The technique of adding relevant external information directly into a prompt before sending it to an AI model.
Context injection is the mechanical step at the heart of RAG — taking retrieved content and inserting it into the prompt so the model can use it to answer.
