# Agent memory · Glossary
The mechanisms that let an AI agent retain and recall information beyond a single model call, such as the current task’s history, facts learned about a user or records of previous outcomes.
[Glossary](/glossary) · Agentic systems and generative AI

# Agent memory

     The mechanisms that let an AI agent retain and recall information beyond a single model call, such as the current task’s history, facts learned about a user or records of previous outcomes. Memory is usually implemented in external storage rather than inside the model.

## Why it matters

     Without memory, an agent repeats questions, loses track of multi-step tasks and cannot learn from earlier outcomes. With poorly designed memory, it may retain personal data too long, recall outdated facts or carry one user’s context into another’s session.

     Memory design is therefore a data governance question as well as an engineering one: what is stored, for how long, who can see it, and how it is corrected or deleted.

## In practice

     For example, a B2B software company’s support agent might keep a short-term record of the current ticket, a longer-term summary of each customer account’s configuration and known issues, and no personal details beyond the ticket’s retention period.

## Where Rodan fits

     Rodan designs memory, state and retention for agents built through [Applied AI Engineering](https://rodan.io/applied-ai-engineering), with the same access controls as the systems the agent works with.

## Related terms

- [AI agent](/glossary#ai-agent)

- [Context window](/glossary#context-window)

- [Vector database](/glossary#vector-database)

- [Data retention](/glossary/data-retention)

- [Agentic workflow](/glossary/agentic-workflow)
HTML: https://rodan.io/glossary/agent-memory