Google Meet AI Voice Translation
A concept prototype for Google Meet that enables live speech translation, designed to let people communicate naturally across different languages.
Google Meet AI Voice Translation
Google Meet AI Voice Translation
A concept prototype for Google Meet that enables live speech translation, designed to let people communicate naturally across different languages.
A concept prototype for Google Meet that enables live speech translation, designed to let people communicate naturally across different languages.
Role: UX & Interaction Designer
Role: UX & Interaction Designer
Team Size: 2
Team Size: 2
Timeline: 3 months
Timeline: 3 months
Tools: Figma · Veo3 · ElevenLabs · FigJam
Tools: Figma · Veo3 · ElevenLabs · FigJam
FOCUS
FOCUS
Enabled seamless real-time voice translation in Google Meet prototype using Veo3 and ElevenLabs.
Validated project direction with Google I/O 2025 Speech Translation launch.
Enabled seamless real-time voice translation in Google Meet prototype using Veo3 and ElevenLabs.
Validated project direction with Google I/O 2025 Speech Translation launch.
CONTEXT
Google Meet is Google’s widely used video conferencing platform, known for features like screen sharing, live captions, breakout rooms, and deep integration with Google Workspace. It enables teams, classrooms, and communities to connect seamlessly across devices.
PROBLEM & OPPORTUNITY
Google Meet makes virtual collaboration easy with features like captions, breakout rooms, and screen sharing. But when conversations cross languages, users struggle, misunderstandings pile up, focus is lost, and external tools like Google Translate break the flow.
Through user research, we saw a clear need: people want real-time translation that’s accurate, fast, and effortless to turn on, without disrupting the meeting.
How might we help people who speak different languages communicate naturally in Google Meet by translating speech in real time?
FINAL DESIGN SOLUTION
Google Meet features that let users enable translation with consent, follow conversations in real time, and review transcripts after the call.
01
PRIVACY CONSENT
Shows a clear consent screen before enabling translation.
Built user trust by making privacy transparent and giving people control before turning the feature on.
02
LIVE TRANSCRIPTION & TRANSLATION
Translates speech in near real-time with live transcripts.
Reduced cognitive load and kept conversations seamless by giving users the speed, accuracy, and visibility they expected.
03
POST-CALL TRANSCRIPT
Provides a downloadable transcript after the meeting.
Ensured clarity and follow-up, helping users avoid misunderstandings and easily review important points after multilingual conversations.
MY REFLECTION & LEARNINGS
01 | Designing for Cognitive Load
I learned how even small delays or unclear UI elements can overwhelm users during live calls. Designing for low cognitive load became a key priority, ensuring features supported the conversation rather than distracted from it.
02 | Building Trust Through Transparency
I discovered the importance of privacy and consent when introducing AI. By making translation opt-in and upfront, I saw how design choices could directly influence user trust and comfort.
03 | Turning Research Into Design Decisions
I practiced translating user frustrations (searching for transcripts, juggling Google Translate, missing information) into clear features like live translation, visible toggles, and post-call transcripts. This reinforced how research directly drives product design.
04 | Giving Users Control
I saw how much users wanted autonomy, from choosing when translation starts to customizing transcripts after the call. Designing with flexibility and control in mind helped improve engagement and satisfaction.
05 | Communicating New Concepts Simply
I practiced framing a new, AI-driven capability in a way that felt intuitive and familiar, so it fit naturally into Meet’s existing ecosystem instead of feeling like an add-on.
FINAL SCREENS
CONTEXT
Google Meet is Google’s widely used video conferencing platform, known for features like screen sharing, live captions, breakout rooms, and deep integration with Google Workspace. It enables teams, classrooms, and communities to connect seamlessly across devices.
PROBLEM & OPPORTUNITY
Google Meet makes virtual collaboration easy with features like captions, breakout rooms, and screen sharing. But when conversations cross languages, users struggle, misunderstandings pile up, focus is lost, and external tools like Google Translate break the flow.
Through user research, we saw a clear need: people want real-time translation that’s accurate, fast, and effortless to turn on, without disrupting the meeting.
How might we help people who speak different languages communicate naturally in Google Meet by translating speech in real time?
THE DESIGN SOLUTION
Google Meet features that let users enable translation with consent, follow conversations in real time, and review transcripts after the call.
01
PRIVACY CONSENT
Shows a clear consent screen before enabling translation.
Built user trust by making privacy transparent and giving people control before turning the feature on.
01
PRIVACY CONSENT
Shows a clear consent screen before enabling translation.
Built user trust by making privacy transparent and giving people control before turning the feature on.
03
POST-CALL TRANSCRIPT
Provides a downloadable transcript after the meeting.
Ensured clarity and follow-up, helping users avoid misunderstandings and easily review important points after multilingual conversations.
03
POST-CALL TRANSCRIPT
Provides a downloadable transcript after the meeting.
Ensured clarity and follow-up, helping users avoid misunderstandings and easily review important points after multilingual conversations.
02
LIVE TRANSCRIPTION & TRANSLATION
Translates speech in near real-time with live transcripts.
Reduced cognitive load and kept conversations seamless by giving users the speed, accuracy, and visibility they expected.
02
LIVE TRANSCRIPTION & TRANSLATION
Translates speech in near real-time with live transcripts.
Reduced cognitive load and kept conversations seamless by giving users the speed, accuracy, and visibility they expected.
MY REFLECTION & LEARNINGS
01 | Designing for Cognitive Load
I learned how even small delays or unclear UI elements can overwhelm users during live calls. Designing for low cognitive load became a key priority, ensuring features supported the conversation rather than distracted from it.
02 | Building Trust Through Transparency
I discovered the importance of privacy and consent when introducing AI. By making translation opt-in and upfront, I saw how design choices could directly influence user trust and comfort.
03 | Turning Research Into Design Decisions
I practiced translating user frustrations (searching for transcripts, juggling Google Translate, missing information) into clear features like live translation, visible toggles, and post-call transcripts. This reinforced how research directly drives product design.
04 | Giving Users Control
I saw how much users wanted autonomy, from choosing when translation starts to customizing transcripts after the call. Designing with flexibility and control in mind helped improve engagement and satisfaction.
05 | Communicating New Concepts Simply
I practiced framing a new, AI-driven capability in a way that felt intuitive and familiar, so it fit naturally into Meet’s existing ecosystem instead of feeling like an add-on.
FINAL SCREENS
Role: UX & Interaction Designer
Team Size: 2
Timeline: 3 months
Tools: Figma · Veo3 · ElevenLabs · FigJam
FOCUS
Enabled seamless real-time voice translation in Google Meet prototype using Veo3 and ElevenLabs.
Validated project direction with Google I/O 2025 Speech Translation launch.
CONTEXT
Google Meet is Google’s widely used video conferencing platform, known for features like screen sharing, live captions, breakout rooms, and deep integration with Google Workspace. It enables teams, classrooms, and communities to connect seamlessly across devices.
PROBLEM & OPPORTUNITY
Google Meet makes virtual collaboration easy with features like captions, breakout rooms, and screen sharing. But when conversations cross languages, users struggle, misunderstandings pile up, focus is lost, and external tools like Google Translate break the flow.
Through user research, we saw a clear need: people want real-time translation that’s accurate, fast, and effortless to turn on, without disrupting the meeting.
How might we help people who speak different languages communicate naturally in Google Meet by translating speech in real time?
FINAL DESIGN SOLUTION
Google Meet features that let users enable translation with consent, follow conversations in real time, and review transcripts after the call.
02
LIVE TRANSCRIPTION & TRANSLATION
02
LIVE TRANSCRIPTION & TRANSLATION
Translates speech in near real-time with live transcripts.
Reduced cognitive load and kept conversations seamless by giving users the speed, accuracy, and visibility they expected.
Provides a downloadable transcript after the meeting.
Ensured clarity and follow-up, helping users avoid misunderstandings and easily review important points after multilingual conversations.
03
POST-CALL TRANSCRIPT
03
POST-CALL TRANSCRIPT
Shows a clear consent screen before enabling translation.
Built user trust by making privacy transparent and giving people control before turning the feature on.
01
PRIVACY CONSENT