The export of an OCR'ed document contains not just the values for each bounding box, but also the bounding box's x,y, coordinates, width, and height. This can be an extremely useful feature when dealing with structured forms / documents where the location of text contains valuable information.