References: Agents that retrieve their own information: agentic RAG

Source material

Source curriculum (structural mirror, cited as further study):
• Microsoft, "Agentic RAG" (AI Agents for Beginners, Lesson 05)
  Author: Microsoft Cloud Advocates
  Lesson page: https://github.com/microsoft/ai-agents-for-beginners/tree/main/05-agentic-rag
  License: MIT
Clawdemy's lessons are original prose that follows the pedagogical arc of this
source. We do not reproduce or transcribe it; we cite it as the recommended
companion. All rights to the original materials remain with their authors.

Read this next

• Agentic RAG (Microsoft) by Microsoft Cloud Advocates. The practitioner version of this lesson, including iterative retrieval loops, self-correction, the boundaries of agency, and runnable samples. MIT-licensed. Start here to see agentic retrieval built in code.

Going deeper on the retrieval black box

This lesson treats retrieval as a black box on purpose. If you want to open it, these are the pieces inside.

• Retrieval Augmented Generation for Knowledge-Intensive NLP Tasks (Lewis et al., 2020). The paper that introduced RAG. The primary source for the retrieve-then-generate idea this lesson starts from.
• Model Context Protocol. An open standard for connecting models to external data sources and retrieval tools, the kind of plumbing a retrieval tool rides on.

Adjacent topics

Where this leads inside this track.

• Planning: breaking a goal into steps. The next lesson. So far the agent decided one move at a time; planning is deciding an ordered sequence of steps before acting, for tasks too large to solve one reaction at a time.
• Giving agents memory. The previous lesson. Retrieval is often how persistent memory gets pulled into a run, so the two lessons describe two halves of the same information flow.
• The tool-use design pattern in depth. Earlier in the track. Because retrieval is just a tool, that lesson’s discipline on describing tools well applies directly to the retrieval tool.