To delete a specific agent memory, target the stored source and its derived artifacts, not just the next prompt. Find the delete handle, disable or remove summaries and embeddings, record an audit event, rebuild affected retrieval indexes, and verify the old fact no longer returns in a fresh agent question.
Last updated: June 20, 2026.
Why is “forget this” not enough?
Telling an agent “forget that” can influence the current conversation, but it does not necessarily remove stored artifacts. If the memory exists as a source item, summary, embedding, tag, or retrieval candidate, the next session may retrieve it again.
The GDPR right to erasure is about erasure under qualifying circumstances, not polite model behavior. Even outside GDPR scope, users expect deletion to affect storage and retrieval.
What do you need before deleting?
You need a delete handle. The handle should identify the source artifact and the derived artifacts created from it:
| Artifact | Why it matters |
|---|---|
| Source document or transcript turn | Original evidence behind the memory |
| Summary | May still expose the deleted fact |
| Embedding | May retrieve the deleted fact by similarity |
| Tags/entities | May route future retrieval toward the fact |
| Retrieval logs | May need retention policy or redaction |
| Supersession edges | May need a replacement rather than deletion |
Provable Forgetting explains the broader lifecycle. This post is the operational checklist.
How do you inspect the memory before deleting?
Start with source lineage. Ask:
- Which source created the memory?
- Was the memory manually saved or extracted from a transcript?
- Which derived artifacts exist?
- Which library and tenant can retrieve it?
- Was it already superseded?
- Does the answer cite it directly?
The NIST AI Risk Management Framework is useful as a governance reference because deletion is a risk-control process, not just a UI action.
What is the safe deletion sequence?
Use this order:
- Freeze writes for the target memory while deleting.
- Mark the source artifact deleted or disabled.
- Delete or tombstone derived summaries, embeddings, and tags.
- Remove the memory from active retrieval indexes.
- Write an audit event with actor, time, source id, and action.
- Run negative retrieval checks.
- Keep only retention-required logs under policy.
If the fact changed rather than needed removal, supersede it instead. Deleting history when the real need is current-state correction creates audit gaps.
How do you verify it worked?
Run four checks:
| Check | Expected result |
|---|---|
| Direct content lookup | Deleted source is unavailable or marked deleted |
| Full-text search | Deleted phrase does not return as active evidence |
| Vector or hybrid search | Similarity search does not return the deleted artifact |
| Agent answer | Agent says it cannot find active source-backed evidence |
If any check returns the old fact, inspect the derived artifacts. The bug is often an orphaned summary or embedding.
What if the user wants correction, not deletion?
Deletion is not the right response to every bad memory. If the underlying fact changed, create a current source-backed replacement and mark the old memory superseded. That preserves the audit trail while preventing the old fact from being treated as current evidence.
Use deletion when the memory should no longer be available at all: a user revokes consent, a sensitive artifact was imported by mistake, or a source must be removed from active retrieval. Use supersession when the memory is historical but no longer current.
What should your product expose?
A usable memory product should expose a deletion surface with stable handles, source lineage, affected derived artifacts, and verification status. A button that hides a memory from the UI but leaves its embedding active is not enough.
At minimum, expose:
| Control | Why it matters |
|---|---|
| Source id | Targets the original evidence |
| Derived artifact count | Shows what else must be removed |
| Actor and timestamp | Supports audit review |
| Retrieval verification | Proves the memory stopped returning |
| Undo or retention policy | Separates accidental deletion from policy erasure |
What should you read next?
Use the Memory Failure-Mode Catalog for the symptom map, then read Provable Forgetting and RAG leaks docs to wrong users for adjacent lifecycle and permission issues.