In this paper we present language technology enhancements to audio-based information services (i.e. services where information is presented using spoken language). The enhancements presented in the paper addresses two issues for audio-based services: 1) interaction with the service is rigid and 2) the ability to listen to summaries is limited. Our developments allow for more natural and efficient control of the service and means that facilitates skim reading. Using speech dialogue instead of traditional buttons provides means for more advanced navigation in the audio material. Vector space techniques are used to collect the most relevant sentences in a text and allows for skim reading of varying depth.