talk-data.com

Everyone’s talking about AI agents! But what are they, and how do you build one? This talk cuts through the hype. Drawing on my experience building a GenAI platform, I’ll show that powerful agents are within reach, no advanced degree required. We’ll define agents simply: LLMs + tools + memory. Then we’ll build an agent with the OpenAI Python SDK, using coding basics you know: functions, loops, and conditions. I’ll show how you can enhance your agent with a knowledge base using Elasticsearch as a tool. By the end, you won't just understand agents; you'll be fully equipped to build your own.

Join this session to see how Microsoft and Elastic are partnering to deliver secure, AI-powered solutions for regulated industries. You’ll learn how to combine Elastic’s new Serverless offering on Azure with Microsoft’s Agent Framework to build intelligent, conversational experiences over embedded document data. We’ll walk through a real-world-inspired demo that shows how to ingest and semantically search sensitive documents using Elastic’s vector search and OneChat capabilities—then expose that functionality through Microsoft Teams using a custom-built Agent. The solution is enhanced with Microsoft’s Trustworthy AI capabilities to ensure responses are safe, secure, and compliant—all within a fully Azure-native environment.