What worked well?
What worked well was that Alexa heard what I said most of the time. As long as I spoke in a clear, not super quiet tone she was able to answer or repeat what I said. Also when you asked generated questions like “what is the weather?” Alexa has a whole programmed in depth answer. Asking other questions like definitions for words or some history questions she was able to say aloud all the information she had about this subject. She responds very well to quick commands and simple questions.
What was difficult to use?
I personally did not find many difficulties in using the product because it is pretty straight forward with just saying the commands to Alexa. I did ask Alexa to give me a recipe for baking a cookie but it is hard to follow a recipe from just hearing it. I feel you can get lost very easily and make a mistake while trying to use this feature. Alexa also struggled with finding certain music to play or only my dads music since Alexa was hooked up to his phone. But overall the basic commands of asking Alexa and understanding what she said was simple and the users were able to use it successfully.
What were some unexpected behaviors you observed?
My human test subjects were actually having more fun than I thought interacting with Alexa. I timed them for thirty minutes to see how their interactions would change over time but they seemed interested in what kind of responses they could get out of Alexa. At first the users were just asking anything to see what she said but then it became more of a game. I also was surprised in some of the responses Alexa seemed to know. We had to look up a lot of them but the program has more of a personality then I thought. She was able to answer personal questions and at a basic level interact with the users and have very simple conversations.
What about the experience was exceptionally flawed?
One of the flaws the exceriance had was a lack of personal interaction. Alexa works fine when asking her to do commands but anything more than that was not feasible. Alexa can also not hold conversations or when asking long phrases she was easily confused in what you wanted. Alexa also misunderstood some of the time in what the users wanted. The user asked where there were places to eat and Alexa redirected somewhere in a different state. Then when the user added more information and tried again the Alexa was still confused. The user found it faster to just look up the information instead of asking a couple times to get the information they wanted. The music system is a bit flawed with only redirecting to amazon music or just to the main users music account. We tend to use other platforms for music but as a portable speaker it would be nice to have simpler and quicker actions to connect to other music sources.
What about the voice user experience could be improved?
An option that would help better the experience is having the ability to connect to Alexa by bluetooth on multiple devices. This would allow users to connect and control music preferences on different streaming platforms or just to have more control over it in general. This also would allow users to connect and disconnect as they please. What also could be improved is some of Alexas responses. Alexa sometimes gives too much information or not enough so there needs to be a decent balance. Users want to hear what they need to know but also other information without getting bored. Last thing to be improved is to make it smoother with conversations. We still need to improve our voice interface platforms to have a little more personality and less like a robot that just sits in the corner of the room.
Usability test video
In this video you can see how the users interact with Alexa over an half an hour. I wanted to see if a longer time would change overtime on how the users interacted with the device. It was interesting in seeing what kind of information the users wanted and how Alexa responded.