Agreed, sometimes there are visual references for new rows such as a cell border (in excel language) or change in background colour. Other times it might be that one of the ‘columns’ is only present at the top of each data row therefore everything below it should be assumed to be the same row until there is another ‘hit’ in the designated column.

Might sounds fancy, but I imagine a table snipper that “learns” from snips on a document for the next snips on the document/ other documents in the worksheet. For instance, I have images of tables in which new rows are shown by bullet points, and then the lines that don’t have a bullet point are the continuation of the previous row. It would be helpful that the Table Snipper after a few correction of lines suggests me to consider this new rule.