Enhancing existing form-based dialogue managers with reasoning capabilities

In this paper we present an approach to complement speech dialogue applications written in VoiceXML with reasoning capabilities simplifying the integration of domain knowledge into the traditionally coded dialogue flow. One of the characteristic enhancements of our approach is the use of a dedicated reasoning component for processing the knowledge items acquired during the dialogue. The approach is also novel in that, instead of building on top of a specialized research dialogue architecture, an industry-standard dialogue description language has been chosen as the basis. This enables enhancing existing voice application rather than starting from scratch. The paper describes a first implementation of the proposed interface between the traditional form-based dialogue manager and the reasoning module and includes the analysis of a sample dialogue conducted with a prototype version of the system.