There is a new feature in AI Builder's 'Create text with GPT using a prompt' to be able to take a PDF document or image as the input to extract the information out of it. It is currently in preview mode.
I was able to create a model and test it. But when I put it into a flow having the document coming from sharepoint, I am getting an error. Please see the screenshot. The Copilot suggested that there is something wrong with the file format, but I can use the same file in the model and tested it successfully. Do I need to process the PDF document into a certain format ?