Introduction¶
Build a complete agentic AI solution — from cloud infrastructure to a production-ready, network-isolated application — using Microsoft Fabric, Microsoft Foundry, and Copilot.
Who this is for¶
- Developers building AI solutions
- Solution Architects evaluating Microsoft Data and AI stack
- Workshop Teams running hands-on sessions
Note
Assumes basic familiarity with Python, the Azure portal, and CLI tools. No prior Foundry, Fabric, or Copilot Studio experience needed.
How the tutorials connect¶
Each part produces artifacts that the next part consumes:
- Getting Started — Provision Azure infrastructure
- Data Foundation — Lakehouse, Semantic Model, Ontology, Search index
- Copilot Agent — Copilot Studio agent published to Teams
- Foundry Agent — Foundry with IQ tools
- API & Web App — FastAPI backend + React chat UI
- Network Isolation — Private endpoints for production
Get started¶
Before diving into the tutorials, you'll need to clone the repository and open it in VS Code:
Each tutorial is driven by GitHub Copilot skills — type a skill command like /miq-0a-infra-setup in Copilot Chat and follow the guided steps. Full setup instructions are in Tutorial 0: Getting Started.
What to expect¶
| Aspect | Details |
|---|---|
| Time | 30–60 minutes per part |
| Style | ~30% concepts, ~70% hands-on |
| Each part | Context, step-by-step instructions, expected outputs, verification |
| Two paths | GitHub Copilot skills in VS Code, or manual steps on each page |
Technologies¶
| Technology | Role |
|---|---|
| Microsoft Fabric | Lakehouse, Semantic Model, Ontology, Data Agent |
| Microsoft Foundry | Agent runtime & model hosting |
| Azure OpenAI | LLM inference (GPT models) |
| Azure AI Search | Document retrieval (RAG) |
| Copilot Studio | No-code agent creation |
| Microsoft Teams | Agent chat interface |
| FastAPI + React | Backend API & frontend chat UI |
| Azure App Service | Application hosting |
| Azure Private Link | Network isolation |
Next steps¶
Start with Tutorial 0: Getting Started.