Re-Ranking is a second-stage retrieval process that reorders initially retrieved results using more advanced AI models to improve relevance before presenting information or generating responses.
Voice AI platforms use Re-Ranking to improve Retrieval-Augmented Generation by selecting the most relevant documents from an initial candidate set, resulting in more accurate and trustworthy answers.