Turning real conversations into structured language learning

Overview

Scripted is a mobile app designed to support intermediate language learners in developing conversational fluency through real-time conversations and personalised feedback.

My role

• Led design of the recording experience and interaction flow (solo, group and AI conversations)
• Designed key interfaces including home, recording, summary-first feedback, and AI conversation screens
• Redefined the achievement system based on usability testing (independent extension)
• Contributed to navigation structure, user flows, and iterative decision-making within a collaborative team

Product Strategy

Scripted is an app designed as an accessory to conversational practice for intermediate language learners.

The strategy focused on:

• Supporting real conversations between learners (friends, family, partners)
• Making practice more accessible through optional AI, solo modes and HelloTalk integration
• Providing real-time conversation assistance without disrupting flow
• Creating a system that encourages continued engagement over time

Concept Evolution

We explored four distinct directions to propose a solution to our problem domain, each targeting a different method of language learning.

• Recording app: Capturing real conversations and generating transcripts
• Media immersion: Learning through curated content such as films, music, games and short-form media
• Vocabulary trainer: Focusing on saving and revisiting difficult words for targeted practice
• Community platform: Enabling social interaction through messaging and daily shared content

Key Insight

Initial testing revealed that while all concepts addressed aspects of learning, none fully supported real conversational fluency.

The recording concept was the strongest, but limited by reliance on conversation partners.

Final Direction

We evolved the solution by combining:

• Real conversations recorded to produce a transcript (later extended to include AI and solo practice modes)
• HelloTalk integration rather then recreating community features
• Short form lessons based on the transcript insights

This expanded accessibility while maintaining focus on real conversational interaction.

Usability Testing & Iteration

We conducted usability testing on the high-fidelity prototype to evaluate how well the experience supported real conversational practice and post-conversation feedback.

Participants: 8 intermediate language learners

Method: Think-aloud protocol, task-based testing

Tasks: Record a conversation, review feedback, complete a lesson

Key Findings

1. Feedback felt overwhelming

   Users found transcript review dense and difficult to engage with.

   “I would just scroll down, look at a couple of the mistakes, and exit out.”

2. Progress and achievement are essential

   Users were motivated by visible progress, but the experience lacked positive reinforcement.

   “I like the weakness levels and being able to see [the bar] go from red to yellow to green as I practice.”

3. Onboarding is needed

   Some icons and features were unclear without guidance.

   “In the onboarding you could explain what the different features and icons are.”

Testing Conclusion

Key considerations

• High use of LLMs: Requires strong guardrails to ensure accuracy and trust
• Privacy & transparency: Critical due to recorded conversation data
• Engagement: The experience needs stronger reward and motivation systems

What worked well

• Strong concept: Addressing a clear gap in intermediate language learning
• Personalised value: Individual lessons based on real conversations
• Balanced, seamless user flows: Easy to navigate and supports conversation without disruption

Opportunity

The product is functional and intuitive, but lacks a sense of reward and progression.

This insight led me to explore how the product could better support motivation and long-term engagement.

Enhancing Engagement Through Achievement Design

To address the lack of motivation identified in testing, I redesigned the feedback system to focus on progress, reinforcement, and habit formation.

Solution — Achievement System

I redesigned the feedback layer to introduce a system of:

• Achievement cards
• Milestones
• Goal-based progress tracking

Key Features

Design Rationale

This approach applies principles of behavioural design:

• Positive reinforcement increases repeat behaviour
• Visible progress supports motivation
• Small wins reduce perceived effort over time

The system reframes language learning as continuous progress, not error correction.

Outcome (Conceptual)

While not implemented in the final prototype, this extension demonstrates how the product could evolve to support long-term engagement.

It directly responds to usability findings and introduces a scalable motivational layer across the product.

The achievement system was designed as a continuous loop across key moments in the product:

• After conversation → reward
• Before next session → goal setting
• During use → reinforcement
• Over time → progress tracking

Reflection

• Accessibility constraints shape product direction — Solo and AI modes emerged from the need to support users without consistent partners.
• Usability does not equal engagement — The achievement redesign addressed motivation, not just clarity.
• Summary-first feedback reduces overwhelm — Prioritizing the summary over the full transcript improved clarity and scannability.
• Low-interruption interaction design — Designing the recording flow required minimizing on-screen interaction, enabling users to stay focused on conversation without disruption.