Virtual Personal Assistant

The Background

A question I've been curious about is whether there is a way to create a successful virtual personal assistant that pops up on screen. There have been multiple attempts that did not work. For example, Clippy, the Microsoft helper, and Apple's Knowledge Navigator concept from 1987. There have been successful personal assistants recently, like Siri and Alexa, but they are only voices in the air with no tangible facial or body features. My goal is to figure out if people would be open to a figure popping up on their screens while flying and their impressions of the face and bird icons I have created and a text only test. The screens these are on are part of my AstraBoard concept, which can be seen under the same name under Interactive Media II.

The methods of research that were used to test these characters were the Directed Storytelling, Elito, Semantic Differential, and Desirability Testing methods listed in Bella Martin's and Bruce Hanington's Universal Methods of Design.

Test 1


Test 1 was a face on screen with text for when it speaks. Interviewees were uneasy with it as it seemed to be listening in on them.

Test 2

Blue Jay

Test 2 contains a blue jay character, to test if an animal would be better received when it pops up on the screen.

Test 3

Text Only

Test 3 was performed using text only on the screen, which would have an accompanying voice but no icon.


Interviews resulted in half of the participants liking the text only screen better because it was the least distracting and didn't seem like it was watching them like the characters do, and half for the characters due to relatability. 

Solution:: Have character or text only as preferences for the AstraBoard, and the ability to turn them on or off, because it all depends on a person's personal preference. Some people are comfortable with virtual personal assistants, and some aren't, so it should be their choice whether they have them on their screens or not.