Microsoft Speech UI

2005 | Microsoft's Speech User Interface.

The Windows Vista Speech UI project I was part of was one of the most interesting UI challenges I had the opportunity to work on while at Microsoft. We did intensive usability research to come up with most of the design decisions. Given that humans are accustomed to utilizing speech as a form of communication with other human beings who understand contextual information, the problem of speech as a User Interface is not a simple one to solve. Repeatedly, we saw our designs fail due to a lack of clarity in the user's mind on what constitutes a proper command. The use of a larger grammar set for commands including synonyms allows users to have a more natural communication style but at the same time it causes more errors because the system's performance decreases as more terms are added to the grammar set. Users tend to have very little tolerance for mistakes when giving speech commands. They expect the command to work and they experience a lot of frustration when this is not the case. This means that instead of continuing to learn how to use speech, they give up and go back to using the mouse and keyboard.

Because of these challenges we decided that for the time being users are better off learning how to use the right commands in speech. At the moment, only users who can't use or have a limited ability to use the mouse and/or keyboard tend to rely on speech and they are more willing to spend time learning how to use a speech user interface. Our First Time User Experience instruction made it easier for users to learn and remember the proper commands. We spent a lot of time and effort designing a simple and easy to use instructional First Time Use application following the classic user centered cycle of design, prototyping, user testing and design iteration.