Smarter Bixby: AI for Seamless Bilingual Experience

Bilingual users often face friction with Bixby because it only supports one active language at a time. This creates difficulties during everyday tasks such as turning off alarms, setting timers, or sending quick messages.

My Role
UI/UX Designer

Project
Personal Project

Tool Used
Figma
Adobe XD
After Effect

Timeline
14 Days (2025)

Overview

Problem Statement

Bilingual users often face friction with Bixby because it only supports one active language at a time. This creates difficulties during everyday tasks such as turning off alarms, setting timers, or sending quick messages.

Goal

To design a bilingual-friendly Bixby experience that allows seamless language recognition and task execution without manual switching.

Design Process

Understand - Research - Define - Ideate - Prototype - User Testing - Presentation

Target Users

  • Bilingual young professionals who frequently switch between two or more languages (e.g., English & Korean, English & Mandarin, German & English, Spanish & English) for work and daily tasks.

  • International students who rely on Bixby for managing alarms, timers, messaging, and other everyday activities in a multilingual context.

Research Methods

  • Online research: Review existing voice assistants and multilingual support features.

  • Online survey: Collect quantitative data on bilingual users’ experiences and pain points.

  • In-person interviews: Gather qualitative insights, observe real usage, and understand user frustrations in context.

How Might We Questions

  1. How might we enable Bixby to automatically detect and switch between languages without user intervention?

  2. How might we design contextual experiences where Bixby understands mixed-language commands in natural conversation?

  3. How might we reduce bilingual input errors so users can dictate or message fluidly across multiple languages?

User Research

Background Research

Research goal is to understand the challenges bilingual users face when using voice-controlled devices in daily activities, including alarms, timers, workouts and messaging, in order to identify pain points and opportunities to improve multilingual voice recognition reliability and user experience.

Research 1 - Current System (Insights / Problems)

  1. Single-language limitation → Bilingual users struggle.

  2. Low recognition accuracy → Critical tasks fail.

  3. Poor contextual understanding → Misaligned user intent.

  4. Weak multilingual input → Texting and messaging issues.

  5. Outdated compared to competitors → Reduced trust and usage.

User Interview & Survey

Research 2- Survey & Interview Summary

How strong do you feel about your satisfaction with Bixby’s speech recognition?

1.5

Out of 5 (low)

Do you feel you can use Bixby smoothly with voice commands?

14.5 %

can use bixby smoothly with voice

Bixby failed to understand in

  1. Setting timers

  2. Turning off alarms

  3. Scheduling calendar events

  4. Sending messages

How many times did you have to repeat the same command to Bixby?

3 times

on average

Did you have to do tasks manually when Bixby failed?

70 %

complete the tasks manually

How satisfied are you with Bixby’s speech recognition accuracy?

1.3

Out of 5 (low)

Research Findings & Pain Points

  1. Alarm & Timer Failures → Bixby often doesn’t respond to bilingual commands.

  2. Messaging Errors → Mixed-language dictation leads to frequent mistakes and manual corrections.

  3. Context Failures → Commands like pausing a workout fail, forcing manual interaction.

Prototype

Wireframes