REGEN introduces a new benchmark dataset aimed at enhancing the capabilities of large language models (LLMs) in generating personalized recommendations through natural language interactions. By augmenting the Amazon Product Reviews dataset with user critiques and contextual narratives, REGEN allows for more nuanced conversational recommendations that adapt to user feedback. The study demonstrates how models like LUMEN can effectively integrate recommendation and narrative generation, paving the way for more intuitive user experiences.