Power Platform Community Forum Thread Details

Hi all,

I would like to ask why copilot cannot read PDF or images. I tried searching on the web if it can read letter from the image or pdf.

I try to create topic and create generative response based on the knowledge that is sourced from image or pdf, but it can't retrieve any information.

Please let me know if there is any solution to this.

Thank you

Categories:

General topics

Copilot studio could read "real" PDF : to understand : PDF generate with low code pdf generator often are just low quality image inside a pdf.

Copilot studio has, at the moment, no OCR capabilities. But PDF, real one, work very well with copilot studio and are the main use of it. When PDF don't work it's often because they are not well created one.

About image : it depend, by design it's not supposed to be readed since no OCR capacities.

BUT instead of create you'r knowledge in sharepoint with the big button sharepoint, just above, in the middle of the screen of adding knowledge there is a small sharepoint one, try this one : it allow advanced index with image and pdf , but with limitation (32mo pdf, and 200 files per source)

But to answer first question : PDF are supported.

Other reason to not work :

- they are to deep in you'r sharepoint tree (10 nested folder is really to deep : where a human canno't go, a bot cannot too)

- just wait : you content wasnt index : did you have the green check for ready for your source ?