Just in Case
Designed and Developed an AI Legal App to Help 1M Sri Lankans
Find Justice Without Leaving Home.
My Roles
Reseacher
UI/UX Designer
Developer
Duration
3 Months (Jul - Oct 2024)
Tools/Technologies
Figma
React Native
Node.js
PostgreSQL
Hugging Face
LangChain
FastAPI
Clerk
Gmail OAuth
Why This Mattered
With the insights gathered, I reframed the problem: the issue was not just access, but also a lack of clarity, trust, and control.
Users did not want to simply book a lawyer, they wanted to know if they even needed one. They needed language support and someone who could explain things simply, privately, and affordably.
This led to the central goal of the project:
How might we simplify access to legal support so that Sri Lankans, regardless of background, can act on their rights confidently?
Overview
The legal system in Sri Lanka is complex, expensive, and intimidating, especially for rural and low-income communities. Legal literacy is low, consultation costs are high, and tech-based solutions are rare. To address this challenge, I designed a mobile-first platform that brings AI-driven legal support directly into people’s hands. JustinCaseApp is an AI-powered mobile platform created to deliver instant legal guidance, lawyer booking, and legal education, all tailored to the Sri Lankan context.
The Design Framework
I lead with intention, not by script.
This isn’t a checklist, I tweak the process to fit the mission. I don’t follow a rigid formula. Like a master chef adapting recipes to elevate the dish, I used this process as a flexible foundation and adapt it as needed to deliver impact.
Research and Findings
To ground the design in real needs, I began with user research that combined surveys and interviews with people from both urban and rural regions of Sri Lanka.
I gathered over 50 survey responses and conducted 5 in-depth interviews with a mix of legal professionals and everyday citizens. The goal was to uncover what stops people from seeking legal help and what support they need.
Empathy Map
To further understand user motivations and emotions, I created empathy maps around three key user types based on the research. Here's a overview of one:
User Type: Friend wrongly accused of a traffic violation
Say & Do
See
Pain Points
Gain Points
Proto Personas
Based on patterns in the data, I developed key personas to guide design decisions. I used Proto Personas instead of Qualitative or Statistical Personas to quickly capture assumptions and early insights about users. Below are 3 primary personas.
Competitor Analysis Overview
To position JustinCaseApp meaningfully in the legal tech landscape, I conducted a review of global platforms and tools. While many existing solutions serve professionals or offer narrow services, JustinCase fills a unique gap: AI-powered, mobile-first legal support tailored for everyday people in Sri Lanka.
Research-Informed Design Decisions
Katz et al. (2016)
→ Proved AI can predict case outcomes with high accuracy. JustinCase adopts this to guide users in making confident decisions early.
Schweighofer & Liebwald (2021)
→ Showed how NLP can simplify complex legal text. JustinCase applies this to explain user rights and laws clearly.
Free et al. (2013)
→ Demonstrated how mobile tech boosts access in underserved sectors. JustinCase applies this in the legal domain across Sri Lanka.
Findings
I identified five core barriers that shaped the final design direction:
Lack of legal literacy
People often don’t know what their rights are or when to act.Fear of cost
Legal help is perceived as expensive and reserved for the elite.Language and tech barriers
Most platforms are not built with Sinhala/Tamil-first interfaces or mobile-first users.Access to trustworthy lawyers
Users don’t know who to trust or how to evaluate lawyers.Overwhelming complexity
The process feels bureaucratic, filled with legal jargon and emotionally taxing.
These insights helped frame a product that isn't just a booking tool but a guide, translator and bridge to justice.
Building the Solution
With user insights in hand, I focused on building an MVP that simplified access to legal help, built trust, and kept the experience mobile-first. Every feature was designed with clarity, cultural context, and usability in mind.
Core Features (Working in MVP)
AI chatbot gives legal advice via NLP
➝ Delivers instant answers so users avoid delays and skip long legal explanations.
Lawyer-matching & appointment booking
➝ Connects users with the right lawyer in minutes so no one wastes time scrolling directories.
Feedback and rating system
➝ Helps users choose trusted lawyers based on real experiences so they feel secure booking.
Legal glossary and article hub
➝ Boosts legal literacy so users know their rights and ask smarter questions.
Minimal, mobile-first UI
➝ Reduces overwhelm so even first-time users navigate easily and feel in control.
Google-auth login
➝ Quick onboarding via familiar tools so users jump in without extra steps.
What I Did?
Simply Everything :)
Behind the Build
React Native Front-End
➝ Built a cross-platform mobile app that runs smoothly on both iOS and Android, ensuring wide accessibility. The modern, minimal UI makes the app intuitive even for first-time users.
Node.js Back-End
➝ Powers all backend logic with scalable performance, handling multiple user sessions efficiently so users get a fast and reliable experience even during peak traffic.
AI Chatbot with Hugging Face + LangChain
➝ Delivers personalized legal answers in real-time using natural language processing (NLP). Built with a curated document base and fallback responses to ensure clarity and trust.
FastAPI for API Layer
➝ Acts as the communication bridge between the AI, database, and front end, enabling seamless data flow and low-latency response times for chatbot and booking actions.
PostgreSQL Database (Hosted on Render)
➝ Stores user, lawyer, and consultation data securely with high performance. Hosted on Render to allow cloud-based scaling as the user base grows.
Strapi CMS as Admin Center
➝ Linked directly to the PostgreSQL database, giving admins a clean interface to manage lawyer profiles, legal content, glossary terms, and user activity — all without touching code.
Clerk + Google Auth for Authentication
➝ Provides secure, modern login via Gmail, allowing users and lawyers to authenticate easily through trusted third-party services.
Reflection
It was intense, messy, and sometimes a overwhelming project. But more than anything, it gave me clarity.
It taught me how to:
Think like a systems designer, considering equity, ethics, and access
Build for edge cases and low-resource users without sacrificing functionality
Communicate across tech and non-tech domains
Prioritize value over features and build trust over flash
If I had more time, I would have partnered with a real legal aid organization and expanded the NLP training into Sinhala and Tamil, which remains a goal for future.
This project reaffirmed my belief that design isn't just about interfaces, it's about building bridges where none exist.
#Nevertheless, this can be monetized and scaled globally with regional customization to become a truly worldwide solution.
