Accordingly, there is a need for a more convenient and thorough process for building a data base of subword transcriptions and associated telephone numbers, and more particularly, for a process of building a data base using speaker independent templates.