AI language learning app by Tenmin

Project summary

Team:
Product team
Timeline:
April 2025 - July 2025
Project duration:
4 months
Role:
Lead designer
Team members
4 members
Product area:
Mobile app
Tools & Skills:

Introduction: What is Tenmin?

Tenmin is an innovative AI language tutor that enables real-time conversational practice. The app provides immediate feedback on pronunciation, personalized lesson content, and interactive speaking exercises designed to build practical fluency faster than traditional language learning methods

My role

As UX Lead for Tenmin, I shaped the user experience and visual direction of this AI language learning app:

  • Developed a scalable design system and partnered with the PM (Mauricio Rivero Pooley) to redefine the app's overall style
  • Partnered with the PM (Mauricio Rivero Pooley), and CEO to balance user needs with business goals
  • Crafted the home page experience, focusing on clarity and user engagement
  • Simplified complex AI interactions into approachable learning moments
  • Guided two designers working on the review page, ensuring consistency

My work helped transform advanced language technology into intuitive experiences that feel approachable and encouraging to learners

Design process

The problems

Course Selection Paralysis

New users feel overwhelmed when choosing their first course, creating friction during onboarding and risking abandonment before experiencing Tenmin's AI tutoring

Unclear Learning Method

Users don't understand Tenmin's structured approach from the current design, undermining confidence and reducing willingness to invest time in the platform

Generic Visual Identity

Tenmin's design blends in with competitors instead of standing out, failing to communicate professionalism and decreasing user trust and motivation

How might we (HMW)

Based on the three core problems identified, we synthesized our findings into two focused How Might We questions that would guide our design solution and help the team concentrate efforts on the most impactful areas

User Interview

To build Tenmin with real users in mind, the whole team conducted in-depth interviews with 10 potential language learners from students mainly.

We explored their learning backgrounds, personal goals, current pain points, and preferences for tutoring.

This research helped us validate the core problems and informed our How Might We questions, ensuring our solutions addressed genuine user needs rather than assumptions

Competitor analysis

Mauricio and I analyzed 10 language learning apps to identify market gaps

We discovered that most competitors struggled with clear course presentation and relied on stress-inducing streak systems.

However, apps with engaging AI characters built stronger user loyalty, while review methods remained generic across the market.

These insights guided us to prioritize intuitive course navigation and create a distinctive AI tutor experience

    User journey map

    Following user interviews and competitor analysis, we mapped the user journey to understand how language learning needs evolve over time and proficiency levels.

    For beginners (Phase 1-2), we discovered key motivations like personal tutor connections and immediate feedback, while frustrations centered on vocabulary overload and embarrassment.

    Intermediate learners (Phase 3-4) seek for authentic conversations and real-world applications but struggle with invisible progress and balancing grammar with practical communication

    This mapping helped us focus exclusively on beginners and intermediate learners, aligning with the app's core vision

    Information architecture

    I developed an information architecture to address the key problems uncovered in our competitor analysis.

    Our IA work prioritized logical content grouping and visual guidance, with flexibility to iterate based on user interview results

    Design iterations

    Several thoughtful iterations to find the optimal balance between information and clarity:

    The final version focuses on essential information with a cleaner presentation. Users found our final design significantly more intuitive and approachable than both our earlier versions and competitor offerings

    By moving detailed information to post-click screens, we created a more inviting entry point while still providing all necessary context when users were ready for it

    Final designs

    App structure: Home, Explore, and Review
    Home page restructure

    Before: The original home screen featured cluttered lesson cards with excessive information, creating decision paralysis and cognitive overload for users. Its design lacked the youthful energy central to our brand identity

    After: After three thoughtful iterations of the lesson card component with the Lead PM, Mauricio Rivero Pooley , our redesigned home page now presents a clean, intuitive interface that preserves minimalism while significantly reducing cognitive load. The streamlined design better aligns with our core brand concept: youthful and professional, while making course selection more approachable.


    Setting up

    The streamlined setup now avoids duplicate questions and reduces decision fatigue.

    For example, instead of asking users to self-identify their language level and then separately choose courses, we now collect level information once and automatically display only relevant course options that match their proficiency, allowing users to start their language learning journey with minimal friction

    Note: The header and the description below the header are written by another designer


    Introducing a personalized AI tutor: Ten

    Ten is our charming 3D AI tutor that brings learning to life with cute, colorful visuals inspired by Microsoft Fluent Emoji.

    Through playful animations, color-changing emotions, and bouncy interactions, Ten creates a fun, engaging educational experience that makes users feel like they're learning from a lively, enthusiastic companion rather than a traditional AI assistant. Working collaboratively with Mauricio Rivero Pooley across several iterations, we developed a 3D style consistent with the redesigned app's overall design language

    Ten introduces itself during setup by asking users to teach it their name pronunciation, a thoughtful solution for Korean, Japanese, Vietnamese, and other names often mispronounced in English

    Flashcard and casual chat with AI

    We have introduced intuitive swiping functionality with clear initial instructions, transforming traditional quizzes into engaging flashcard experiences that reduce learning stress while adding playful interaction.

    Most importantly, Ten spontaneously sends surprise messages to users, creating delightful unexpected moments that spark curiosity about what he might say next. These random conversations mirror real-life friendships where friends chat about anything, providing natural daily conversation practice while keeping users motivated and eagerly returning to interact with their AI tutor.

    Note: The flashcard concept was designed by another designer (Juna Han) and the PM (Mauricio Rivero Pooley)

    A/B Testing results

    We tested two home screen designs with 10 participants:

    40%

    Increase in perceived professionalism by users

    Professionalism:

    • Original design: 3/5 average
    • Redesigned version: 4.2/5 average

    80%

    Increase in perceived youthful energy by users

    Youthful Appeal:

    • Original design: 2.5/5 average
    • Redesigned version: 4.5/5 averag


    Our redesign successfully balanced professionalism with a significantly more youthful, approachable feel which creates a better first impression for users!

    Takeaways

    01 Leading the team

    As the lead designer, I directed the overall strategy while collaborating effectively with PM, designers, and CEO

    02 New AI Tutor

    Designed a new AI tutor experience with enhanced interactions that create deeper user connections and make learning more enjoyable

    03 Simplicity

    Redesigned the homepage for better clarity and intuition, helping users quickly understand

    04 Measurable wins

    80% increase in perceived youthfulness and 40% boost in professionalism through A/B testing