PACIFY
- Conversation Design
OVERVIEW
Pacify is a voice user interface that helps users manage and relieve their stress on a daily basis. For this project, we worked from initial research to final voice prototype.
GOAL
1. To research and understand user's needs for stress management VUI
2. To design an VUI that meets user's needs for daily stress management
DELIEVERY
PROJECT PROCESS
Pitch
Design
Interaction Model
Prompts List
Flow Diagram
Persona&Sample Script
Testing
Iteration
Final Prototype
My Role
Researcher & Designer
This was a 4 week project in collaboration with another two designers.
I conducted user interviews and user testings, developed bot persona, created flow diagram, interaction model, prompt list, sample script for key interactions, and voice prototype.
Demonstration Video
RESEARCH
We conducted surveys and interviews to understand people’s preferences for stress management methods and their expectations of using a stress management voice interface.
#
Most Used Stress Management Methods
117 Survey Responses
Techniques Interested in Learning More About
6 Interviews
What do user want in voice design?
Users prefer mature female human voice, displayed as an experienced and friendly adviser.
DESIGN
Meet Pacify
Pacify is an empathetic, attentive and informative VUI, that can help you manage and relieve your daily stress through stress level assessment, guided exercise and music & literature recommendation.
Pacify has a warm yet calm tone, with word choices that are conversational and jargon-free.
Development Documents:
TESTING & ITERATION
testing prototypes: guided exercise, stress assessment, entertainment recommendation.
We built three audio prototypes with Botmock,
and tested with 9 participants.
Findings from User Testing:
1. Voice & Tone
The tone of the bot makes the participants feel robotic and distant. Participants expect a human-like friendly tone.
2. Transparency about Data
For the on-boarding process, participants are confused and hesitant about providing personal information such as age and if they’re working with medical professionals
3. Transition between Bot and Human Audio
Participants feel erupted when transitioning from the bot to music or guided exercise because of different voices or background music. Participants expect a smooth hand-off in between.
4. Fall-back Options
Participants hope to be able to assess the menu at any time and be reminded of the menu options whenever they need to make a choice, such as to continue the session.
FINAL PROTOTYPE
FUTURE STEPS
- Currently this skill is available for beta-test on Alexa
- For Stress Assessment, we are using the method of questions. We hope to adopt the StressScan method in the future, so the bot will be able to detect the user's stress level by their heart rates.
- At the moment, the entertainment recommendation is sorted by popularity. We hope to personify the recommendations for users in the future.
WHAT I LEARNED
- Take advantage of Alexa built-in intents
- There are so many different testing methods!
- Set up the proper expectations with user. Let them be aware of the bot's capability.
- Sometimes we have to give up some features for the scope of the project
- Always have a fall-back pattern and a back-up solution for users (even if it means human agents). Don't frustrate your users!