| Topic | Details |
|---|---|
| Best for | Keeping the Knowledge Base trustworthy instead of noisy or half-processed. |
| Start here | Check scope first, then status, then retry if needed. |
| Outcome | Knowledge entries that retrieve more reliably later. |
| Scope | What it affects |
|---|---|
| Chat + Image | Both answers and visuals. |
| Chat only | Planning, answers, and text-heavy workflows. |
| Image only | The visual side of output. |

A healthy knowledge base is not just a long list. It is a list you can still trust later.
| Status | What it means | What to do |
|---|---|---|
| Processing | The source was accepted and is still being prepared. | Wait unless it appears stuck for longer than expected. |
| Ready | The source is available for later retrieval. | Leave it alone unless the content itself is weak. |
| Failed | The ingestion did not finish correctly. | Retry first. Rebuild only if the retry keeps failing. |

Upload flows are the ones most likely to need patience, status checks, and occasional retries.

Keep the header in view while you confirm which brand and knowledge area you are updating.

The roadmap is a useful checkpoint when you are deciding what still needs attention after an import or interview.

Pasted notes are another common source type, especially when a retry means cleaning up the source text first.