Can ChatGPT do careful and thorough read of pdf and word docs?

I’m trying to get pro ChatGPT to do something and it is failing miserably… wondered if anyone can suggest how to get it to work.

I have ten CVs in pdf/word, and a spreadsheet with each of ten people listed. Files are zipped and named to link to sheet.

I want ChatGPT to read them and fill in two columns with “organization” and “job” to list what everyone currently does. This is a test batch and ideally I’d like it to rattle through about 400 for me.

It doesn’t work no matter what I try… it “thinks” a lot, but either leaves stuff blank, fills it incorrectly with other data from the sheet (there’s already a column with some other organizations listed against people, eg where they went to university, and it just copies that).

It has now admitted that it can’t parse the text from the files – saying it thinks they are scanned pdfs (not true) or its word reader isn’t working.

Is this a setting problem? Can anyone suggest?

It would be incredible if I could automate this deeply boring task!

Leave a Reply