Using a Speech Technology Agent as an Interface for E-Commerce

Abstract

This paper outlines an experiment set out to determine the usefulness of a Speech Technology Interface Agent for E-Commerce transaction tasks. There are a limited number of speech technology agents in use on the Web at present and there are none used in the E-Commerce scenario that combine both speech synthesis and speech recognition. The results of this experiment indicate that users would find such speech technology agents acceptable and useful for non-trivial tasks but that there is a need for technical improvements with regard to implementation and also a greater assessment of optimal task division between speech and traditional hyperlink navigation modalities. The latter has implications for all agent interaction on the Web as user acceptance and the usability of interface agent interaction is given little priority in the current research literature.