Policy Pilot

Inspiration

Policies and regulations are often lengthy, complex, and difficult to understand. We wanted to build a solution that allows users to ask questions in simple language and quickly find reliable information without going through large policy documents manually. This inspired us to create PolicyPilot, an AI-powered policy assistant.

What We Built

PolicyPilot uses Retrieval-Augmented Generation (RAG) to retrieve relevant information from trusted policy documents and generate context-based answers with source citations. The system is designed to reduce hallucinations by grounding responses in the available policy content and providing a fallback when relevant information cannot be found.

The platform includes a conversational chatbot, OTP and Google authentication, user and admin roles, and an admin interface for managing policy information.

How We Built It

The frontend was developed using React.js, while the backend was built with Java, Spring Boot, and Spring Cloud Microservices. We used MongoDB for data storage and implemented Spring Security, JWT, and OTP authentication for security. The AI layer combines RAG with an LLM to retrieve and generate policy-based responses.

Challenges & Learning

One of the main challenges was ensuring that the AI did not generate unsupported information. Designing the retrieval and fallback mechanism helped us understand the importance of grounding AI responses in reliable sources.

We also learned how to design microservices, integrate AI with backend APIs, manage authentication securely, and build a complete end-to-end AI-powered application.