This project aims to create a chatbot with RAG (Retrieval-Augmented Generation) technology so that the responses generated will be more specific and can adjust to the context we want. In making this project I used LLM in the form of text-bison@002 from Vertex AI developed by Google. While for the vectore store in this project I used Pinecone. Then for the dataset that I use as an experiment is a dataset that I scraped from the Prudtial Indonesia website.