This is a good example of how "percent employees of underrepresented groups" isn't a very useful metric. They're counting all US jobs in that category, and page

This is a good example of how "percent employees of underrepresented groups" isn't a very useful metric. They're counting all US jobs in that category, and page 211 of the report shows that it's only 67% URM for service workers, laborers, and "operatives". Once you hit professionals it's 80% White or Asian.