A Model for the Development of Speech Applications

Speech recognition has matured over the past years to yield acceptable recognition rates. While analyzing the scientific and IT-related literature one could get the impression that speech applications now get into mainstream. The technology seems technically mature for widespread adoption, from call-center automation up to ambient intelligence environments. When digging deeper into the domain of speech applications and their deployment to business application one might puzzle, why the acceptance is not as high as expected and which technical, human-computer interaction, and business aspects must be considered to get a well designed speech application for a specific domain to gain superior user experience. This paper presents a model for architecting speech applications to be more efficient in regard of development time and to increase user acceptance.