RAG-Based Insurance Underwriting Assistant powered by LLMs, LangChain & FAISS
UnderwriteAI automates insurance policy underwriting by extracting risk insights from PDF documents using a Retrieval-Augmented Generation (RAG) pipeline — similar to enterprise underwriting platforms used in InsurTech.
| App Interface | Risk Assessment |
|---|---|
 |
 |
| Underwriting Report | RAG Q&A |
|---|---|
 |
 |
| Feature | Description |
|---|---|
| 📄 PDF Processing | Upload & parse insurance policy documents |
| 📊 Risk Assessment | Automated risk scoring (0-10) with identified risks |
| ✅ Approval Recommendation | Decision + Confidence Score + Reasoning |
| 🔍 Missing Information Detection | Flags incomplete policy data |
| 📝 Underwriting Report | Full structured report generation |
| 💬 RAG Q&A | Ask natural language questions about the policy |
- Frontend: Streamlit
- LLM Framework: LangChain
- Vector Store: FAISS
- LLM: Gemini / OpenAI
- PDF Parsing: PyPDF2
- Language: Python 3.10+
# Clone the repo
git clone https://github.com/Sneha8271/UnderwriteAI.git
cd UnderwriteAI
# Install dependencies
pip install -r requirements.txt
# Add your API key in .env
GOOGLE_API_KEY=your_key_here
# Run the app
streamlit run app.py