62% NPS
Writing feature
Validated active beginner practice in testing.
01 Case Study
Kaizen Languages.
Designing an AI-powered language learning experience around confidence, motivation and retention.
- Role
- Founding Product Designer
- Timeline
- Approximately three years
- Platform
- Native iOS and Android
- Team
- Founders, developers, language specialists, AI stakeholders
- Focus
- AI conversation, writing, retention, gamification and subscription conversion
My role
As Founding Product Designer, I owned the product experience end to end—from customer research and product strategy through UX, interaction design, visual design, testing, analytics and delivery.
- Product strategy and roadmap influence
- User research and testing
- UX and interaction design
- Visual design and illustration
- Prototyping and developer documentation
- Product analytics and experimentation
- Retention and conversion optimisation
02 Product challenge
Making AI language learning useful for complete beginners.
Kaizen used conversational AI to help people practise speaking with digital tutors. The challenge was not simply to create a chat interface. We needed to make unfamiliar technology feel human, provide enough structure for complete beginners and build habits strong enough to support both learning outcomes and a sustainable subscription product.
03 Product-level strategy
Connected decisions across three years of product work.
Prioritise novice learners
Rather than concentrating only on advanced content, the product focused on beginners who needed confidence and structure first.
Define a human visual direction
I tested photography, illustration and hybrid routes, then adapted the strongest evidence to an early-stage budget.
Turn AI messaging into a lesson system
Conversation became a structured learning environment with progressive support, recovery and measurable iteration.
Introduce Japanese writing support
Hiragana and katakana practice gave novice learners a stronger foundation before harder lesson progression.
Build practice and retention mechanics
Practice, shipped streak mechanics and progress feedback helped learners return. Recovery concepts such as freezes were explored for a later phase.
Experiment commercially
Free-access limits, paywall hierarchy and pricing emphasis were tested against customer value and business sustainability.
04 Combined impact
Strongest defensible outcomes.
Across three years, the strongest gains came from connecting learning behaviour, product confidence and commercial experimentation rather than treating them as separate activities.
05 Explore work
Explore three years of product work.
Six feature stories cover the major product areas I shaped as founding designer. Each page goes deeper into the problem, evidence, design decisions and outcomes.
01Defining how an AI tutor should feel
I explored four contrasting art directions and tested how photography, illustration and hybrid approaches affected trust in an unfamiliar AI product.
An evidence-led recommendation adapted to early-stage budget constraints.
Read feature story
02Designing conversation as a language lesson
I developed the core AI conversation model, then iterated the lesson experience using observed frustration, funnel behaviour and repeated product versions.
Lesson completions increased by 25% after introducing Suggestions.
Read feature story
03Helping complete beginners learn Japanese writing
I challenged the proposed advanced-content roadmap and designed a progressive system for learning hiragana and katakana.
Created a clearer progression route for the product’s novice audience.
Read feature story
04Building a personalised practice system
I designed practice tools that helped learners revisit vocabulary, recognise weak areas and continue learning outside structured lessons.
Connected lessons, revision and confidence-building into one learning system.
Read feature story
05Designing habits without punishing learners
I introduced shipped streak and reward mechanics to encourage regular learning, while exploring recovery concepts for a future phase.
Retention increased by 35%, with 40% day-one and 22% week-one retention.
Read feature story
06Improving premium conversion through experimentation
I tested free-access limits, subscription hierarchy and pricing emphasis to balance customer value with business sustainability.
Approximately 13% relative conversion uplift and more than 20% growth in six-month-plan selection.
Read feature story
06 Reflection
What three years as a founding designer taught me.
Kaizen required me to move continuously between product strategy, detailed interaction design, learning behaviour and commercial decisions. The strongest outcomes came from treating those areas as one connected system rather than separate design activities.