VERA
The scientific approach to AI Language Apps
PROBLEM STATEMENT
Globalization saturates our world at the speed of light, lubricated by transportation and technological advances. Youtube, Netflix, Instagram along with other platforms quickly update their audience with top entertainments around the world, instigating interests to explore and learn about new countries, people, and cultures. As a result, the demand to pick up new languages soars, followed by an influx of language apps in the marketplace.
However, without any guiding standards, these apps often vary greatly in quality and contents. Some are beautifully designed but poorly structured, some focus on vocabulary but forfeit grammar, some fail to address speaking with importance, and the others... let’s just say there’s currently no perfect app.
The K-Pop fandom’s exponential growth over the years in social media
As humans though, we strive for perfection. Which is why this study exists - to attempt at giving a name and a face to what could most likely be the perfect language app.
And to make things easier, we named it Vera. That’s half of the problem solved.
A language app that offers complete training in all language skillsets with high level of customization and real-time assistance of AI.
HYPOTHESIS
Language learning apps are all over the place, with each specializing in some disciplines but neglecting the others. Most users keep an average of 2-4 different apps on their phone just to master the full scope of a language (typical example: one app for vocabulary, one app for grammar, another app for connecting with native speakers). This leads to distraction, inconsistency, and inevitably confusion.
A one-size-fits-all app eliminates this problem. Users stay focused throughout their progress. Lessons follow leveling order. Performance recording and feedbacks become more consistent and effective. Everything gets pushed and elevated with the assistance of AI.
Why
How
Anything that sells is a well-molded combination of science (structure) and art (aesthetics). Science lends us insights into building something that is logically sound and sustainable while art gives the structure an appealing appearance.
For the first part, Structure, let’s tap into the human’ general learning process, the language learning process, and AI’s possible roles in this grand scheme.
Basic Learning Styles
According to teacher Neil Fleming, there are 3 types of learners: auditory, visual, and kinaesthetic. If we cater our teaching methods to suit with the users’ preferred learning style, they might benefit tremendously.
Opportunities
Offer lessons in both text and video format
Offer practice sessions with each new lesson
Offer one-on-one chat function or chat/call ability with friends
The 4 Skillsets of Languages
There are 4 skills that one must master when studying a language: listening, speaking, reading, and writing. Speaking and Listening have more to do with conversations while Reading and Writing determines literacy. Altogether, they are measurements and tools for one to accurately and fluently express himself and understand others.
Opportunities
Cover all 4 skillsets in depth
Assessment and feedbacks to users’ performance within each skillset
AI in Learning
AI’s ability to learn from inputs and build upon them makes it a great virtual teaching assistant.
Opportunities
Provide catered lessons and feedbacks that match users’ learning styles, level of proficiency, areas of interests, and availability
Provide live translation
Conversational Bot
Conversational AI mimics human conversations and interactions, capable of communicating naturally with users through texts or voice. This allow users to practice speaking & listening at any time and at their own pace.
Amazon’s SocialBot project
Conversational AI is an interesting and tricky subject, for sure. There is so much more to learn about and build upon, which is exciting and promising for those who know how to leverage this resource. Amazon’s SocialBot project is a great inspiration for many more breakthroughs to follow.
Explore the amazing Amazon’s SocialBot
COMPETITION ANALYSIS
DUOLINGO
Duolingo reigns as the most popular language app. It’s gained favor among users for its gamified learning, friendly and attractive design.
Strengths
Gamification, beautiful interface, vocabulary building through repetition
Weaknesses
Limited grammar teaching, over repetitiveness, impractical sentences, limited speaking
AI Features
Personalized learning path, simulated real-life conversations
DROPS
Drops engages its users with vivid images and gamified learning
Powerful visuals, gamification
Lack of auditory and kinesthetic learning factors, limited grammar teaching, limited speaking
Vocabulary learning through visuals, spaced repetition system, personalized learning path
TANDEM
Tandem is favored when it comes to chatting with people who are interested in learning your language and teaching you theirs.
Language immersion
Inconsistent and unsorted user-generated content, reliance on other users’ engagement
In-app translation, matchmaking algorithm connecting users with language exchange partners
Now the fun part.
I conducted some one-on-one interviews to learn about users’ opinions about the language learning process, then dive deeper into their experience with some of their go-to language apps. The goals are to:
Record general learning styles
Explore motivations and pain points
Capture people’s experience with using AI in learning
Using these data, we can:
Create a learning app that technically works
Create a learning schema that appeals to different types of audience
Incorporate AI as a study assistant in an elegant, non-intrusive manner
USER INTERVIEWS
THE INTERVIEWS
INTERVIEWEES
14
AGED
17 - 54
QUESTIONS
6
What languages are you learning, and why?
When/how often do you learn them?
What are the best ways for you to learn new languages?
Which language learning platforms are you using? What do you like and dislike about each?
What have you found to be the trickiest parts when learning a new language?
What is your experience with learning using AI?
THE QUESTIONS
EMPATHY MAP
To make better sense of the interviews’ findings, I inserted the data into this empathy map.
KEY INSIGHTS
Results of the interviews revolve around a few common themes.
They reflect users’ needs, motivations, and pain points.
Understanding “why” gives people insights into the subject at hand, facilitating information retention and further discovery.
The interviewees value apps that are thorough with their explanations of new concepts, grammar points and vocabulary usages.
Understanding “Why”s
9/14 users mentioned that some apps lack structural knowledge and therefore, hard for them to follow and retain information. Traditional resources like textbooks are sometimes the preferred channel thanks to their well planned-out lessons.
Structured Lessons
Repetition was considered a great way to learn vocabulary. Most interviewees enjoy the gamified system in some apps; however, 1 of them said at the end she wasn’t able to retain the vocabulary in any meaningful way, and another expressed his concern for gamification’s being nothing more than muscle reflex and blind memorization.
Vocabulary Learning
Almost all subjects believed that immersion - the quality of being surrounded by the people or channels that use, speak, and broadcast the language on a daily basis - is crucial to understanding the language better, staying motivated, and getting improved.
10/14 users are using 1 or more separate platforms on top of others that enables them to communicate with native speakers. They believed that this is the closest way they could get a hold of accents and colloquialisms.
5 users brought up how awkward and impractical some sentences are in the apps. This reinforces the importance of learning from real life.
Immersion
5 users who got positive experience with AI said they actively used it on a daily basis to answer questions, summarize information, and generate ideas. 4 others mentioned having some experience, though so far left unimpressed.
When asked about implications of AI in language apps, 11 showed excitement. 2 felt skeptical, voicing their doubt about AI’s ability to replicate real life conversations and materials.
AI Learning
USER PERSONAS
ERIN
the experienced traveler
ALEX
the cultured youngster
SITE MAP
With the requirements in mind, I created this Site Map that arrange data in a way that makes sense with users’ mental model.
5 Main Tabs
Permanent Link
| Dashboard
| Account
| Study
| Practice
| Connect
| Vera
USER FLOWS
Learn new Vocabulary
2. Learn New Grammar
3. Translate
SKETCHES
PROTOTYPE
VERA, in flesh
Home screen
Onboarding is designed to learn about individual users’ goals, motivations, and hobbies, all of which will be utilized in the future to create customized learning experiemce.
ONBOARDING
“How am I performing?”, “How can I do better"?” are the common concerns among language learners.
To answer these questions, Vera provides users with a performance tracker of their 4 skills as demonstrated from their exercises, tests, and interactions with Vera chatbot. This tool is a meticulous analysis of users’ learning progress, reinforced with personalized feedbacks and tips for improvement. In this manner, learners can see the “why” behind every moving parts, thus understand better their strengths & weaknesses and know how to move forward with specific guidance.
TRACK YOUR PERFORMANCE
INDIVIDUALIZED LESSON PLANS
Users have the freedom to follow scheduled lesson plans or explore other options.
Users’ inputs during onboarding including study goals, availability, personal interests are later combined with users’ demonstrated level of competency to create customized lesson plans.
Vocabulary and Grammar, the 2 building blocks of languages, make up the main study contents which can be accessed through Study tab or Today’s Lesson Plan.
In the video format, an instructor would explain new words and grammar points in depth, build the lessons through images, verbal repetition, and conversations. This is great for auditory and visual learners.
Additionally, the text format is available to serve traditional learners who prefer read/write learning style. It also consolidates the knowledge from the video in a more easily absorbed, note-friendly output.
Every lesson comes with its own set of practices so all users can brush up on what they just learned.
STUDY
(VOCABULARY & GRAMMAR)
Learning New Grammar
Learning New Vocabulary
Practice comes hand in hand with Study. It reinforces the knowledge one just accumulated through repetition and application. Through practice, new words and concepts register meaningful relationships with the human brain, therefore boosting information retention.
Practice has its own tab and is also accessible through Today’s Lesson Plan. It is made up of Skills - lesson-based exercises and My Library - users’ saved contents.
PRACTICE (LISTENING, READING, WRITING)
SKILLS
With the Grammar and Vocabulary garnered from previous lessons, users now apply them for the mastery of 4 skills, Listening, Speaking, Reading and Writing. The former 3 skills are available under Practice tab, while Speaking can be more efficiently performed using Vera conversational chatbot.
Mandatory practice are tied with lessons
Remember giving out your interests during onboarding? Vera took your answers into account and fill up “Based on your Interests” section with contents that you like, for example, Music, Culture, and Sports. This view will be updated all the time to keep things fresh and engaging!
”Based on your Interests” are Optional practice exercises that users can choose to do in their free time
Here users can revisit and practice their saved vocabulary and grammar points.
2. MY LIBRARY
Practice Vocabulary in the form of text or flashcards.
At the end of the day, language is but a means to connect people. Therefore, Chats and Forum become the most important features of this application, in hope that here is the place where users find compassion, joy, motivations, and endless inspirations.
CONNECT
Getting out there. Partnering with people who share the same interests. Connecting with native speakers. Asking and receiving help. Express yourself, learn new things, spread love.
Immerse.
An interactive dictionary at the reach of your hand.
Translation through voice and images makes Vera a great on-the-go tool as one adopts brand new adventures, makes new friends and immerses himself in rich, beautiful cultures that the world has to offer.
VERA TRANSLATION
Translation through Voice
Translation through Image
VERA CONVERSATIONAL BOT
Available 24/7, Vera conversational bot acts as a teaching assistant and a study partner. The bot can be accessed using voice or text-based chat.
As an AI conversational chatbot, Vera possesses extreme flexibility and resourcefulness. Its ability to converse in a variety of topics, answer to learner’s problems, and assess users’ Speaking and Writing skills make it a powerful representation and epitome of Vera the app.
While it is our top priority for now to develop a Conversational Bot that talks and feels as naturally and humanly as possible to encourage engagement, it is our future endeavor to encapsulate in it human traits and quirks, like a sense of humor, local accent, and language pattern.
DESIGN NOTES & FUTURE IMPLICATIONS
Vera is a powerful app. It got so much potentials that during brainstorming I had to keep asking myself, “Is this going to be too much?” It was tempting to include all the best things in your product but doing so risks ruining the app’s structure and usability.
Throughout the design process, I made conscious effort to make sure that all features were absolutely necessary, each had a place and meaning in the information architecture, and the app as a whole was navigable.
Functionalities
It was learned through the user interviews that although most users explicitly enjoyed learning through gamification, they would reach for more conventional platforms for in-depth knowledge like grammar and writing, and other communication-based apps for listening and speaking.
Certainly, it’s worth exploring gamification alongside other formats like texts, videos, virtual interactions,… that help deliver and retain knowledge more effectively. This way, we know users are coming back not just for the fun but the real sustainable values gained from app contents.
With Vera going the non-gamification direction, I feel like it’s a more realistic journey of language learning. That being said, it does not lack the fun or reward element. Vera appeals to its users for its engaging and relevant topics of interests, coming-of-age contents, and especially, direct involvement in the learning community alongside AI interactions.
It’s a very much lively system.
Gamification
In the future, it might be interesting to see Vera work alongside with textbooks as an auditory supplement and teaching assistant.
While not included in the prototype, in-app games between users and friends (accessible through Chat) is a desired feature. Simple mechanic games like Hangman or Crosswords are fun and competitive ones that let users learn while enjoying themselves.