Tableau Knowledge is an enterprise knowledge management solution that helps AI agents understand your organization's data and business context. It connects trusted Tableau Cloud content and business knowledge into live knowledge graphs, giving agents context so they consistently provide accurate and explainable answers.

This intelligent network of knowledge and assets enables agents to understand not only what your data contains, but also what it means to your business.

For example, Tableau Knowledge can help an agent interpret revenue, profit, or regional definitions according to the business logic your organization has established, rather than making assumptions.

Note: Tableau Knowledge is currently available in English only.

Licensing and roles

Tableau Knowledge is available for Tableau+ licenses on Tableau Cloud.

Your Tableau Cloud site role determines the actions that you can perform:

Key concepts

To better understand Tableau Knowledge, become familiar with these common terms:

Benefits of Tableau Knowledge

Tableau Knowledge provides visibility across your organization’s knowledge, allowing you to optimize it for agent consumption. AI Agents that consume this knowledge provide detailed, accurate answers to your questions.

Tableau Knowledge helps you and AI agents to:

• Use consistent business definitions. Add business preferences that define terms and business rules for the whole knowledge graph. Connected agents can use this shared context consistently.

• Ground answers in trusted content. Agents use trusted business definitions and calculations from semantic models, metadata, relationships, and business preferences in the knowledge graph, rather than relying only on raw data.

• Understand the context behind the answers. Tableau Knowledge helps agents use relevant data assets and business definitions when responding to questions.

• Improve graph quality. The Tableau Knowledge Graph Quality Scorecard and related improvements identify opportunities to improve the knowledge graph, such as missing descriptions, owners, certifications, and connections between assets.

• Maintain governance. Tableau Knowledge works with Tableau Cloud access controls to help protect access to Tableau content.

Tableau Knowledge users

Tableau Knowledge supports the people who manage organizational knowledge and the AI agents that use it.

Creators

Tableau Creators build and maintain knowledge graphs. They select Tableau Cloud content, upload supporting documents, add business preferences, and address recommendations for strengthening the graph.

For example, a creator can:

• Create a knowledge graph.

• Add Tableau Cloud content to a knowledge graph.

• Upload a glossary or other supporting document.

• Add a business preference that applies to the entire graph.

• Review graph health and optimization suggestions.

• Set the default knowledge graph for the site.

Knowledge graph managers can also use the Tableau Agent for Knowledge to ask questions about a knowledge graph and identify ways to improve it.

AI agents and their users

AI agents including Tableau Agent for Knowledge in Tableau Cloud and any agent running over MCP query Tableau Knowledge graphs to answer questions for users.

When used as a consumer, an agent reads from the knowledge graph. It retrieves relevant business context including: assets, metadata, relationships, and business preferences to ground its responses in your organization's data and terminology.

How Tableau Knowledge works

A knowledge graph represents the network of knowledge across your business—it connects enterprise data, business definitions, and context to show how they relate to each other. When you create a knowledge graph it is automatically generated based on the selected sources and inferred relationships. The knowledge graph is presented in two ways:

• Graph View—each entity type is represented as a node connected to all nodes it is related to.

• Explorer View—each entity type is represented in a hierarchical tree where related nodes are represented as nested nodes.

Both views have the same functionality. The only difference is the presentation. Note that the nodes that appear in the knowledge graph are only the nodes that you are permitted to see.

When you connect Tableau Cloud content to Tableau Knowledge, the knowledge graph represents the included assets and their relationships. The graph is generated live once you select the data to include. When you create a knowledge graph, Tableau Knowledge automatically infers relationships between sources and includes them in the graph. You can also add any Tableau asset such as workbooks and published data sources (PDS) and business preferences to provide more guidance about how your organization defines and uses its data.

For example, a graph can connect a revenue field to its published data source, related workbook, and a business preference that explains how your organization defines revenue.

Default knowledge graph

When you use an agent to query Tableau Knowledge without specifying a graph, Tableau Knowledge refers to the default knowledge graph. The same concept applies to the Tableau Agent for Knowledge. If you don’t specify a graph in your question, the Agent answers your questions from what it knows about the default graph. In the home page dashboard, the specified graph (default or other) is indicated. All metrics and reporting correspond to the indicated knowledge graph.

Entity types

Tableau Knowledge graph nodes can be of any of these Tableau Cloud entity types:

• Published data sources
• Workbooks
• Dashboards
• Sheets
• Business preferences
• Embedded data sources

Business preferences

Business preferences are natural-language instructions that explain organization-specific logic, terminology, and how your organization uses data. They help agents interpret information consistently. For example, a business preference might define how to answer the question when asked about the best performing or largest accounts. In this case you could add a preference to instruct the Tableau Agent to use the accounts with the highest total_amount.

Business preferences apply globally across the entire knowledge graph. The business preference can be in the form of simple instructions in natural language or an uploaded document. If you use a document, each preference must be in its own paragraph, one business preference per paragraph.

Supported document formats are:

• PDF (.pdf)
• Microsoft Word (.docx)
• Plain text (.txt)
• Markdown (.md)
• Comma-separated values (.csv)

Note: If you added business preferences to a Tableau core workbook which is part of a Tableau Knowledge graph, the preferences are ingested as business preferences and associated with the workbook node. These preferences take precedence over the global business preferences and cannot be edited inside the knowledge graph. For more information about business preferences, see Improve response quality for Dashboard Q&A.

Knowledge graph quality scorecard and suggestions

Tableau Knowledge provides a health score and recommended actions to help you evaluate and strengthen a knowledge graph.

Suggestions can identify issues such as:

• Assets without descriptions
• Assets without owners or certifications
• Disconnected assets
• Incomplete business context

Use these suggestions to determine where additional documentation, context, or governance can improve the quality of agent responses. The higher the score on the scorecard the more agent ready your graph is.

Tableau Agent for Knowledge

Get insights on your knowledge graph quality, ask questions about any knowledge graph, or get help with managing knowledge graphs by starting a conversation with the Tableau Agent for Knowledge. You can do this on any screen or in the dedicated Chat Menu. Remember to specify the knowledge graph you want to ask about; if you don’t, Tableau Agent for Knowledge uses the default knowledge graph to derive its answers. To get started, ask a question in the chat window or choose a suggested question.

Example use case

Peter, a Tableau data analyst, receives a question from Mary in Sales: “What NQLs did my org create?” While working on the Sales Pipeline Overview dashboard Peter queries the Tableau Agent.

Tableau Agent queries the default knowledge graph and identifies the relevant Leads field, regional sales dashboard, and business context. It also identifies that NQL refers to new leads based on the business preferences provided in the graph. It can then provide an answer that is based on the organization’s Tableau Cloud content and terminology, rather than a generic interpretation of the question. The Agent supplies the correct answer and Peter sends Mary the response.

See also

• Tableau Cloud Help
• Tableau Cloud: Tableau MCP Server
• Salesforce Help: Tableau

Other articles in this section