Problem Statement

India is home to 22 official languages and over 120 regional languages, yet there is no effective, structured, and inclusive app to help people overcome language barriers for everyday communication. Migrants, professionals, students, and travelers struggle to adapt to new linguistic environments, leading to miscommunication, social isolation, and reduced opportunities.

Project Goal

To bridge linguistic gaps for students and professionals in Maharashtra by enabling Hindi and English speakers to confidently learn and use Marathi for everyday, workplace, and official interactions.

Solution

Developed Boli - a scenario-driven, culturally-aware language learning app that delivers modular lessons, real-world practice, and adaptive support.

• Personalized learning paths focused on the most relevant social and professional scenarios.

• Bilingual content, voice recognition for pronunciation, and cultural etiquette modules.

• Offline-capable, intuitive design to ensure accessibility in low-connectivity environments.

In This Section of Research &
Landscape

Secondary Research
Literature Review, Competitor Analysis
Competitor Analysis
Industry Research, Market Insights
Primary Research
Survey, Interview, Building Persona, Understanding pain points,
Understanding User
Persona, Empathy Mapping, User Journey Map, Contextual Enquiry

Secondary Research

In Brief
Studied 30 papers and articles to understand language use in India. Language and literacy play a big role in social development goals. India has a rich history of multilingualism with many people speaking two or three languages. Mother tongues are important for cultural identity and have different categories.

Government policies like the Three-Language Formula encourage multilingualism but more focus is needed on diverse local languages. The MuRIL AI model supports 17 Indian languages, helping representation for those with fewer resources.

Languages in India mainly belong to groups like Indo-Aryan (northern/central), Dravidian (south), and Sino-Tibetan/Austroasiatic (northeast and tribal areas). Multilingualism benefits business, education, media, and cognitive skills but a digital divide remains. More local content creation is necessary.

There’s a three-level language hierarchy: elite, regional, and marginalized languages. To improve technology like speech recognition, better linguistic datasets and models are needed. Key themes include language mixing, policy, education, inclusivity, culture, diversity, and equality.

Competitor Analysis

Primary Research

The Interview process started with creating a research plan outlining 3 key objectives:

1. Problem

2. Research Goals, Objectives and Questions

3. Hypothesis

Research Objectives:

• Understand real pain points and gaps in existing solutions.

• Understand how competitors navigate this gap.

• Learn about limitations and efforts put by users to deal with the problem.

Primary Research Insights

Students and Professionals in India navigate a complex multilingual environment, engaging with diverse regional languages and dialects daily. However, their language learning needs are varied and evolving, and existing solutions often fail to address the cultural and contextual nuances essential for effective mastery. 

Building Persona

My user persona funneled all the information from survey user interview responses by generating a user persona that I felt would be the best candidate to create a prototype for. The persona helped guide the project forward during the design and testing phases. Continually cross-referencing this document ensures my design meets the parameters for our outlined project brief.
Based on my user interviews I know two important findings,

1. Learners face significant challenges due to the complex linguistic landscape, including dialect variations, pronunciation nuances, and cultural contexts, which are often overlooked by existing language learning tools.

2. There is a strong demand for culturally relevant, scenario-based learning modules with bilingual support and offline accessibility.

In This Section of Defining Problem and Ideating solution

Defining Problem
Contextual Enquiry, 5 Whys, 5W1H, Fishbone diagram
Feature roadmap
Feature List, How Might We?
Task Flow
Card Sorting, Task Analysis and User Flow
Information Architecture
Organizing features.

Contextual Enquiry Model - Flow Diagram

Contextual Enquiry Sequence Model

Contextual Enquiry Cultural Model

5W 1H

Fishbone Diagram

Feature Roadmap

Feature List

How Might We?

How might we create an engaging, accessible, and practical language-learning app for Indian languages, using real-life conversations and similarity-based learning, to help people overcome language barriers?

Card Sorting

I collected feedback from 5 different users and had them sort 34 cards. 2 of them were closed, and 3 were open.

• Pronunciation

• Grammar

• Listening

• Speaking

• Reading

• Writting

• Add language

• Create a situation

• Daily life situations

• At work

• At school

• At college

• Travelling

• Ice breakers

• In case of Emergencies

• learning track (Beginner/Intermediate/Advanced)

• Translation

• Practice

• Learn

• Learning for Work

• Common Workplace Phrases

• Ai chatbot

• maintaining streaks

• user profile

• Edit profile

• Interactive lessons

• Native language

• Target language

• Rewards

• daily challenges

• AI-driven personalized lesson recommendations

• Offline

• Script and Alphabets

• Basic vocabulary

Categories:
Practice
Profile
Learn
Home
Game

Information Architecture

In This Section of Wireframing

Ideation and Wireframe
Contextual Enquiry, 5 Whys, 5W1H, Fishbone diagram
Design Guide
Typography, Colors, Grid system, Buttons.
Mid - High Prototypes
Card Sorting, Task Analysis and User Flow
Final Prototype
Organizing features.

Ideation

Moodboard

Style Guide

Mid Fidelity Wireframe

Prototype Screens

Screens

Prototype

(Click to Interact)

In This Section of Usability Testing

Defining Tasks
Scenario, Objective, Success & Failure criteria, Start and end action.
Usability Test Plan
Moderator script, Consent form, SUS, SEQ and NPS questionnaires.
Mid - High Prototypes
Card Sorting, Task Analysis and User Flow
Final Prototype
Organizing features.

Defining Tasks

Usability Test plan & Notes

Pre Test Preparation:
1.1. Moderator script
1.2. Consent form
1.3. Set up
1.4. Pilot test

During Test:
2.1. Welcome, consent, intro, think-aloud.
2.2. Pre-test questionnaire (demographics, prior experience).
2.3. Task briefing, actual task runs.
2.4. Collection of metrics (time on task, errors, success/failure) and think-aloud notes.
2.5. Administer SEQ and SUS after each task.

Post Test:
3.1. Post-task questions (debrief, open-ended feedback).
3.2. Net Promoter Score (NPS).
3.3. Acknowledgment and Regards.
5 target users volunteered to participate.
These Interviews were conducted in person, 1:1.

Single Ease Question

How difficult users found the tasks?
1= Very difficult
7= Very Easy

System Usability Scale

Net Promoter Score

Insights from the users during Usability test

“ I was curious about will this word come in this situation. I would like to see same word used in different sentences, situations. I would have liked more Activities.

“The word did not have it’s meaning, so took tike to understand.”

“Can’t I just have lifelines while playing the quiz, just like KBC format.”

“Should have included conversational, etiquettes, currency, numbers.”

“I wanted to download thew whole page, word with usage of the word.”

Opportunities

1. Expand and Contextualize Content

Issue: Users want more usage examples, sentences, and broader phrase coverage.
Improvement:

• For each word, provide at least two example sentences drawn from real life.

• Enrich scenario modules with varied, practical expressions relevant to daily life, work, and emergencies.

• Let users suggest or vote on new scenarios/phrases to be added.

2. Make Save/Download/Categorize Features Intuitive

Issue: Users like saving and downloading but want more organization.

• Improvement:

• Visual feedback after saving/downloading (“Added to Phrasebook!” with animation).

• Let users organize saved words/phrases in custom lists or by category.

• Add a “Review Later” section on the home screen for fast revision.
  
  

3. Reduce Friction and Prevent Errors

Issue: Participants experienced lack of audio feedback and navigation issues.
Improvement:

• Rethink navigation links.

• Provide an always-visible “Back” or “Help” button that lets users recover from mistakes.

• Integrate audio input and output.
  
  

4. Break Complex Tasks into Manageable Steps

Issue: High error rate and time in Task 3; users overwhelmed by multi-step scenarios.
Improvement:

• Introduce a step-by-step guided flow for scenarios (e.g., “Ready to order? First, choose your dish. Next, add a drink. Finally, review and practice the conversation.”)

  
  

5. Enhance Pronunciation and Contextual Feedback

Issue: Users wanted instant, actionable feedback after pronunciation.
Improvement:

• Provide real-time indication of correct/incorrect pronunciation, and clear tips for improvement (“Try again: stress on the first syllable”).

• Incorporate audio-based role-play to simulate real conversations (e.g., “The app acts as the waiter; you practice responses and receive feedback.”).

Learnings

#1 User-Centered Research

• Conducting thorough primary and secondary research was crucial to understand the multilingual realities and cultural nuances of the target users.

• Real-life linguistic challenges guided the creation of relevant, context-driven learning scenarios.

#2 Iterative Prototyping

• Rapid prototyping enabled quick testing of features such as audio pronunciation tools, bilingual UI, and scenario-based lessons.

• Iterations were informed by user feedback, balancing educational depth with usability.

#3 Balancing Complexity and Usability

• Designing for multilingual inclusivity meant addressing complex phonological, morphological, and syntactic differences without overwhelming users.

• Simplified, intuitive navigation and bite-sized learning modules proved essential to maintain engagement and reduce cognitive load.