Using an example of watsonx Assistant calling a customizable Q&A with RAG Python function on watsonx.ai, we explain the architecture for how these products can be used together. We explain how to build an assistant with call outs to invoke the watsonx.ai REST API to pass a question and get an answer, and to subsequently pass end user feedback on the answer to be recorded and analyzed in watsonx.ai. We show how this scenario works end to end in the user interface from the point of view of a person using the assistant that is powered by the LLM on watsonx.ai. Finally, we explain the setup and deployment options for using watsonx Assistant + watsonx.ai as software or as a service on the IBM Cloud.