The Schützenberger methodology can enumerate the number of words of given length $k$ for unambiguous context free languages by a generating function. But not all regular languages are unambiguous context free languages, so the Schützenberger methodology can not be applied to them.
My question: Is there a procedure to give for every regular language a generating function that enumerates the number of words of length $k$ of this language?