Research process

This was an exciting problem to take on, as such a system would allow users to incorporate non-verbal information into a conversation with a virtual assistant. However, users have never used an in-car voice assistant with eye-tracking capabilities, so foundational groundwork needed to be done to understand how and why users might interact with such a system.

Fortunately, our technical team in Germany had created a life-sized prototype using a highly immersive video projector that wrapped around a parked vehicle. My team ran studies to gather initial feedback and test how the system might operate.

However, I am located in Montreal and didn’t have access to a working prototype, so I needed to be creative with how I collected feedback. My solution was to host an online study where I had participants watch a video from behind a car dashboard and speak freely about what they saw. I was careful to provide a vague introduction to the product idea, without biasing how they might use it by making use of metaphors to real world situations without the use of technology.

By transcribing their audio, I discovered what felt natural for users to try at first. For example, many of them were trying to describe the buildings using visual reference by asking questions like “What’s that yellow building on the corner.” By observing patterns in user’s initial interactions with the system, I found discrepancies in our expectations and technical abilities.

Through my initial experiments with prototypes, I was able to gather research questions from which I could base my first design explorations.