Power BI: Chat with Your Data Schema

Key insights

• Simplify the Data Schema and Chat with Your Data: guide Copilot to prefer the most useful fields and measures so answers and visuals match business intent.
  Set this to shape which parts of your semantic model Copilot prioritizes when users ask questions.
• Where to set it: open Power BI Desktop and use the Prep data for AI button, then the Simplify data schema tab to select or deselect fields and measures.
  Changes persist when you publish the model to the Power BI Service.
• Not a security feature: this setting does not hide objects from users and does not replace Row-Level Security (RLS) or object-level controls.
  Hidden fields, hierarchies, and relationships still affect model behavior and access rules remain enforced.
• How Copilot behaves: visuals can override schema choices and Copilot may fall back to implicit calculations if a measure is deselected.
  Re-enabling a measure can change Copilot’s answers and the visuals it builds.
• Best practices: deselect technical IDs and keys so Copilot prefers user-friendly fields, keep clear, named measures, and add synonyms or verified answers where helpful.
  These steps reduce confusion and improve the relevance of results.
• Benefits and limits: simplifying the schema improves accuracy and cuts hallucinations while speeding ad‑hoc analysis across reports and Fabric sources.
  Prerequisites and limits: enable Q&A on the semantic model, current support is English, and some scenarios still need the full model to answer correctly.

Keywords

Power BI data schema, Chat with your data Power BI, Power BI semantic model, Power BI Q&A chat, Power BI data modeling, Natural language queries Power BI, Power BI schema visualization, Power BI conversational analytics