Chi app is a voice assistant app created for the diverse immigrant groups in the United States who struggle with accessibility to voice assistant app due to accent and language barrier.
Timeline
June 2022 - October 2022
Problem
My role
UX Researcher,UX Strategist, UI/UX Designer
Communication Barrier Among Immigrants.
Immigrants and non‑English speakers often struggle with mainstream voice assistants because of accent bias and limited language support. Most systems misinterpret regional accents and native languages, turning simple tasks into frustrating experiences. This lack of inclusive language coverage makes everyday interactions feel inaccessible for many users.
The Solution
App for accessing languages and making requests in your dialect.
Supports multilingual interactions and understands regional accents to reduce errors.
Delivers responses in users’ preferred languages for clearer, more confident communication.
Builds trust by honoring natural speech patterns rather than forcing users to adjust.
Multilingual & Accent-Inclusive Voice Recognition
Accessible Voice-First Interaction & Controls
Offers multiple ways to activate voice input for flexible, natural use.
Enables hands-free interaction, reducing the need to read or type.
Improves accessibility for users who struggle with written English or complex interfaces.
WHITE PAPER RESEARCH
A way to achieve goal with 95% success…
Starting with the white paper research, I drew directly from a key peer‑reviewed article to ground my understanding of language accessibility and dialect‑aware design. Its evidence guided me toward clear gaps in current voice technologies and shaped the direction of my recommendations.
Research Focus
The effect of language on user’s ability to communicate with voice assistant
The research explored how language, accent, and cultural context affect users’ ability to successfully interact with voice assistant applications.
“Current automatic speech recognition (ASR) systems achieve over 90–95% accuracy when used with native speakers of the language, but the level of accuracy decreases significantly when the same ASR system is used by a non-native speaker.”
TESTING + IMPROVEMENTS
Competitive Analysis + The Gap
Competitors provide poor support for diverse accents and non‑native speakers
To identify accessibility limits in voice assistants, I reviewed popular platforms—Siri, Google Assistant, and Alexa—and looked for common usability gaps. I found none supported diverse accents or non-native speakers.
Google Assistant
Siri
User interview
When voice assistants recognized accents and tones, interviewees succeeded three times more—a problem for both immigrant young adults and native-born Americans.
I interviewed immigrant teens and native-born Americans to study how accent bias and language limits impact voice assistant use. Affinity mapping revealed recurring pain points and accessibility gaps.
RESEARCH QUESTIONS
Can you describe the last time you used a voice assistant?
What was the most frustrating part of that experience? Why?
What made you want to try using a voice assistant in the first place?
How do you usually try to get voice assistants to understand you?
Tell me about a time when a voice assistant actually worked well for you—what made it successful?
Based on the trend of my affinity map, I’ve noticed if there is no accurate accent or intonation recognition, users won’t have a successful or reliable experience with voice assistants, regardless of their background.
More Accessible Voice Input Feedback
I made the interface show what the assistant heard so users get clearer, immediate feedback.
I cut confusion by offering simpler correction choices.
THE FINAL SCREENS
The final product
THE MAIN INSIGHT
Interviewees often reported voice assistants misinterpreting their accents and intonations.
Alexa
Toba, the Junior Nurse
DESIGN
22 year old | Junior Nurse
User Story
“I spend a huge amount of time code-switching during day-to-day activities. Now I have to do the same with my voice assistant”
Goals
Ensure voice requests are understood without altering her accent.
Use voice assistants efficiently during busy shifts. Speak to tech naturally, reflecting her cultural identity.
Lower cognitive load by relying on voice instead of typing or repeating.
Motivation
Wants technology to recognize and respect her accent and culture
Prefers time-saving, stress-reducing tools for fast-paced settings (e.g., healthcareSeeks efficient, reliable voice tools for daily tasks
Needs adaptable tech that fits her, not the reverse)
Pain Point
Voice assistants often misinterpret her speech.
Feels forced to fake an accent or over-enunciate to be understood
Repeats simple commands, wasting time and causing frustration and embarrassment when voice tools fail publicly or professionally.
Set back + A new direction for accessibility
I started the project assuming users had varied accents, but research proved my assumptions wrong. I learned to do thorough user research and pay attention to overlooked real-world experiences. I also found voice UI needs different skills than regular UX, so I adopted voice interactions designed for accessibility.
2 major Improvements in my design
Based on various feedback from mentors, I continually iterated my design over the span of 6 weeks with three major improvements:
Improved the interaction flow to better support users with diverse accents
I corrected the assistant’s misunderstanding so users know what it thought.
I reduced interruptions by adding simple options like “Try again” or “Tap to correct.”
Conclusion + Lesson Learned
What I learned
ChiApp was my first end-to-end UX project, and I approached it with strong curiosity and excitement around solving a real-world accessibility problem. The project challenged me to think critically about inclusive design, especially for users whose needs are often overlooked by mainstream voice technologies.
I discovered how different VUI and VUX are from traditional UX, especially in how they rely on conversational flows, error handling, and non-visual feedback.
I learned that designing for voice-first interactions requires understanding speech recognition limitations and accessibility needs that aren’t visible in standard UI design.
I realized how essential thoughtful user research is, especially when designing for users with diverse accents whose experiences are often misrepresented or overlooked.
I gained a deeper appreciation for accessibility-driven design and how even small interaction details can significantly impact user confidence and trust.
Thank you for reading!
I hope you enjoyed exploring this project! I’m always open to feedback, collaboration, or just a good conversation about UX. You can reach me at eskaychinaza8@gmail.com