This entry will continue from the previous blog. The focus will be on the strategy employed to develop the prototype. As mentioned before, the prototype service is needed for validating the proposed architecture for building IVR systems. The development involved three major phases excluding the requirements analysis phase. These are:
- Building the service ontology – This ontology captures information about services in Grahamstown. The ontology answers questions such as: ‘where to get a free service?’ and ‘which non-governmental organisation offers X ?’
- Designing the VoiceXML application – This part of the work focused on the actual dialogs that constitute the service. Since IVRs are notorious for causing frustration to users, the goal was to attempt to create an application that provide users with a satisfactory experience 😉 .
- Implementing the whole system – The focus here was on bringing all elements together such that a user can call, put a request, have the ontology queried and finally have a response returned back to the user. Gaining “perfect” fluidity to this seemingly simple goal involved a number of things, which I shall not discuss in this blog (partly because I am tired but mostly because I don’t want focus to be removed from the value of employing a strategy that encourages divide-and-conquer).