It is possible to achieve using m//g.

my $line = 'This line has a hit here and a hit there.';
my $word = 'hit';

my $count = 0;
my $hilit = '';
while ($line =~ /(.*?)(?:\b($word)\b|\z)/sgi) {
   $hilit .= $1;
   if (defined($2)) {
      ++$count;
      $hilit .= "[$2]";
   }
}

print "$hilit\n";
print "$count occurrences of $word\n";
[download]

/c would indeed allow you to simplify the above.

my $line = 'This line has a hit here and a hit there.';
my $word = 'hit';

my $count = 0;
my $hilit = '';
while ($line =~ /(.*?)\b($word)\b/sgci) {
   $hilit .= "$1[$2]";
   ++$count;
}
$hilit .= substr($line, pos($line));

print "$hilit\n";
print "$count occurrences of $word\n";
[download]

But s///g is much simpler.

my $line = 'This line has a hit here and a hit there.';
my $word = 'hit';

my $count = (my $hilit = $line) =~ s/\b($word)\b/[$1]/gi;

print "$hilit\n";
print "$count occurrences of $word\n";
[download]

Note that if $word can contain characters other than those matched by \w, \b may fail and the contents may be treated as a regex instructions (e.g. $word="foo.bar" would match foolbar).

Update: Fixed a bug in first snippet.