Published at: 2026-10-15
Knowledge Ingestion Rules
| Source | Configuration |
|---|---|
| Knowledge Articles | Map 1.0 articles to AI items |
| Network Disk Files | Sync folders under “Knowledge Base Files” |
| CRM Objects | Filter + field mapping; new data after enabling |
| Work Circle Sharing | Capture shares by topic |
| Training Courses | Sync courses/transcripts/materials |
Overview
Knowledge ingestion rules determine which sources can feed knowledge into the knowledge base and how each source is mapped. Configuring the right sources reduces manual entry and keeps the knowledge base in sync with existing materials.
Supported ingestion sources
| Source | What is ingested |
|---|---|
| Training assistant course | Course content from the training assistant is converted into knowledge. |
| Fenxiang network disk file | Files (Word, PDF, PPT, Excel and others) from the network disk are ingested as documents. |
| CRM object | Data records from CRM objects are ingested according to the configured field mapping. |
| Knowledge article | Knowledge articles from the legacy knowledge base are imported. |
| Work circle share | Content shared in the work circle is ingested as knowledge. |
Configure an ingestion source
- Open the knowledge ingestion rules in the backend.
- Choose the source type to configure.
- Set the scope and mapping (for example, which folders or categories the ingested content goes to, and which fields map to title and content).
- Enable the rule and test with one item.
- Verify that the ingested item appears in the target space and is searchable.
Ingestion behavior
- Ingested items inherit the permission scope of their target space or folder.
- Content is synchronized periodically or on demand depending on the source; schedule accordingly for frequently updated sources.
- Duplicate detection runs on ingested content; configure merge rules to avoid duplicates.
Legacy backend notes
- These ingestion rules are part of the legacy 1.0 backend and are mainly for enterprises with existing custom configurations.
- For standard Knowledge Base 2.0, use the front-end creation and import methods (online document, upload, AI drafts, system import, OpenAPI) instead.