DualDive: AI Language Learning App
End-to-end UX | AI Ed-tech | Mobile App
DualDive is an innovative language learning platform that leverages AR and AI to make the experience more engaging and realistic.
Impact: ​
Dualdive yields a Usability System Score (SUS) of 84, improving learning vocabulary by 24%, reading by 12.7%, speaking skills by 9.9% and writing by 8.5%.
Role
UX Designer
Mixed Methods Researcher
Prototyping​
High Level Goals
Empowering user with personalized learning choices
Improving learning relevancy and accuracy
Reducing drop-off rates
Enhancing language learning experience
Timeline
6 months
Tools
Figma, Adobe Suite, Miro, MiniTab
Initial Research Question
Do current language learning apps effectively help users achieve their goals?
Specific Goals
-
Identify what's missing in current apps
-
Find out how many users quit and why
-
Find ways to improve learning
Key Metrics
-
Customer NPS
-
User engagement metrics
Product Reports Findings
47 % Users Quit in initial few months
User Interviews
Let's understand the User Journey
To understand the user journey, I conducted a semi-structured user interview of people who quitted using language learning applications and platforms in the initial few months. I interviewed 15 participants for the study, between ages 18-45. After collecting data, I realized that most users were facing difficulties similar to someone most of us know, Emily from Emily in Paris!
​
Emily Cooper, a 26-year-old marketing professional from Chicago, moves to Paris for her dream job at a luxury fashion brand. As she navigates the city’s charm and chaos, she struggles with the language barrier, cultural differences, and the pressure of impressing her French clients. With a mix of ambition, humor, and adventure, she embraces Parisian life—one misstep (and croissant) at a time.​
"Ugh, French is so hard! How am I supposed to contribute to work if I don't know the product or context?"
"Did I just order a steak... raw? Again? Why does everything sound so similar? What am I even saying?"
Current user journey
Expected user journey
Key Pain Points
What are we solving for?
Lack of Flexibility
-
In choosing content
-
Setting the learning pace
Redundancy
-
Repetitive content
-
Irrelevancy
No Progress Tracking
-
No real insights into learning parameters
Process
How did I identify these issues?
-
Conducted a quick heuristic markup of the top 3 used products for language learning out there and noted areas for improvement based on established design principles​
-
Analyzed user-interviews to uncover issues and pain points based on verbal feedback, observed body language and behavior of the users.
-
Used previously done research and statistics to support the findings from my testing since it was a small-scale study.
Exploring ideas
Turning data into game-changing ideas
What if I could label the whole world around me as I see it? Won't it make my understanding more meaningful and memorable?
While I was exploring ideas to make an app to solve the discovered pain points, I saw a chance to explore AR and AI to improve the UX of the language learning platforms. I created design explorations by creating flowcharts and sketching low fidelity screens to pitch. This sparked conversations with peers about making a product that can demonstrate the current needs of language platforms and join the trending AI wave!
​
I still wanted to keep the basic learning principle same, so I explored how kids learn language in school for the first time. This gave the product's AR interface more clarity. Kids classrooms are filled with labels for each object! That's it, that was my clue!
Information Architecture
Giving structure to the ideas
Early Stage Testing
Checking for early-stage usability
Checking for usability at early stage helped me pin point learnability and adaptability of the product. 10 participants were tested on 14 tasks with each session lasting for about 60 minutes. System Usability Scale (SUS) questionnaire was used to score the product. The data was extracted through user-session replays of verbal feedback and suggestions, observed user behavior and facial expressions.
​
The complete usability report is available here.
Metrics for Testing
Task Time
Success Rate
Error Rate
Number of assists
Early stage iterations
I iterated initially the low-fidelity wireframes from the feedback received from the early-stage usability testing. These wireframes helped me solve core usability and adaptability issues. As users were already acquainted with various language learning applications out there, keeping the base structure similar was a necessity. As I dived deeper into the design stage, I decided to segregate features into different activity labels, making decision making easier.
Prioritize
User's prior experience (also known as procedural memory and semantic memory) should not be disturbed or confused with, keeping basic UI and user navigation straightforward.
Enable
New features with a proper onboarding, keeping learnability easy and intuitive. Eliminating restricted learning to more free learning, helping users to learn faster and relevant skills.
Elevate
Existing navigation and progress checking by adding cognitive and learning parameters for better user understanding and control.
Final Design
This is what the product looks like! 🎉
After getting feedback on the designs at multiple stages, I restructured the experience and gave it a few final touches to refine the look and UI of the application. Focusing on the 3 pain points discovered in the early stages, I focused on features that eliminate the frustrations of the users.
Improving Onboarding
Onboarding questionnaire help users to modify learning process according to their preference, making it easier to tailor the experience. This includes setting the learning pace, tone, frequency and theme.
Highlights
-
Improved flexibility of learning
-
Allowing users to choose thematic content
-
Setting the pace and tone of learning, depending on the situation you wish to apply and your history with the language
AR Learning Feature
-portrait.png)
This feature allows users to learn everything around them, whatever they put the camera lens on! Not just that, users can hear how to pronounce, understand use case and also record and practice pronouncing the words with real-time speech feedback on the card top-right. On top of that, they can add the object as a flashcard (add a tag to it) and revisit later to practice.
"Finally I can snap new product around me to know what they are!"
​
"I can finally tell my landlady that the plumbing is bad. I don't need Gabriel to translate every situation and complicate my life more."
Users can also start a public challenge or invite friends for private challenges to improve vocabulary. User can join a public challenge that are active and also share their progress and achievements in their feed.
Highlights
-
Improved vocabulary
-
Improved word-meaning-context association
-
Eliminating redundancy and improving relevancy
-
Improved application in daily life
AI Chat and Progress Tracking
AI Chat enables users to prepare for scenario-based conversation in real-time or in advance. You can practice and learn ahead of time. It will help you navigate your way through every situation, not knowing the language perfectly and also help you prep for some upcoming work situations.
​
A detailed progress tracking will help you keep a track of your actual learning parameters, including, reading, listening, speaking, grammar, pragmatic competence and cultural competence. It also tracks cognitive parameters such as attention, memory and speed to keep updating your learning track with the data analyzed.
Highlights
-
Enabled scenario-based conversation practice
-
Tracking specific language parameters
-
Tracking cognitive parameters
"I'm gonna be so ready to impress everyone tomorrow at the product launch party with my impressive conversational skills."
​
"Now, I can confess my feelings to Gabriel without them sounding robotic."
Impact
The knock-your-socks-off moment
The AR language learning feature designed to improve language learning improved the parameters, making learning more effective and engaging.
24 % increase
Vocabulary through AR learning
12.7 % increase
Reading through AR learning
9.9 % increase
Speaking through AR learning
8.5 % increase
Writing through AR learning
Reflection
Learnings and Takeaways
Emphasis on research
Invest time in thorough user research. Building user personas and profiles is key to empathizing with users and designing effective solutions.
Agile iterations
It is important to test more to design better solutions. It’s easy to get stuck and comfortable with your designs to identify flaws.
Show, don't tell
It is important to share ideas, good or bad. They are the basis to start something great!
Extras!
A few fun screens too good to not show
