The typical structure of aRAG pipeline involves several stages which collaborate in order to retrieve information and produce a response. First, the documents or business data are gathered and arranged in an indexable repository, usually a vector database. On entering a query, the retrieval stage will search through the available information and will find those pieces of information that are more relevant to the input query. These pieces of information are then passed to the generative AI as additional context. Lastly, the model generates a response based on this context.