Does OCR make sense for digitally generated PDFs?

Scanned PDF files usually consist of one raster image for each page. The OCR engine can recognize the text in this image and make the document searchable. But what about digitally generated documents?