Build Your First AI Agent in Python: From Prompt to Tool Use | Course #2

Build your first customer-support AI agent with Python and Azure OpenAI. In this lesson, we connect a model to two simple Python tools and inspect the data behind its answers. Our assistant can look up an order with get_order_status() and retrieve product details with get_product_info(). We use fictional local data so you can compare every tool result with the assistant's reply. In this video: Understand the chat UI, FastAPI backend, and Python agent Read order and product records from Python dictionaries Describe tools and their required inputs to the model Configure an Azure OpenAI connection using backend environment variables Follow a model request through Python tool execution and back to the model Inspect actual tool names, arguments, and results in Agent Activity Test order 104 , then try your own sample queries Stack: Python, FastAPI, Azure OpenAI Responses API, and a reusable React + Vite interface. The lesson focuses on Python; the frontend is provided as course infrastructure. This version uses a plain Python tool loop, without LangChain or LangGraph. Before starting: Follow the project's README, configure your Azure endpoint, API key, and deployment, and start both local servers. Never put your API key in frontend code or share it. Playlist: Agentic AI with Python: Build a Customer Support Agent Previous lesson: What Is Agentic AI? How Agents Work & What We'll Build. Next lesson: How AI Agents Use Tools: Tool Calling & the Execution Loop. Prerequisites: Basic Python, a configured Azure model deployment, and the project setup described in the README. #AgenticAI #Python #AzureOpenAI #AgenticAI #Python #AIAgents #AgenticAIExplained #AIAgentsForBeginners #PythonAI #ArtificialIntelligence #LLM #ToolCalling #FunctionCalling #CustomerSupportAI #HumanInTheLoop