Abstract
This paper describes the results of an experiment in which the behaviors of callers to two different natural language prompts in a simple voice user interface were measured. Callers into a call center heard one of two natural language prompts. In one of the prompts, the interface identified itself to the caller by name (“Hi, I'm Bill”) while the other prompt did not. Results show that this simple personification led to significantly better user performance in several areas. First, more callers interacted with the system if the interface identified itself by name. Second, callers provided more complete statements of what they needed to the system. Third, callers generated fewer vague statements that would have required the system to re-prompt them for more information. Finally, callers who heard the named-prompt uttered more words on average than those who heard the un-named prompt. Implications for voice user interface design are discussed.
Get full access to this article
View all access options for this article.