Exploration and assessment of proactive use cases for an in-car voice assistant

Authors: Maria Schmidt, Daniela Stier, Steffen Werner, Wolfgang Minker

Abstract:

More and more people tend to use voice assistants on their smartphones,at home or in the car. These assistants help people to find their way in a foreigncity, call their friends or add items to the shopping list – mostly in a user-initiatedmanner. Whereas there have been several scientific publications on adaptive andproactive Spoken Dialog Systems (SDSs) in the past decade (e.g., [1], [2]), mostof these primarily focused on the technical implementation, but none of them dealtwith basic research on which proactive use cases a user of the respective assistantmight actually wish for. In order to compensate for this, we explore and assesspotentially suitable proactive use cases in this work as follows. First, we conducteda creative workshop in which expert teams brainstormed, rated, selected, elaboratedand finally presented eight proactive use cases in a role play manner to internalstudy subjects. Second, according to the previous subjects’ ratings, we selectedand elaborated four primarily proactive use cases in order to present them to 20 external subjects in an interview study in a non-moving car. All of the latter usecases achieved high acceptance rates, whileProactive Parkingboth got the highestacceptance and the highest additional value score.