One of the more surprising aspects of developing voice-first applications has been just how easy it is to inadvertently induce anxiety in your user.
As an example, one of our earlier MVPs was a voice-first Gmail client that let you respond to any email you received via a push-to-talk interface. Our target customer was the sales executive who spent a lot of time in their car travelling between meetings.
As the majority of our users would be driving we decided to always reduce the concentration required to respond to their emails. So after recording our user’s response we would just display a countdown timer before automatically sending the email. Less concentration required – perfect!
However, after a couple of days trying to change ourselves we had our most unified negative response to a UX update.
We each reported unhappiness with the quality of the emails we were recording. Watching the time counting down had forced us to stop thinking about our responses and solely on how quickly we were saying it. One seemingly useful UX change had inadvertently made us all feel anxious that we weren’t thinking fast enough.
While we have since moved on from building that particular MVP, this experience has formed one of our fundamental voice design principles: Always give users the time to think about what they are saying and their communication quality will improve.