Memory Triage for AI Agents

An AI agent that remembers everything is not smart - it is cluttered. When every fact gets the same priority, the agent spends tokens retrieving outdated preferences, stale project contexts, and one-time corrections that no longer apply.

The fix is not better storage. It is intentional forgetting.

The 100% Retention Problem

Most agent memory systems are append-only. Every interaction adds new facts, corrections, and preferences. Nothing gets removed. Over weeks of use, the memory grows into a sprawling collection where critical instructions sit next to trivial observations.

When the agent loads context at session start, it has to process all of it. A memory file that was 2KB in week one is 50KB by month three. The signal-to-noise ratio drops continuously.

How Memory Triage Works

Instead of retaining everything, a triage system ranks each memory by two factors:

1. Access frequency - how often does this fact get retrieved and used in actual tasks?
2. Semantic relevance - how connected is this fact to the agent's current active projects and goals?

Memories that score high on both dimensions stay at full priority. Memories that score low on both are candidates for decay.

Natural Decay Is Not Data Loss

Letting memories decay does not mean deleting them permanently. It means moving them out of the active context window. A decayed memory still exists in storage - it just is not loaded by default. If it becomes relevant again, it can be retrieved on demand.

In practice, a well-tuned triage layer shows about a 57% decay rate for memories older than 30 days. That means more than half of what an agent "learns" in a given month is not relevant enough to keep in active memory.

Implementing Memory Triage

A practical approach:

• Tag memories with timestamps and access counts at creation time
• Run a weekly triage pass that scores all memories against current project context
• Archive low-scoring memories to a separate file that is not loaded by default
• Keep a retrieval index so archived memories can be found when specifically needed
• Set hard caps on active memory size to force prioritization

The Right Amount of Forgetting

Human memory decays for a reason - it keeps the most relevant information accessible. AI agent memory should work the same way. Perfect recall is not the goal. Useful recall is.

Fazm is an open source macOS AI agent. Open source on GitHub.

More on This Topic

• Agent Persistent Memory and Knowledge Graphs
• Long-Term Memory Separates Toy from Useful Agents
• Embeddings vs Tokens in Agent Memory