Speaking Buddies is an innovative proof of concept prototype for an interactive conversational agent, designed to enhance foreign language speaking practice for adult learners.

It addresses the need for more engaging and effective online speaking practice tools, a demand that has grown significantly in the context of the recent pandemic.

Overview

Role

UX Researcher UX/UI Designer

Duration

9 weeks

Result

I developed a user-centred language learning tool that:

  • offers direct conversational practice with AI avatars,

  • increases user engagement and confidence in speaking a foreign language,

  • addresses the real needs of language learners by focusing on coherent speaking practice rather than phrase memorisation,

  • set a strong foundation for future development and testing,

  • identified potential for expansion into web platforms, standalone products, and immersive technologies such as AR/VR.

The process

  • Existing language learning apps rely heavily on gamification and pre-recorded interactions that are not always representative of real-world speaking practices.

  • Users often learn set phrases without developing coherent speaking skills.

  • The demand for more engaging and effective online speaking practice tools has grown significantly due to the recent pandemic.

  • Users need more than gamification features and pre-recorded interactions to learn how to speak a foreign language eloquently and confidently.

Challenge

Problem Statement

Current language learning tools often fail to provide realistic and engaging speaking practice, leaving adult learners unable to build confidence and fluency. Learners need interactive, personalised, and accessible solutions that prioritise natural conversation over memorisation, addressing the growing demand for effective online language speaking practice.

I conducted mixed methods research to supplement academic literature, including:

  • An online survey with 18 adult Duolingo users from different backgrounds, focusing on their preferences and pain points in current language learning apps.

  • Two sets of semi-structured interviews with:

  • 3x Duolingo learners and

  • 3x English Foreign Language professionals

The purpose of these interviews was to gain insights into both user experience and industry perspectives.

Design and Development

  • I used speech recognition technology to create a natural user interface.

  • I implemented lip-syncing AI avatars to mimic real conversational scenarios.

  • I designed subtle facial expressions and optional subtitles for enhanced understanding and comfort.

User Personalisation

  • I collected user information such as the learner's first language and learning goals to customise the experience.

  • Learners could choose between multiple languages and avatars.

Accessibility and Inclusivity

  • I ensured the design adhered to WCAG regulations with contrasting colours, large fonts, and legible buttons.

  • I included features like session recording and optional subtitles.

Prototyping and Testing

  • I created high-fidelity wireframes and prototypes using tools like Figma.

  • I conducted user testing to refine the prototype based on feedback.

I developed additional features based on research findings, such as mini-group sessions for enhanced interaction and learning diversity.

Action

Key Features

Direct Conversations with AI Avatar: Users engage in natural dialogues with a synthetic AI avatar, choosing sessions to practice the language of their choice. The integration of lip-syncing and subtle facial expressions aims to improve comprehension and user comfort.

Enhanced Interaction: The app offers optional subtitles for better understanding, along with the capability to record sessions for pronunciation and interaction review.

Flexible Learning Options: Includes innovative features like mini-group sessions, catering to diverse learning preferences and enhancing the overall experience.

Personalisation: Collects user data to tailor the learning experience, including language preferences and learning goals, and offers customisable avatars.

Accessibility: Ensures compliance with WCAG standards through features like contrasting colours, large fonts, legible buttons, and options for session recording and subtitles.