The WebChatAgent Data Sources interface explained
Know what every source type, status and management action means before changing the knowledge base.
Data Sources is the control center for knowledge used by retrieval. Website, file, text and connected-workspace sources share a list but have different editing and refresh behavior.
Status is operational: pending means wait, indexed means usable, error needs inspection and re-indexing required means no current embeddings are available.
What you will have at the end
- A clear choice between all source types
- A repeatable status and quota review
- Safe edit, re-index and delete procedures
Before you start
- A chatbot with at least one source or permission to add one
- A list of content owners and update frequencies
Source lifecycle from input to retrieval
Adding a source starts extraction and indexing. The resulting passages count toward the character-based content quota and become available only after a successful status.
Editing scope or changing provider can require a new index. Deleting a source removes its knowledge from future retrieval.
01–05
Set it up step by step
Review all source types
Choose Website, Documents, Text, Confluence or Notion by ownership and refresh need.
Open Add Data Source and compare the choices. Use Website for public pages, Documents for supported files, Text for focused facts, and native connectors for approved Confluence or Notion content that should synchronize.
Read source status correctly
Do not test before indexing completes.
Distinguish Indexed, Pending, Error and Re-indexing required. Open source details when an error occurs and re-index immediately after a provider switch.
Use search, type filters and categories
Find the correct source in a large knowledge base.
Search by name or URL, filter by source type and assign stable categories such as Products, Support or Legal. Categories should express ownership or topic, not temporary status.
Inspect quota and edit settings
Control scope before quota becomes a problem.
Review indexed characters and remove duplicates before upgrading. Website sources can change crawl settings; text can be edited; files need re-upload for new content; connectors can change selected spaces/databases and sync interval.
Re-index or delete safely
Refresh changing content and remove obsolete versions.
Use re-index after source changes and wait for completion before testing. Before deletion confirm no unique required facts live only in that source; then delete and run a regression question that previously used it.
Example & result
See the practical test and its result
Every tutorial includes a fixed input, the expected outcome and a transparent record of what was actually verified locally.
Practical example: The WebChatAgent Data Sources interface explained
This exact scenario was completed with the temporary tutorial account.
Exact test input
Filter for “tutorial-demo” and verify website and PDF status.
Expected result
Both sources remain visible and show Completed.
What was actually verified
The real source table showed the website and chatbot-training-demo-knowledge-base.pdf as Completed.
Tips & tricks
Make the setup reliable
Test with realistic examples, record your baseline and change one setting at a time. That makes real improvements visible.
Index one language per duplicate page set
The chatbot can reply in the visitor language; indexing translated duplicates wastes quota and may reduce retrieval precision.
Assign an owner to every source
Ownership makes outdated policies and abandoned connectors easier to detect and remove.
When something does not work
Troubleshooting
Check status, permissions and test data systematically before changing the model or prompt.
The expected option is missing
Confirm the account plan, feature permissions and selected chatbot. Paid or beta features can be hidden when prerequisites are not met.
The test result is inconsistent
Reset the test conversation, keep the input identical and change one setting at a time so the cause remains measurable.
Ready for a production-style test
Create a quarterly source inventory with owner, category, refresh method and one verification question for every critical source.