Agentic Systems with Microsoft Agent Framework | بالعربي

AI agents are moving beyond demos and proof-of-concepts into real-world systems that need to be reliable, observable, scalable, and production-ready.
In this session, we’ll explore how to design *production-grade agentic systems* using Microsoft Agent Framework — starting with an important question: *when should you use an agent, and when is traditional code or a workflow a better choice?*
We’ll cover the core architecture of Microsoft Agent Framework, tool-enabled agents, agents vs workflows, multi-agent orchestration, state management, human-in-the-loop patterns, observability, and the challenges of moving from POC to production.
### What we’ll cover
* From AI assistants to agents and agentic systems
* When to use an Agent — and when not to
* Microsoft Agent Framework architecture and core concepts
* Agents, tools, and external systems
* Agents vs Workflows
* Multi-agent orchestration
* State, checkpoints, and human approval
* Reliability and observability
* Moving from POC to production
* Hands-on demo: *AI Production Release Guardian*
The session will focus not only on how to build agents, but on how to design systems where *agents, workflows, and traditional software work together effectively in production.*
Whether you're a developer, AI engineer, cloud architect, or technology leader, this session will help you move beyond simple agent demos and understand how to design agentic systems for real production environments.
*Join us and discover what it really takes to build AI agents that are production-ready.*