Automation glossary

AI Memory

AI memory is how an AI system keeps information across turns or sessions, either within the current context window or by saving and later retrieving facts from storage.

In plain language

Short-term memory is what the model can see right now; long-term memory is notes it saves and looks up later.

Example

An assistant remembers a client’s preferred report format by saving it to a memory store and retrieving it in future conversations.

Context Window · Retrieval-Augmented Generation (RAG) · AI Agent · Vector Database

Common misconception

A model does not remember past conversations on its own; memory exists only when the application stores and supplies it.

Implementation guidance

Decide what is worth remembering, let users see and delete stored memories, keep sensitive data out, and expire stale information.

Primary references

← Browse every glossary term