100xElevate — AI Customer Support Chatbot

Developed an AI-powered customer support assistant for the 100xElevate digital agency website to provide instant, accurate, and context-aware responses to customer inquiries. Instead of relying on generic LLM-generated answers, the assistant uses a Retrieval-Augmented Generation (RAG) architecture to retrieve relevant information from the agency's internal knowledge base, including service documentation, FAQs, pricing details, and company policies, ensuring every response is grounded in verified business content.

The backend is built with FastAPI and implements a complete RAG pipeline using LangChain. Documents are processed through recursive text chunking, transformed into vector embeddings using Google Generative AI Embeddings, and indexed for semantic similarity search with PostgreSQL (pgvector) and FAISS. When a user submits a query, the system retrieves the most relevant context before generating a response with Google Gemini, resulting in highly accurate, source-aware, and hallucination-resistant answers.

The frontend is developed with Next.js, TypeScript, and Tailwind CSS, providing a modern and responsive chat interface that integrates seamlessly into the agency's website. This solution improves customer engagement, reduces repetitive support requests, and enables visitors to receive reliable information 24/7 while maintaining consistency with the company's official documentation.