Skip to content

Tutorial part 1: Build the data foundation

In this tutorial, you generate sample data and create a complete data layer on Microsoft Fabric: a Lakehouse with Delta tables, an Ontology with entity types and relationships, and a published Data Agent.

Prerequisites: Tutorial part 0 completed, DEPLOY_FABRIC=true in scripts/.env

Skills

Run these skills in Copilot Chat:

Skill What it does
/miq-1a-generate-data Generate sample tables and documents with Azure OpenAI
/miq-1b-create-fabric-items Create Fabric workspace, Lakehouse, Semantic Model, Ontology, and Data Agent
What's happening in this module

Generate data -- Azure OpenAI creates realistic CSV tables and PDF documents that match your chosen industry. Azure OpenAI overview

Create Fabric items -- The script creates a Fabric Workspace, loads tables as Delta format into a Lakehouse backed by OneLake, defines an Ontology with entity types and data bindings, and publishes a Data Agent as an MCP endpoint that downstream agents call to query your data.

Supporting documentation

Next steps

Continue to Tutorial part 2: Create a Copilot agent.