A team of four design students and I redesigned Dragon Mobile Assistant, a popular NUI personal assistant for Android. We worked closely with professional virtual assistant designers to receive mentorship throughout the process.
Expert evaluations and task analysis
Analysis of Google Play user reviews
Focus group discussions
Analysis of survey data
Independently-run survey of 200+ mobile assistant users
- 50+ features
- Divided into Must Have, Should Have, and Nice-to-Have Features
- Concepts developed based on background research and features
What do you like about your current mobile assistant?
- Voice recognition
- Hands-free capabilities
What do you dislike about your current mobile assistant?
- Voice recognition
Most frequent uses:
- Voice calling
- Feeling of segreation
- Often forgotten
- Need to be “open”
- Poor app integration
- Robotic conversations
- Microphone icons
- Famous for being wrong
In-car operation & safety
Feedback while speaking
- Text message readback
- Correct speaking errors
List of available features
Our prototype goes by the codename Tamagotchi- referencing the personality that came with the toy we all played with in the 90's.
Tamagotchi will be accessible as a chat head on every screen.
In most mobile assistants, when a user requests to access another app (such as Spotify), the assistant brings them to the new app and then no longer be available.
With this chat head model, Tamagotchi will be with users every step of the way and allows for seamless app transitions.
Tamagotchi takes into account user patterns.
If the user frequently texts Joe on Facebook Messenger, that will be where Tamagotchi crafts a new text when text platform isn’t specified.
But, if the user decides to stray from the norm, he can easily correct the assistant.
Tamagotchi allows users to customize faces and voices -- making it a truly personal assistant.
So far, this page only documents half of our design process :) The second half of our design process included:
- Usability tests
- Low - medium fidelity prototypes
- Summative testing
- High fidelity prototype
- Formative testing
- High fidelity prototype iterations