The conversion process begins by finding and loading an XML document into the Sous-processus Arbortext PE. If the input document is not XML, it is imported from its native format to XML, and the resulting XML document is loaded.