References: Giving agents memory

Source material

Source curriculum (structural mirror, cited as further study):
• Microsoft, "Memory for AI Agents" (AI Agents for Beginners, Lesson 13)
  Author: Microsoft Cloud Advocates
  Lesson page: https://github.com/microsoft/ai-agents-for-beginners/tree/main/13-agent-memory
  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

• Memory for AI Agents (Microsoft) by Microsoft Cloud Advocates. The practitioner version of this lesson. It goes further than the conceptual short-term-versus-persistent split here, into memory implementations and finer-grained types (working, episodic, entity, persona memory), plus a pattern for an agent that decides on its own what is worth saving. MIT-licensed. Start here if you want to see memory built in code.

Going deeper

A short, durable list.

• Anthropic: Build with Claude (context and prompting). Practical guidance on what to put in a model’s context, which is the short-term half of this lesson: getting the right information into a single run.
• Model Context Protocol. The open standard for connecting models to external sources of state and data, the kind of plumbing persistent memory often rides on. Relevant once you move from what to remember to where it lives.

Adjacent topics

Where this leads inside this track.

• Agentic RAG. The next lesson. Where persistent memory lives and how an agent pulls the right piece at the right moment is a retrieval problem. Agentic RAG treats retrieval as a tool the agent chooses to call.
• The tool-use design pattern in depth. The previous lesson. Loading and saving memory typically happens through tools, so the tool-definition discipline from that lesson applies directly to memory operations.
• Building trustworthy agents. Later in the track. Persistent memory is stored personal data; what an agent remembers, and for how long, is part of building an agent users can trust.