I haven’t done this in a while, but IIRC there is a program called SubRip that rips the subtitles off a DVD then creates text based titles using OCR. There will be some text errors to correct, but may be better than doing them one by one. At this point in the process you will have to get the text into an XML format. Now I would look at a program like Annotation Edit which might be able to import the SubRip file and then export an XML.
Not sure if this will work, but worth a look see if you still need the workflow.