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.
Required Editions
| Available in: Tableau Next |
Tableau Agent with Data Analysis can be used in these surfaces.
- Tableau Next: Lightning experience pages, dashboard pages, metric details pages, Q&A Calibration, Semantic Model Builder
- Salesforce: Home page, Record page, Mobile app
- Slack: Metric cards, channel, thread, direct message, Agentforce for Slack (including agent scoping for semantic models), Tableau Next MCP integration for Slack
- Third Party: MCP, Embedding SDK
Tableau Agent with Data Analysis is supported in these contexts.
- Tableau Next Dashboards: Single dashboards with multiple semantic models. Context: Global filters including initial filters and embedded filters.
- Tableau Next Metrics: Single metrics, record pages in Lightning experience.
Context: Metric name, description, measure, aggregation, time dimensions, active filters.
Tableau Agent with Data Analysis supports these visualization types. Users can ask the agent to change the visualization type provided.
- Bar chart, line chart, table chart, donut chart, scatter plots, and heatmaps.
Supported Questions Types in Tableau Agent
Tableau Agent with Data Analysis can answer the following types of questions.
| Category | Context | Examples |
|---|---|---|
| Descriptive Questions | Top N list | "Show the top 10 opportunities by amount" |
| On the fly measure aggregations (Sum, Count, Avg, Min, Max) | "What's the average number of open tickets across accounts?" | |
| Breakdown of measure by dimensions | "Breakdown total sales by region, account type and opportunity stage " | |
| Trends | "Show our sales trend over the last 3 quarters" | |
| Period over period change | "What's the MoM change in closed opps in the last 6 months?" | |
| Measures | Drivers and contributors | "What are the key drivers for the open pipeline?" |
| Comparison Questions | Compare a measure over time | "Compare the number of open opportunities this year and last year" |
| Diagnostic Questions | Top and Bottom Contributors |
"Which regions contribute most to the sales growth in the current quarter?" "What contributes to the sales growth in the current quarter?" "What's driving sales?" |
| Statistical Analysis | Distribution Analysis | “What's the distribution of deal sizes for closed opportunities?” |
| Values Filtering (All Questions) |
Filtering: Dimension Values Exact value matching is supported when value is wrapped in quotes |
"Show all open opportunities in Canada And Brazil" Exact matching: “Show ACV for “ACME” account” |
| Dates Filtering (All Questions) |
Filtering: Dates (Year, Quarter, Month, Week, Day) Last N and To-Date |
"How many opportunities were created in the last 2 quarters? " |
| Calendars |
Fiscal Calendar Standard Calendar |
"How many opportunities were created in the last 2 fiscal quarters?" |
| "My" Questions | Assets associated with current user (depends on the semantic model configuration) |
"What are my sales?" "What are my open tickets?" |
| Organizational Hierarchy | When a role or manager hierarchy is defined in a semantic model, the agent can recognize questions based on that organization. |
"What's my team's revenue?" "How is my region performing?" |
| Multi-Turn Conversation Memory | Session Memory |
Question1: "Breakdown number of opportunities by stage" Question2: "Also show them by region" |
| Metadata Questions |
How measures or dimensions are defined What questions can be answered with the semantic model |
"How is the open pipe metric defined?" "What questions can you answer?" |
| Change of Visualization | Override the default visualization |
"Now show it as a bar chart and break it down also by state" "Breakdown number of opportunities by stage as donut chart" "Change to Line Chart" |
Exact Matching Syntax for Questions
Use exact matching syntax to control how the agent searches for values in your data. By using quotes around values, you can specify an exact match or a broader Value is contained search.
- Use quotes for exact matches when you know the precise value.
- Omit quotes for broader searches when you want to find variations.
- Be aware of business preferences, as the preferences may adjust matching behavior for known abbreviations.
When an exact match query returns no results, the agent notifies the user that no match was found and suggests performing a search with contained values.
| Category | Context | Examples |
|---|---|---|
| Exact Match (Quoted Values) | When a question includes a value in quotes, the agent will only return records that match exactly (case-insensitive). |
Question: Show me accounts for "Bank of America" Returns records with the exact name only "Bank of America" (not "Bank of America Corp", "Bank of America NA") |
| Contained Match (No Quotes) | When no quotes are used for a value in a question, the agent finds all records that contain the value anywhere in the field. |
Question: Show me accounts for Santander Returns: "Santander", "Santander UK", "Banco Santander", "Santander Consumer" |
| Override Phrases | Uses fuzzy matching when the phrase "similar to" is used. |
Question: Show accounts similar to "Bank of America" Override Keywords:
|
| Mix of Quoted and Unquoted Values | Provides results with exact and contained matches |
Show closed opportunities for "PayPal Holdings" Response: "PayPal Holdings" (exact only), and "SAP", "SAP America", "SAP SE" (contained match for closed opportunity) |
Business Preferences (data model rules that define how certain values should be interpreted) may override the standard exact matching behavior.
For example, given a business preference rule:
US = United States, US should be treated as equivalent to UNITED STATES
The question "Show accounts in the US", may trigger an exact match even though "US" is not quoted.