What is agent memory in AI?
Agent memory is the notes an AI assistant or agent saves, such as your preferences or past results, so it can use them later instead of starting from zero each time.
In plain English
A language model forgets everything once a conversation ends. It can only see what is in its context window right now. Agent memory fixes this by saving short notes outside the model, such as “prefers vegetarian restaurants” or “the team list is these six people”. Later, the most useful notes are put back into the context window, so the AI can use them.
An example
You tell an assistant once that your reports go to your finance team on Fridays. Next week you just say “send this week’s report”, and it already knows who gets it and when.
What to watch out for
Memory can keep out-of-date or wrong details, and it holds personal information. Check what your assistant has saved, delete what you don’t want kept, and turn memory off for sensitive work. See how the AI pieces fit together.