Shanghai Ganglian Zongshi Large Model and Xiaogang AI Assistant

25,000,000

• Co-developed the Zongshi Commodity Large Language Model in collaboration with the Shanghai AI Laboratory’s “Shusheng Puyu” large model.
• Released Xiaogang AI Assistant 2.0, supporting intelligent content creation, customer service automation, and smart operations.
• Developed a commodity price prediction model using comprehensive commodity datasets and satellite remote sensing as alternative data sources, achieving a prediction accuracy rate of over 80%.

• In internal applications, the Zongshi model and Xiaogang Assistant reduced costs and improved efficiency by 30%, covering content creation, automated customer service, and operational management.
• In external applications, the model provided intelligent services such as price data queries, market opinion aggregation, industry encyclopedias, and trade opportunity identification, helping enterprise users save 30% of their time.
• By December 2024, the project had served 848 enterprise clients and generated an additional RMB 50 million in annual receivables.
• The business has since been spun off as Shanghai Ganglian Technology Co., operating independently, with plans to seek financing and pursue a listing on the STAR Market (Sci-Tech Innovation Board).

This project advanced the digital transformation of the commodity sector, becoming the industry’s first large-scale application of a domain-specific large model. By building both a specialized large model and an AI assistant, it enabled Shanghai Ganglian to upgrade its business model and established a benchmark case for AI empowerment in traditional industries. With the spin-off subsidiary operating independently and preparing for a public listing, the project demonstrates strong sustainability and growth potential in capital markets.