is_utf8 does not say anything about the encoding of a string, just how it's stored internally.

$ perl -le'
   use Encode qw( encode is_utf8 );
   my $x = "\xC9ric";
   print is_utf8(encode("iso-8859-1", $x))?1:0;
'
0

$ perl -le'
   use Encode qw( encode is_utf8 );
   my $x = "\xC9ric";
   print is_utf8(encode("UTF-8", $x))?1:0;
'
0
[download]

You determine the charset of an HTML page by looking at the Content-Type header of the HTTP response.

$ perl -le'
   use LWP::UserAgent;
   print join " ",
      LWP::UserAgent->new
         ->get("http://www.google.ca/")->content_type;
'
text/html charset=ISO-8859-1

$ perl -le'
   use LWP::UserAgent;
   print join " ",
      LWP::UserAgent->new
         ->get("http://www.microsoft.com/")->content_type;
'
text/html charset=utf-8
[download]

Since this doesn't work when the HTML is stored in a file, the http-equiv tag was created.

$ perl -le'
   use LWP::UserAgent;
   my $html = LWP::UserAgent->new
      ->get("http://www.google.ca/")->decoded_content;
   print for $html =~ /(<meta[^>]*>)/ig;
'
<meta http-equiv="content-type" content="text/html; charset=ISO-8859-1
+">
[download]

That said, you don't need to do ANY of this. You just use HTTP::Response's ->decoded_content method and it will decode the content for you. Then, if you need it encoded, just use encode unconditionally. (Although you may need to adjust the http-equiv header...)