Given that in your examples the operands are in the character class [a-z] you could split on the boundary between operand and operator using look-behinds and look-aheads. You split either at a point preceded by the character class [a-z] and followed by the negated character class [^a-z] or vice versa, like this

use strict;
use warnings;

print
   map {qq{$_->[0] -- $_->[1] -- $_->[2]\n}}
   map
   {
       [
          split m
          {(?x)
             (?:
                (?<=[a-z])(?=[^a-z])
                |
                (?<=[^a-z])(?=[a-z])
             )
          } 
       ]
   }
   map {chomp; $_}
   <DATA>;

__END__
a=b
a!=b
a<b
a<=b
a>b
a>=b
[download]

which gives this output

a -- = -- b
a -- != -- b
a -- < -- b
a -- <= -- b
a -- > -- b
a -- >= -- b
[download]

Given a more complex set of operators you would probably be better off taking grinder's approach of setting up a regular expression that matches and captures any operator. As complexity increases a parser solution becomes more appropriate.

I hope this is of use.

Cheers,

JohnGG