In week 3, you will evaluate enterprise AI platforms and vendors using weighted criteria while distinguishing between agent, copilot, and workflow automation paradigms. You will learn core agentic architectures including ReAct, plan-and-execute, single-agent, and multi-agent systems, and implement a hands-on MCP server with secure read-only and write-with-approval workflows. Additionally, you will apply data governance principles like RACI and data cataloging to assess how data quality impacts RAG reliability. Finally, you will identify and mitigate agent risks including hallucination, tool misuse, prompt injection, and EU AI Act triggers.