I would say that NIEM's alignment between XML and RDF is *very* important for use of the data. 

XML and XML Schema don't address a lot of issues fundamental to understanding data. What does a block of XML mean? What does type extension mean? What does an element containing another element mean? By defining the interpretation of NIEM data based on RDF, we get a real semantic model that explains a lot about the meaning of any given block of data. 

But a lot of people don't care about that level of detail about the meaning of data — it's too philosophical, too esoteric.

For them, the XML data looks like a straightforward use of XML - elements with sub-elements, types with base types, IDREFs linking to IDs - all clearly named and not too hard to understand. 

However, the rigor that the XML–RDF alignment provided helped to ensure that a lot of things were done in a consistent manner across a very large number of data definitions. And that diligence helps make a big pile of data understandable. The alignment to RDF benefits everyone who uses it, even those who don't care at all about RDF.