Annotations are not converted to Word during the conversion process; only page content is converted. If you need to include annotations, you can flatten the PDF prior to converting. Flattening will "burn" the annotattions down onto the page.
The text may have been there already and was just under the image of the logo. To the best of my knowledge, no OCR is performd during conversion.
Grids are very hard to convert. My guess is that the grid is not vector based but an image and that the converter is favoring the text over the lines. If you can share the original PDF, I can probably elaborate.
Note: The purpose of the converter is to create an editable Word document from the PDF. It will not be able to preicely recreate the layout of the PDF because the two document models are completely different. Unfortunmately, this means that the conversion service won't be appropriate for all document types.