Microsoft · AI-103
Multi-agent solutions and approvals
AI-103 · Lesson 8 of 15 · 59:13
YouTube loads only after you press play. This chapter opens at its timestamp in the full course. Watch on YouTube ↗
In this lesson
Prepare for the AI-103 exam (Azure AI Apps and Agents Developer Associate). In this lesson: when to use several agents, Microsoft Agent Framework orchestrations (sequential, concurrent, handoff, group chat, Magentic), agents as tools and A2A, semi-autonomous workflows with human approvals (approval_mode, MCP require_approval, iteration caps, routines), and monitoring and evaluating deployed agents (server-side tracing, the monitoring dashboard, continuous evaluation, agent evaluators, cluster analysis). Includes three original practice questions.
Preview the topics
- A support system starts with a triage agent. Depending on the conversation, a billing or a returns specialist should take over and talk to the customer directly. Which orchestration fits?
- A refund agent built with Agent Framework must never issue a refund without a person’s OK, while its other tools run freely. What should you do?
- An agent understands requests and finishes most tasks, but traces show it often calls the refund tool with the wrong order number. Which evaluator measures this?