Designed a transcript feature that allows users to capture, store, and revisit key conversation segments
Findings
Recruiting various participants who is representing lie target audience. Developing a research plan with clear instructions and tasks. Moderated the research sessions effectively.Developing a research plan with clear instructions and tasks.
Problem Statement
We identified that users struggled to revisit important parts of conversations, leading to low re-engagement and shallow session depth.
Hypothesis
If we allow users to extract and revisit key parts of conversations, we can improve retention and session depth.
Target Audience
| Segment | Who | Pain Point |
|---|---|---|
| Primary | 25-45 working adults, migrants | Family abroad, mixed-lang office chats |
| Secondary | 15-25 youth/students | Friends different states, socialise |
Declaring assumptions
Business Assumptions
Customers have a need to chat freely with family/friends in their own language without copy-paste frustration
These needs can be solved with real-time audio-to-transcript translation in native script
Customers will pay through ads after experiencing translation value
product will beat competitors because seamless audio→chat flow vs manual Google Translate
Free users will share app virally after first successful cross-language chat
Marketing will work because Tamil pride messaging resonates with target audience
Customers are willing to keep parallel WhatsApp + translation app usage
Customers will pay through ads after experiencing translation value
User Assumptions
Where does product fit in workflow? (WhatsApp→Translate gap)
Between WhatsApp voice messages and Google Translate copy-paste—faster audio→translated chat
What problems solved? (Lang barrier, slang fail, copy-paste)
Will users trust local-only voice processing?
Will users share translated transcripts?
Will users tolerate 2-3s translation delay?
Budget Android phones smooth experience?
Will users keep both WhatsApp + our app?
Family groups > individual chats usage?
Users understand "optional translate" button?
Youth discover via friends vs ads?
Prioritized assumptions
Market Validated - 60%+pain = build
Privacy green light - NPS above 8 = Safe
Tech validated - 82%+accuracy = go
Growth engine - K-factor>0.3 = organic
UX confirmed - less than 3s equals mass market
UI locked - 70% native script
Onboarding ready - 95% completion
Launch ready - No tutorials needed
Model confirmed - No WhatsApp switch needed
Segment found - Groups first features
Quantitative Research
"I used a combination of user interviews and surveys to validate assumptions. Interviews helped me understand user behavior and pain points qualitatively, while surveys allowed me to quantify those insights at scale."
User Persona's
Transcript Feature
UX Research
Transcript
Need for accessibility | Improved searchability | Increased efficiency
Brand Awareness
Support
Help
Regular exposure | Interactive communication | Organic reach
Design Thinking
Accessibility
Problem Statement
Use clear and concise labels in the mic settings menu to identify the language options. Include native language names for broader understandability.
Voice in Feeds and Conversations
Settings
Visiblities
Speech Recognition Accuracy | Privacy Concerns | More Natural Interaction
Integrations
Tech stack and DB
Chart Overview
User Journey
Visual Design & Interaction
Typography
Colors
Prototyping
Skeleton and Mockup
A/B Tests
Where can I reach you?
What are you looking for?
Give me a quick idea of what you need
You’ll hear from me soon.
All works © Surya Narayanan 2026
