Hands-On with Microsoft Agent Framework: Build Your First Agentic Workflow

In this hands-on session, participants will build a small but complete agentic workflow using Microsoft Agent Framework. Starting from scratch, we’ll create a basic AI agent, connect it to one or two tools, and orchestrate them into a multi-step flow that handles a simple real-world task (like triaging a request and drafting a response). Along the way, you’ll learn how to configure agents, pass context, and see what’s happening inside the workflow as it runs. By the end, you’ll have a working example and a practical checklist to start building your own agentic solutions.

Exploring the Workflow Dungeons with Microsoft Agent Framework

This session dives into the “workflow dungeons” of Microsoft Agent Framework: the graph-based orchestration that lets you combine agents, tools, and human steps into robust processes. We’ll walk through how to model complex flows with branches, loops, and approvals, and how to keep them observable and debuggable. Using a realistic scenario (like claims, approvals, or document review), we’ll show how to turn a messy manual process into a clear, testable workflow. You’ll walk away with patterns, pitfalls, and mental models for designing agentic workflows.

Meet the Microsoft Agent Framework: Agents & Workflows

Microsoft Agent Framework is the new way to build AI agents and orchestrated workflows on top of large language models. In this session, we’ll introduce the core concepts—agents, tools, memory, and workflows—and show how they fit together to solve real business problems. You’ll see simple examples of going from a single agent to a small multi-agent flow, plus where the framework fits alongside Azure OpenAI and other AI services. By the end, you’ll know when to reach for Agent Framework and what your next steps are to start experimenting.

The Future of Microsoft’s Agentic AI: Patterns, Pitfalls & the Road Ahead

Agentic AI is evolving fast — and Microsoft is reshaping its AI ecosystem to make multi-agent collaboration smarter, faster, and more powerful than ever.

In this session, we’ll explore the latest and greatest in Agentic AI for the Microsoft stack — from advanced multi-agent collaboration patterns you can use today, to insights on the new CoreAI division, the merging of Semantic Kernel and AutoGen, and what the upcoming Agents Framework could mean for developers.

Through real, working code demos, you’ll see how to design agent workflows that can reason, plan, and collaborate to tackle complex tasks — while understanding the current limits and the opportunities that lie ahead. (if access to the latest bits is provided, otherwise will show this with Semantic Kernel multi-agent stack)

Whether you’re building your first AI agent or pushing the boundaries of what’s possible, you’ll leave with a clear view of where we are, what you can build now, and where Microsoft’s agentic AI is headed next.

Agent Framework: The New Standard for Agentic Development

AI agents are more than conversation—they’re decision makers and actors. In this session, we dive into Microsoft Agent Framework, showing how it empowers you to build robust, production-ready agents with minimal boilerplate. We’ll cover the core building blocks—agents, tools, orchestration—and show why this framework is a next-gen unifier that brings together ideas from Semantic Kernel, AutoGen, and open AI runtimes. Then, we’ll build a working agent live, wiring a tool and driving a simple scenario, to show how little code is actually needed to power something useful. Along the way, you’ll see the elegance, flexibility, and developer ergonomics that set Agent Framework apart. This is not theory: you’ll see it in action, and walk away inspired by how much you can build with so little.

Agent Framework: The Not-So-Secret Superpower — Workflows

Agents are powerful on their own—but together, they become superpowered. In this session, we explore Microsoft Agent Framework Workflows, the not-so-secret capability that turns standalone agents into intelligent, coordinated systems. You’ll discover how workflows connect multiple agents, decisions, and human input into seamless, dependable processes that mirror real-world collaboration.

We’ll start with a simple linear flow, then evolve it into a more advanced workflow that introduces branching, human-in-the-loop interactions, and adaptive control—showing how workflows give agents coordination, resilience, and purpose. We’ll discuss the concepts behind agentic workflows, what makes them different from traditional graph orchestrations, and why they represent the future of enterprise-grade automation.

Through practical examples and case studies, you’ll see how workflows elevate agents from single-task executors to cooperative problem solvers—showcasing the superpower that makes the Agent Framework stand apart: the ability to think, act, and collaborate as one.

Harnessing the Power of AI with GitHub Copilot: Turbocharge Your Development Workflow

Join us for an interactive session on leveraging GitHub Copilot, the AI-powered coding assistant, to supercharge your development process. Learn how to seamlessly integrate Copilot into your workflow, explore its capabilities, and uncover best practices to maximize its potential in enhancing code quality and developer productivity. Don't miss this chance to elevate your coding experience with AI!

requirements: fast Internet , non-filtered traffic to access LLM's and other AI services.
Target: Developers or Team Leads interested in the real abilities of GH Copilot.

Lessons learned on Applied Advanced Agentic AI

In this session, we will peek at some projects and tools I have been working on using the latest models and Agentic AI techniques, showing what works best and gets the job done.
We will see some Agentic AI workflows in action delivering business value, then understand the concepts behind them and, lastly, the code.
This session will focus mostly on Semantic Kernel and some custom implementations for the OpenAI RealTime API.

Inside the Mind of a Graph Agent: Crime Investigations

What does it mean for an AI agent to think in graphs? In this session, we’ll look inside the mind of a graph-native detective agent that investigates a Neo4j crime graph powered by Microsoft Foundry.

Using the POLE model (People, Objects, Locations, Events) hosted on Aura, we’ll follow Holmsworth Marot, a Golden Age–style detective persona, as it connects to the graph through the Model Context Protocol and the Microsoft Agent Framework. The agent uses Cypher to traverse suspects, victims, locations, and events – and then explains its findings in the voice of a Golden Age detective, turning graph queries into narrative crime investigations.

We’ll walk through how the crime graph is modeled, how Cypher tools are exposed to the agent, and how we instrument the agent to reveal the sequence of graph queries and subgraphs it uses while investigating a case. You’ll see an ‘investigation trace’ that shows which Cypher queries were executed, which parts of the Neo4j graph were inspected, and how those pieces are combined into a narrative explanation.

The King’s Council of Agents: Orchestrating Medieval Workflows with the Microsoft Agent Framework

In a far-away kingdom, the King is drowning in petitions. Subjects arrive from every corner of the realm with questions, crises, and impossible requests: “Settle this dispute!”, “Find the truth in these scrolls!”, “Send word to all guilds before sunrise!”

No single royal advisor can keep up — so the King does what any modern monarch would do: he forms a Council of Agents.

In this story-driven session, we’ll explore how to design and orchestrate a guild of AI agents using the Microsoft Agent Framework and its magical workflows, without getting lost in low-level plumbing. Each member of the court represents a real-world agent pattern: the Chancellor as orchestrator/router, the Court Wizard as planner, the Royal Scribe as memory, the Knight-Commander as executor… and the Jester who tests assumptions and makes fun of everyone.

Rather than promising a full end-to-end implementation, we’ll focus on reusable patterns, mental models, and design choices: how to turn messy business, aka royal, problems into well-defined agent roles and workflows, and how to make agents actually perform useful work instead of just giving fancy answers. You’ll leave with a clear picture of what an agentic workflow can look like in your own kingdom — and how to start building your first royal council without losing your head.

Takeaways:
- Map real business problems to specialized agent roles with the Microsoft Agent Framework.
- Design agentic workflows where planners, memory, and executors work together.
- Use the “Jester” mindset to keep agents useful, testable, and grounded in reality (and also, fun)

Between jokes and the story will introduce the Microsoft Agent Framework, what's an agent, a workflow and build a complex workflow that fans-out, iterates, provides feedback and why not, helps save the day. Of course, there will be MCPs and maybe, graph databases as knowledge graph - maybe ;)