Video length: 9 min 49 sec
AI agents use autonomy, tools, data access, and managed conversation state to complete tasks beyond simple chatbot responses, with Microsoft Foundry Agent Service providing the platform to build and integrate them.
Defines AI agents as autonomous systems that can choose tools, look up data, take actions, and scale across enterprise use cases such as customer service, research, and developer assistance.
Highlights production security needs, including least privilege, input validation, monitoring, and human approval for high-risk actions like financial transactions or data deletion.
Explains how Microsoft Foundry Agent Service handles tool calling, state management, conversation threads, generated artifacts, and secure execution loops.
Covers the agent development pattern: connect to a Foundry project, create an agent, define instructions and tools, create a thread, process messages, review responses, and clean up resources.
Summarizes available tools, including Bing Search, File Search, Azure AI Search, Microsoft Fabric, Code Interpreter, Custom Functions, Azure Functions, and OpenAPI-based actions.