Finalist
October 2023
8 min read

PropBot Revolution

We're excited to announce that our team has won the Cyber Cypher 2.0 hackathon! Our project, PropBot, is a multilingual GenAI assistant for real estate that helps users find their dream homes in their preferred language.

We're proud to have developed a solution that addresses the critical need for accessible real estate information in multiple languages. Our AI-powered platform makes it easier for users to search for properties, understand contracts, and get answers to their questions in their native language.

"Our solution addresses the critical need for accessible real estate information in multiple languages," said one of our team members.

We're committed to making real estate more accessible to everyone, regardless of their language or background.

Team PropBot
NMIMS Mumbai
The Big Question

What if finding your dream home was as easy as having a conversation?

This question sparked a 48-hour journey that would challenge everything we knew about real estate technology.

The Challenge That Started It All

Picture this: You're searching for a new home, but you're drowning in listings that don't match what you actually want. Legal documents are written in a language that seems designed to confuse you. And if English isn't your first language? Good luck.

The real estate market wasn't just complex—it was actively hostile to regular people trying to make one of the biggest decisions of their lives.

Pain Points We Identified:

  • Language barriers in property search
  • Incomprehensible legal documents
  • Generic, irrelevant recommendations
  • No way to ask natural questions

Our Vision: PropBot

We envisioned an AI assistant that could understand you—not just your words, but your intent. Something that could read through dense legal documents and explain them in plain English (or Hindi, or Gujarati, or any of 12+ languages).

PropBot wasn't just another search engine. It was going to be your intelligent real estate companion.

The 48-Hour Sprint

Hour 0-12: Research & Foundation

We started by talking to real estate agents and potential buyers. The insights were eye-opening—people wanted to ask questions like "Find me a 3BHK under 50 lakhs near good schools" but had to navigate through endless filters instead.

Hour 12-36: The Breakthrough

The magic happened when we successfully fine-tuned a T5 model for legal document summarization. Suddenly, 50-page contracts became 3-paragraph summaries that anyone could understand.

Hour 36-48: Polish & Pitch

The final sprint involved integrating everything into a seamless experience. We were running on pure adrenaline, but seeing PropBot respond to natural language queries in multiple languages felt like witnessing the future.

Features That Changed Everything

Each feature was designed to solve a real problem we identified during our research.

Multilingual Search

Breaking language barriers in property discovery

12+ language support
Natural language queries
Context-aware responses

Impact: Made property search accessible to non-English speakers

Legal Document AI

Making complex contracts understandable

Contract summarization
Risk assessment
Plain language explanations

Impact: Reduced document analysis time by 75%

Smart Recommendations

Personalized property matching using ML

Behavioral analysis
Preference learning
Dynamic filtering

Impact: Improved match accuracy by 60%

Technologies That Powered PropBot

Our carefully chosen tech stack that made the magic possible.

Python
Transformers
Google Translate API
React
MongoDB

The Impact We Created

75%
Faster document analysis
🌍
12+
Languages supported
🎯
60%
Better match accuracy

Our Journey at Cyber Cypher 2.0

PropBot proved that AI could make complex systems accessible to everyone. The judges were impressed by our technical execution, but what really stood out was how we focused on solving real human problems. While we didn't take home the top prize, the experience and feedback we received were invaluable.

What We Learned

Fine-tuning models for domain-specific tasks requires quality data

We spent significant time curating and cleaning our legal document dataset. The quality of our training data directly impacted the accuracy of our contract summarization feature.

User experience is crucial when dealing with complex information

Even the most sophisticated AI is useless if users can't understand or interact with it effectively. We iterated on our interface design multiple times based on user feedback.

Multilingual support needs more than just translation

Cultural context, legal terminology variations, and regional preferences all matter when building truly multilingual applications.

The Future of Real Estate Tech

PropBot represents our vision for a more accessible and user-friendly real estate industry, where technology serves to bridge gaps rather than create barriers.