Thank you all again for the wealth of information. :-)

I used Data::Dumper to examine the string contents.
You can see that the string is in Perls internal UTF-8 format if the umlauts, accents etc are stored as 1-byte code points (left column in this table: https://www.utf8-chartable.de/ ) and not in the 2-byte UTF-8 format with C2/C3 as first byte.

is_utf8 seems to me just an indicator flag that Perl has either created a variable natively in UTF-8, or it has done a conversion (decode).

It does not mean that the data is actually in Perls native format! One of the servers I use returns data which has to be decode()-d twice to become valid UTF-8 data. So after the first decode() is_utf8 returns true, even though the data still needs another decode() run to become ungarbled. Just this to emphasize the fact that one should not blindly rely on what is_utf8 says.