Skip to main content

AI Agent Solutions

This domain covers designing and implementing AI agents — autonomous systems that can plan, use tools, and collaborate to accomplish complex tasks. It represents 5–10% of the AI-102 exam and is the newest addition to the certification.

AI agents go beyond simple prompt-response patterns. They combine LLM reasoning with tool use (function calling), memory, and multi-step planning. You'll learn to build agents that can query databases, call APIs, process documents, and coordinate with other agents to solve complex problems.

While this is the smallest domain by weight, the concepts here represent the cutting edge of AI engineering and increasingly appear in questions across other domains (especially Generative AI and Knowledge Mining).

What You'll Learn​

  • Design agent architectures with planning, memory, and tools
  • Implement function calling and tool use patterns
  • Build agents that interact with external systems
  • Implement multi-agent orchestration and collaboration
  • Handle agent errors, loops, and safety boundaries
  • Use Semantic Kernel and Azure AI Agent Service

Skills Measured​

  • Design an agent architecture for a given scenario
  • Implement tool use and function calling in agents
  • Implement multi-agent orchestration patterns
  • Configure agent safety boundaries and error handling

Challenges​

#TitleKey Topics
21Agent Architecture & DesignAgent components, planning loop, memory, tool registry
22Tool Use & Function CallingTool definitions, parallel calls, response handling
23Multi-Agent OrchestrationAgent handoffs, collaboration patterns, Semantic Kernel

Prerequisites​

  • Completed Domain 2 (Generative AI) — especially Challenges 12, 16, 19
  • Strong understanding of Azure OpenAI function calling
  • Familiarity with async programming patterns
  • Understanding of orchestration concepts