Document Intelligence Auto-Extraction Not Triggering on Attachment Upload

[Mohamed_009]

Hello, 

I have created a Document Intelligence use case for demo purposes. However, I am facing an issue where the fields are not auto-populating after attaching the document.

Currently, the extracted values are populated only when I manually open the Document Intelligence Workspace, select the fields, and submit the extraction. This manual intervention is required every time, which is not the expected behavior for the demo.

Additionally, in the Document Task record, the status remains in “New” state, and the Extracted Values section is empty, even after the document is attached. The values appear only after I manually perform the extraction and submit it from the workspace.

I would like to understand:

• What could be the possible reasons for this issue?

• Why the extraction is not triggering automatically on document upload?

• Why the Document Task status is not moving from New to Processed/Completed automatically?

• What configurations or steps should be verified to enable auto-population without manual action?

Any troubleshooting guidance or best practices would be really appreciated.

[Matthew_13]

Hi There
 Auto-extraction isn’t triggering because auto-extract/auto-submit isn’t enabled, the task isn’t linked to the model, or the trigger flow isn’t running. Manual extraction works because the workspace forces it. Let me know if this helps!

[Mohamed_009]

@Matthew_13 

Thanks for your reply. 

Just to confirm, are there any specific system properties or configurations that control auto-extract or auto-submit behavior?

Also, could you please share any recommended troubleshooting steps to verify:

• Auto-extract / auto-submit settings

• Model-to-task linkage

• Trigger flow execution or conditions

This would really help in validating the setup and identifying why auto-extraction isn’t triggering in our case. Thanks in advance!

[Matthew_13 · Accepted solution]

Just to confirm, there isn’t a single system property that turns auto-extract or auto-submit on or off. It’s usually controlled by a combination of the model setup, how it’s linked to the record, and any flows that trigger the extraction.

To troubleshoot, I’d suggest checking:

• That the model is published and auto-extract / auto-submit are enabled at the model level

• The model is linked to the correct table and matches how the document is being attached

• Any flows or triggers are active and their conditions are being met when the attachment is added

• The processing or flow logs to see whether the model is running or being skipped

These checks usually help narrow down why auto-extraction isn’t firing in a specific case. I Hope this helps!