Recently, China Coal Development Co., Ltd. and iFlytek Technology Co., Ltd. have cooperated to introduce "AI evaluators" based on large model technology into the enterprise bidding workflow, significantly reducing procurement costs for enterprises. This is also the first officially appointed large-scale model "AI evaluator" in China. According to data released by the China Federation of Logistics and Purchasing, the procurement amount of state-owned enterprises reached 38 trillion yuan in 2022. At present, the traditional bidding process has problems such as low efficiency, cumbersome processes, and insufficient compliance. Traditional bidding involves multiple stages such as issuing bidding documents, evaluating qualifications, negotiating prices, and signing contracts. The complex bidding process consumes a lot of time and human and material resources, reduces bidding efficiency, and can also easily lead to non compliant behaviors such as bid rigging and collusion. Based on this industry pain point, iFLYTEK Co., Ltd. used its self-developed Spark Model as the basis to study the bidding documents and evaluation reports of China Coal Development Co., Ltd., accurately extract evaluation factors, and construct a large-scale evaluation model base for the energy industry. At the same time, through low intrusion integration, the integration with existing bidding and procurement systems can be quickly completed, providing functions such as intelligent assistance in sourcing document preparation, intelligent assistance in bidding business evaluation, intelligent assistance in non bidding business evaluation, and intelligent assistance in negotiation evaluation. This achieves the dual mode application of unmanned intelligent evaluation and human-machine assistance evaluation, promoting the intelligent upgrade of the entire process of enterprise bidding and procurement business. At present, the actual application results show that "AI evaluators" can achieve a 50% improvement in the efficiency of procurement document writing and a 70% improvement in the efficiency of bidding evaluation. (Lai Xin She)