Hi GuiPerl, I haven't read thoroughly through your entire code, but I would suggest that this code:

for my $division (keys %AG) {
    # ...
    for my $person (@{ $AG{$division} }) {
        # ...
[download]

could possibly be made quite simpler (and also slightly faster, but that probably does not make a measurable difference) with the following syntax:

for my $division (values %AG) {
     # ...
     for my $person  (@$division) {
          # ...
}
[download]

The above is untested, but this is a simple demonstration of the same idea under the Perl debugger:

  DB<1> %AG = ( GAHD => [ { name => "foo",}, {name => "bar",} ], GAGD 
+=> [ { name => "baz",}, {name => "foobar",} ] );

  DB<2> x \%AG;
0  HASH(0x6005009f8)
   'GAGD' => ARRAY(0x600500680)
      0  HASH(0x600500740)
         'name' => 'baz'
      1  HASH(0x6005006e0)
         'name' => 'foobar'
   'GAHD' => ARRAY(0x6005007e8)
      0  HASH(0x60005f150)
         'name' => 'foo'
      1  HASH(0x600500848)
         'name' => 'bar'
  DB<3> for $division (values %AG) { for $person (@$division) {print $
+person->{name}, "\n"}};
baz
foobar
foo
bar
[download]

I hope this helps.