Build an AI-Ready Semantic Model

Understand AI-readiness in semantic models

Building the semantic model requires following best practices as you design and define it. Tableau Next also provides tools to help you make the semantic model ready for use with Tableau Agent.

Familiarize yourself with the concepts in Design Semantic Models for AI Readiness.

Select a semantic model

Create a semantic model or use one that already exists. Add data model objects (DMOs) and define relationships. You can do this manually, or with AI-assistance using Einstein or Data Pro in the Semantic Model Builder.

Create calculated fields

If needed, create calculated fields. You can do this manually, or with AI-assistance using Einstein or Data Pro in the Semantic Model Builder. For example, you might create calculated fields for KPIs of interest or for dimension fields that don't physically exist in your data set.

Add descriptions

Add descriptions to reduce ambiguity in objects and calculated fields. You can do this manually, or use AI-generated descriptions.

Define business preferences

Add business preferences to define business-specific knowledge within the semantic model. You will want to provide tailored business preferences for unique jargon, logic, and concepts in your domain. You can revisit this step during optimization as you test and refine the model more.

For example, if you know that the term "ROI" is referring to Advertising Spend, you can write a business preference to let the agent know ROI means Advertising Spend.

Define metrics

Define metrics in the semantic model for high-impact and highly-used calculated fields or field measures. Metrics contain targeted calculations, time-scoped logic, contextual filters, and formatting. They provide context that guides agents in understanding where and when specific logic applies.

Any calculated field that appears frequently in analysis or decision-making is a good candidate to be upgraded to a metric.

Note

As part of this step, also review related visualizations and dashboards you’ve created based on the semantic model. Make sure the data makes sense and names of fields are understandable for your users.

Enable agent readiness

Enable the semantic model for use with an agent created from the Analytics and Visualization template.

Analytics Agent Readiness is a setting that marks a semantic model as mature and ready for use by an analytics agent. This step is required to be able use the analytics agent.

Note Admins can create fallback lists of semantic models per agent for scenarios where the semantic context isn't clear. For more information, see Set Up Agent Scoping for Tableau Agent.

Validate model

Validate and test the model in the Semantic Model Builder to make sure that it’s complete: all required DMOs, relationships, and calculated fields are included. Also make sure metrics and calculated fields aren't duplicated in the model.

Use Semantic Model AI Optimization (Beta) to gain clear visibility into your semantic model's health and AI-readiness, and see suggested recommendations to improve it. This tool runs a series of automated checks against your semantic model and calculates an overall model health for AI-readiness.

Use Q&A Calibration (Beta) in Tableau Next to improve accuracy in Tableau Agent. Test questions in Tableau Agent and validate the agent’s answers against your specific data and rules. The list of Verified questions that you compile and save improves agent accuracy for similar questions. When you mark a question as Inaccurate, you can further fine-tune business preferences, update fields and metadata, and update calculations to improve accuracy and performance. After you make the changes, you can retest the questions until they are accurate.

Note

After using Q&A Calibration, you can also use manual testing to determine if the model is complete and generates the queries you are expecting.