Advancing the modernization of the Chinese-style grain industry requires addressing the high grain losses across the entire industry chain, a core task that lays the foundation for forming a grain-saving and loss-reduction mechanism with Chinese characteristics. Taking COFCO Group as a case study, this paper deeply analyzes its practical work in the field of grain saving and loss reduction across the entire grain industry chain. Firstly, it proposes a new process of "Human-AI Collaboration," combining large language models with grounded theory methods in coding. Secondly, through "Human-AI Collaboration," the case materials are coded, and the results are analyzed. Based on this, the advantages and disadvantages of three mainstream large models as "AI" in open coding, axial coding, and selective coding are evaluated based on their inductive and deductive abilities. Finally, the abstract thinking, logical reasoning, and systemic view abilities of humans are leveraged to refine the final results of open coding, axial coding, and selective coding for the entire grain industry chain. This establishes a theoretical framework for enhancing grain-saving and loss-reduction capabilities at each stage, thereby forming a new mechanism with Chinese characteristics for the entire grain industry chain.