can one speed up a custom aggregate by using native ones

Greetings,

I have an aggregate function which basically concatenates a whole bunch of
rows (that would be the sfunc) and then wraps it in a header and footer
(the finalunc). It is however for fairly large set many orders of magnitude
slower than running string_agg and then the finalfunc.

Is there anyway to use the builtin functions say in the aggregate function
definition. Or any other way to speed them up?

Regards,

Rhys
Peace & Love|Live Long & Prosper