a single semantic query path for questions. Semantic queries interact with the underlying semantic model dynamically, allowing for custom breakdowns, entity comparisons, and deeper exploration. Tableau Agent generates and runs a semantic query against the underlying semantic model of an associated metric or dashboard. It synthesizes the query output data to create a textual answer and creates a visualization based on the semantic query.
Tableau Agent with the Data Analysis subagent:
- Answers questions in a session using the underlying data from the semantic model as a grounded source to query against and analyze.
- Clarifies intent when it encounters complexity or ambiguity in a question to provide a better, more accurate response.
- Plan Your Tableau Agent Implementation for Conversational Analytics
As an analyst implementing Tableau Agent in Tableau Next, you play an integral role in getting the agent to work well for your organization. Successfully enabling conversational analytics in Tableau Next includes five major phases. - Ask Questions with Tableau Agent in Tableau Next
When Tableau Agent with Data Analysis is enabled in Tableau Next, you can ask questions to explore your data. When Tableau Agent is enabled in Slack, you can use it to ask questions about shared Tableau Next metric and dashboard data in that context. - Supported Questions and Surfaces for Tableau Agent with Data Analysis
You can use Tableau Agent to ask questions about your data across Tableau Next, Salesforce, and Slack. This topic lists the types of questions, surfaces, and contexts where Tableau Agent with Data Analysis is currently supported. - Limitations for Conversational Analytics in Tableau Agent
In Tableau Next, agents enabled for conversational analytics have the following limitations related to questions and responses. Conversational analytics are available in agents based on the Analytics and Visualization template with the Data Analysis subagent in Tableau Next. - Tips for Improving Performance in Tableau Agent with Data Analysis
Multiple factors can affect performance in Tableau Agent in Tableau Next. This topic lists tips for improving query performance when asking questions in Tableau Agent with Data Analysis. - Add the Data Analysis Subagent to the Analytics and Visualization template in Tableau Next
The Data Analysis subagent in the Analytics and Visualization template enables conversational analytics in Tableau Agent.
