一句话：提出用带RAG的会话式LLM，为缺乏产品管理经验者生成并验证完整可测试的产品需求文档。
这篇立场论文指出，临床医生、研究者等创新者常缺乏产品管理经验，导致交给工程或制造合作方的需求说明不充分。作者主张构建基于大语言模型（LLM）的系统，生成并验证全面、可测试的产品需求文档（PRD）。系统结合检索增强生成（RAG）提供领域知识，并以对话方式引导用户细化需求，目标是与工程或制造伙伴签约所需的详细规格。作为立场论文，它主要提出系统愿景与动机，未报告实现或实证评估。
This position paper motivates the development of a large language model (LLM) -based system for creating comprehensive, testable product requirements documents (PRDs). The intention is to guide innovators who lack formal product management experience— clinicians, researchers, academics, and other specialists—in producing specifications detailed enough for successful contracting with engineering or manufacturing partners. The LLM would be augmented with a retrieval-augmented generation (RAG) component and set up in a conversational manner, mitigating issues that arise when under-specified requirements are handed to contract firms.