Скачать или смотреть Next Steps Call a Tool with Microsoft Foundry in Python

In this video, we walk through how AI tool calling actually works using Python and the OpenAI API.

Instead of letting the AI directly call external APIs, we give it a function it’s allowed to request. When the model decides it needs real-world data, it asks for that function, we execute it in Python using standard HTTP requests, and then we pass the result back to the model.

✅ This keeps everything clear and predictable:
✅ We let the model decide when data is needed
✅ We stay in control of how the data is fetched

External services are called explicitly, not magically

We use a real example with weather data to show how tool calling, argument extraction, and follow-up responses fit together in a practical workflow.

This video is for developers who want to understand what tool calling really is, without hype or unnecessary complexity.

All lessons on github: https://github.com/TheAndrewHess/Micr...

Chapters:
0:00 Introduction
0:45 Last week
3:08 Lesson4 Updating Our Code
4:24 Adding the Tool Role
8:25 Our Function To Get Weather
10:25 Test our Tool
12:15 Conclusion