← back
Building Code First AI Agents with Azure AI Agent Service — Cedric Vidal, Microsoft
Takeaway
Azure AI Agent Service moves agent state, tools, and data integrations to the cloud, trading flexibility for a much simpler stateful dev model.
Summary
- Cedric Vidal (Microsoft) workshops Azure AI Agent Service via an outdoor-equipment sales analytics use case — natural-language queries produce dynamic tables and pie charts.
- Azure AI Agent Service is stateful: agents, threads, instructions, models, and data sources live in the cloud rather than being managed client-side — pros (simplicity) and cons (lock-in, learning curve).
- Agent capabilities demoed: function calling, code interpreter, file search, and Bing-search grounding for web info.
- Workflow: create-or-update agent → create/reuse thread → run on thread → poll status → display response; instructions attach to agent once.
- Compares with LangChain, LangGraph, Semantic Kernel and other agent frameworks — Azure differentiates via managed state plus enterprise security.
azureagentsmicrosoft
Original description
This workshop offers a hands-on introduction to developing Large Language Model (LLM)-powered AI agents using Microsoft’s Azure AI Agent Service. Participants will build a conversational agent capable of analyzing sales data, generating visualizations, and delivering actionable insights. The session takes a code-first approach using the Azure AI Foundry SDK for Python, and demonstrates how to integrate core Azure services including Azure OpenAI, Azure AI Search, and Azure Storage. Attendees will explore key concepts such as function calling, document grounding, and leveraging code interpreters to generate diagrams. The workshop also covers how to connect agents to external data sources like SQL databases (e.g., SQLite), enabling access to legacy relational systems. By the end of the session, participants will have a solid foundation for building and deploying intelligent, code-first AI agents with Azure AI Agent Service—ready to power real-world applications. About Cedric Vidal Cedric Vidal is a Principal AI Advocate at Microsoft, specializing in Generative AI 🤖, and the startup 🚀 and research 🔬 ecosystems. He is dedicated to promoting AI in startups and facilitating the transition of research and startup products to the market. If you're an AI Startup Founder or Engineer, I'd like to feature your work, come talk to me. Before his current role, Cedric spent 4 years as an Engineering Manager in the AI data labeling space for the self-driving 🚕 industry at Argo AI (now re-spawned as Latitude AI). He also served as the CTO of the Fintech AI SAAS startup Quicksign and worked as a software engineering services consultant for major Fintech enterprises for 10 years. Recorded at the AI Engineer World's Fair in San Francisco. Stay up to date on our upcoming events and content by joining our newsletter here: https://www.ai.engineer/newsletter