1. Introduction
2. Features
3. Usage
4. Technology & Dependencies
5. Development and Contribution
6. Reflection & Growth

LegalFlow, an AI-powered virtual assistant designed to streamline document classification in the legal domain. This project was developed for the Knight Hacks 2023 Hackathon, taking on Morgan & Morgan's LegalTech Innovation Challenge.

LegalFlow utilizes Azure technology to conduct document analysis to classify documents into 6 different categories. By doing so, it simplifies the process of organizing and managing legal documents, saving valuable time and resources.

• Document Classification
  • Into 6 Categories:
    1. Medical Records
    2. Medical Bills
    3. Police Reports
    4. Correspondence
    5. Court Orders
    6. Other
• Chatbot
  • Legal conversations for clients
  • Inquisitive in nature
• Web Interface
  • User-friendly design powered by Streamlit, ensuring accessibility and smooth navigation
• Accessibility Options
  • Text to Speech (TTS) service for ChatBot (LegalFlowAI) responses

• Law firm clients can use to identify and classify documents
• Law firms can utilize this tool internally to sort and filter through legal documents
• Clients may receive generalized guidance & assistance through conversation

• Azure AI Document Intelligence
• LangChain
• Streamlit
• OpenAI
• gTTS

• Learned:
  • Semantic Kernel (First approach)
  • How to deploy a model using streamlit
  • Azure integration
• Room to expand
  • Automation of tasks through chatbot query (such as correspondence & drafting of legal docuements)
  • More accessibility options such as
    • Multilinguality
  • Improve model by enhancing OCR (incorporate formatting, spacing, tables...)

• Richard Brito, Lead Web developer & Designer
• Ian Ordonez, Project Manager and Legal Researcher
• Maverick Reynolds, Lead Langchain & Streamlit developer
• James Trent, Azure and Deployment Specialist