As for the CSV files, perhaps adding a BOM would help?  AFAIK, it is widely used on the Windows platform (even with UTF-8) to both indicate that the files do have Unicode content, and to specify the particular encoding being used (UTF-8, UTF-16le, etc.).

For this, the first thing you write to the file should be the BOM (\x{feff}):

my $fname = 'output.csv';
open my $fh, ">:utf8", $fname or die "couldn't open '$fname': $!";
print $fh "\x{feff}";
...
[download]