Pipeline 3: RAG generative AI

In this section, we will use RAG generative AI to automate a customized and engaging marketing message to the customers of the bank to encourage them to remain loyal. We will be building on our programs on data preparation and Pinecone indexing; we will leverage the Pinecone vector database for advanced search functionalities. We will choose a target vector that represents a market segment to query the Pinecone index. The response will be processed to extract the top k similar vectors. We will then augment the user input with this target market to ask OpenAI to make recommendations to the market segment targeted with customized messages.

You may open Pipeline-3_RAG_Generative AI.ipynb on GitHub. The first code section in this notebook, Installing the environment, is the same as in 2-Pincone_vector_store-1M.ipynb, built in the Pipeline 2: Scaling a Pinecone index (vector store) section earlier in this chapter. The Pinecone index in the second code...