Catsby!
Are you ready to hear what your cat really thinks about you?
About Cats Speak
Cats Speak lets cat owners communicate with their furry friends and answers any queries owners have about the emotions, diet, and activities of their cats. Physically, the Cats Speak device lives inside of this harness. But, Cats Speak uses a voice user interface (VUI), which makes spoken human interactions with computers possible. This device can be used in any environment, but will need access to wifi in order to use key features.
Catsby, the name of the voice assistant, allows owners to talk to their cats by translating every purr and meow to the owner. It will also translate every word into cat-speak to the owner’s cat. Owners can simply activate Catsby by stating its name and the device will start listening. Owners are required to be in the same room as the device in order for the device to respond to voice commands.
VUI Features
Catsby is enabled to do 5 main features: translate, check emotion, check diet, check activities, and list interaction details. Below is an example script for task 1: translation between owner and cat (Jessie).
Task 1. Translate:
Owner. Catsby, say “Hi, Jessie!”
Catsby. Okay, Jessie they said: *Purr Purr*
Cat. *does not respond*
Catsby. It seems that Jessie does not care.
Owner. Ouch, okay then.
Catsby. Sorry, I do not know what that means. Can you say that again?
Owner. I didn’t ask a question. Catsby, say “Hi, Jessie!”
Catsby. Okay, Jessie they said: *Purr Purr*
Cat. *Purrrr*
Catsby. Jessie said: “Hello, Mom.”
Full sample dialogs for all tasks can be viewed here.
Every task can be activated with a key word. If the query or command does not contain any of the key words that activates the tasks, Catsby will prompt the Owner to repeat the query or command before listing out possible tasks.
The key words are:
Task 1. Translate — “Tell/Ask/Say”
Task 2. Emotion — “Feeling”
Task 3. Diet— “Diet”
Task 4. Activity— “Doing”
Task 5. List Prompts— “What”
Purpose of Investigation
The purpose of this investigation is to measure the feasibility of creating a VUI that bridges the language barrier between the owner species and cat species. What are the benefits of using a VUI that screen-driven interactions cannot offer? For this method, I will focus on analyzing the owner’s interactions with the VUI.
The VUI script was tested with a cat owner in the environment of their home. As stated above, the VUI needs to be connected to wifi and the owner is required to be in the same room as the voice assistant, which will typically be somewhere at the owner’s home.
Dialog Flow Chart
For better readability, view the dialog diagram here.
Analysis
Prototype
The benefits of VUIs is undeniable — they are hands-free and eyes-free interactions. With Catsby, owners are able to better interact with their pets without having to worry about getting scratched or bit. Although Catsby needs wifi to operate, owners are not required to interact with their mobile devices in order to interact with their furry friends.
What worked well about the prototype were the key activation words. Overwhelmed by free form of VUI, I used key word activation as a VUI version of buttons you would click on to prompt actions. This made it very easy to treat unknown commands or non-questions as errors, which leads Catsby to prompt the Owners to try again or offer prompting information.
There are some drawbacks to this method. For one, people have different flows of conversation. Although the key words are common, they might not come up naturally in people’s dialog. Two, it assumed that any sentence or statement with those key words are prompting Catsby to activate respective features, which might cause accidental activations. And lastly, I was not able to explore the possibilities of a command or question that contains two or more of the key words.
Overall, the design of the Cats Speak’s VUI is effective. I would need to conduct more usability tests to better understand effectiveness of the key words. In the future, I would need to weigh the pros and cons of using more unique activation words to avoid accidental activations — or if I need to choose more common words to better fit them into conversation flows.
Process
The process of creating an effective VUI, even in a small scale project, was very challenging. Although the first deliverable only required 5 tasks, translating the sample dialog scripts into a dialog flow chart required time and care.
I started writing the sample dialog scripts with only tasks in mind. I followed this up with identifying overlaps in the sample scripts to avoid repetitive items for the flow chart. By doing this, I began to piece together a more efficient structure for the VUI, which enabled me to focus on incorporating confirmative messages in the dialog and an error message system.
In future iterations, I would tackle VUIs with a mixed approach of screen-driven and exploratory UI methods. Since, I thought of VUIs as entirely disconnected from screen-driven UI methods I am familiar with, I easily got overwhelmed by the infinite possibilities of designing VUIs and sidetracked by non-essential elements. Instead of focusing on all the places that can cause an error, I would focus on preventing the creation of those error hotspots by configuring a dialog pattern in the beginning.
