Build a bilingual (English, Swahili), AI assistant for Savanna Bites restaurant (Nairobi), using the Microsoft Foundry Toolkit in VS Code Microsoft Foundry.

1. Fork this repository
2. Clone your fork locally:

git clone https://github.com/<your-username>/Safaricom-Decode-Agents-Workshop.git
cd Safaricom-Decode-Agents-Workshop

1. Open the folder in VS Code:

code .

Then follow the workshop steps starting at workshop/01-overview.md.

1. Explore & Compare Models — use GitHub Copilot Agent Mode and Microsoft Foundry Toolkit Model Catalog
2. Create Agent — set up Biashara Agent in Agent Builder with gpt-4.1-mini via Microsoft Foundry
3. System Prompt — define the agent's personality, role, and language behaviour
4. MCP Tools — connect the local data server for grounded answers
5. Evaluation — run manual and AI-assisted evaluations on agent responses
6. Web App — deploy a Flask chat interface powered by the agent
7. Export Code — export agent code for production with Microsoft Agent Framework

• lab-guide.md: Participant index and sequence
• workshop/: Split workshop sections
• agent/system-prompt.md: Agent instructions template
• facilitator-notes.md: Trainer notes

1. BRK441: Build and launch AI agents fast with Microsoft Foundry and the Microsoft Foundry Toolkit