PDF enabled but not indexed

I would agree I am not seeing any generative replies in the help center or with advanced AI picking up data from the PDFs

Hi ​@Tobias11 & ​@Sean Morrissey, this is Daniel from Zendesk’s product team. As a first triage, can I check if you updated your help center search sources to include your Sharepoint/Google Drive?

You need to go to Knowledge Admin > Settings  > Search Settings > Search Sources and check if the sources related to PDF files are configured. 

Additionally, can you confirm if the PDFs you’re looking to ingest are text or scans?

Thanks!

Hi Daniel,

you are 2 steps to far for me. 

1.) We using it exclusive internally in Knowledge Capture (Side-Bar App) which has the source selected by selecting external source from here (no plan to move the crawler result to helpcenter yet).

2.) It´s kind of a test Server 3 files XLS, 3 files Doc, 3 files PDF

I can see in Result page total items = 6, which already show the 3 PDF are not part of.

Thanks.

Tobias

Hi ​@Tobias11, just to confirm, in your Connections dashboard (Knowledge Admin > Content > External Content) what is the count of items in the Sharepoint connection? 6 or 9?

Thanks,

Daniel

Hi Daniel,

the scenario WAS showing count 6, not 9 so it just “not picked the PDF”.

I repeat the test today and it works now, so for me, no issue again (last test was done 1-2 weeks ago).

I´ll monitor it.

Thanks,

Tobias

Hi ​@DanielPinho,

I dig deeper into our issue, and found following things.

1.) Some Agents uploaded different content (png, rar, zip, pptx) which seems instead of “skipping” it may abort the crawler, can´t find other (previously indexed) files (pdf, doc) in the folder structure, just say “Sync Success” but in the end “not complete” or abort due to whatever issue. (This is hard to reprodcue with on-the fly adding content, no log in Zendesk GUI about “skipped items” or any log about which items are susccessfully migrate, the AI Agent provide much better visibility at this stage.

2.) We have some PDF which are 700 pages long (User Guide for our products) seems the crawler have issue with this document, is it design limitation or should it work? (similar import in AI Agent Advanced is working with this crawler). I can send you such PDF via E-Mail if you like to test out, but I can clearly see it´s not indexed, even I click sync again it didn´t pick it and even if I remove all other content (not supported files) or put it in Root Folder it´s not indexed.

I believe this is why we see some challenges here.

Kind Regards,

Tobias

 

I did not have that enabled.