I don't know anything about Prolog, but it should certainly be easy enough call a Processing application (exported sketch), which picks up the xml file and visualizes it.  I think we may need a little more information though to help you further.  Where are you stuck?

If writing the file is quick, it can be just a matter of looking at the target directory from time to time, and see if a new file appeared.

You could approach this in a few ways.  Here are some questions though.  Is the file name the same every time or the only one of it's kind in the written directory (something that the sketch could always look for and find)?  Is this file / directory always in the same place (again, so the sketch can always look "here" and find it)?

* You could have the Processing app launch the Prolog script, which upon completion, it reads and visualizes it.
* You could have the Prolog script launch the Processing app via command line, upon which it reads and visualizes the file.
* You could write a batch / shell file that launches the Prolog script, then upon completion, launches the Processing app.