Power Platform Community Forum Thread Details

Plain Text

Hi everyone,

I want to build the following process:

Power Pages user

→ uploads a PDF (text, images, charts, graphs)

→ file is stored in Dataverse (Timeline Note or File Column linked to a project)

→ AI Builder Prompt ("Image or document input") analyzes the document

→ extracted information is written into specific Dataverse fields/tables based on AI classification.

I already confirmed that the Prompt works when a PDF is passed directly through a manual Power Automate file input.

My challenge is passing the uploaded document from Power Pages / Dataverse to the Prompt without running into issues like:

"Unable to identify the mimetype input"

Has anyone successfully implemented this architecture?

Which storage method worked best:

- Dataverse File Column?

- Timeline Note (Annotation)?

- SharePoint file + Get File Content?

Thanks!

Categories:

Power Pages PDF Upload -> AI Builder Prompt (Image or Document Input) -> Dataverse Architecture