Lab — Build AI Agents with Microsoft Foundry¶
Welcome to the hands-on lab for the Microsoft Cloud & AI Frontier Week Hackathon — where ideas turn into real solutions.
Throughout Frontier Week, you've explored how AI is transforming organizations. This is where you put that into practice.
In this lab, you'll build, monitor, evaluate, and orchestrate AI agents using the Microsoft Foundry SDK — following a guided, scenario-based experience designed to take you from concept to a working, enterprise-ready multi-agent system.
By the end, you won't just understand how agents work — you'll have built one you can trace, evaluate, and deploy.
Choose Your Scenario¶
All three scenarios use the same five-challenge structure. Pick whichever industry fits your interest.
| Scenario | Domain | What You Build |
|---|---|---|
| 🏭 Factory | Predictive Maintenance | Anomaly Detection + Fault Diagnosis agents |
| 📋 Claims | Insurance Processing | Claims Triage + Claims Decision agents |
| 📞 Call Center | Customer Support | Intent Classification + Resolution Advisor agents |
Challenge Structure¶
Every scenario follows the same five challenges:
| # | Challenge | Duration |
|---|---|---|
| 0 | Setup — Deploy Azure AI Foundry infrastructure | 20 min |
| 1 | Build Agents — Create two AI agents with tools | 30 min |
| 2 | Monitor — Enable GenAI tracing with Application Insights | 20 min |
| 3 | Evaluate — Run systematic quality evaluations | 30 min |
| 4 | Workflow — Multi-agent orchestration via the Foundry portal | 20 min |
Prerequisites¶
- Azure subscription with Contributor access
- Python 3.10+
- Azure CLI (
az) installed and authenticated (az login) - A terminal (bash, PowerShell, or WSL)
Getting Started¶
- Clone this repo and pick a scenario folder (
factory/,claims/, orcallcenter/) - Start with Challenge 0 — it provisions everything you need
- Work through challenges 1–4 in order; each builds on the previous one
- The
agents.pyanddeploy.pyscripts are ready to run — read the README in each challenge folder for what to do