On March 6th, the China Academy of Information and Communications Technology held a sharing meeting on the "Model Room Project" for the implementation of large-scale model applications. Using the model room project as a starting point, we gained in-depth insights into the technological innovation progress and industry implementation practices of large-scale models. As one of the first outstanding case representatives of the "Model Room Project", TRS was invited to share its technical support for the application of Ping An Bank's "Huixiaomiao" Audit Large Model.

The "Huixiaomiao" Audit Large Model is an AI audit system developed by Ping An Bank's Audit and Supervision Department and Financial Technology Department in collaboration with TRS, targeting internal audit vertical scenarios. It is based on an audit think tank to build a high-performance vector database and relies on the bank's self-developed BankGPT platform and TRS's advanced big model scenario implementation capabilities.

The Audit Large Model has developed intelligent Q&A, intelligent audit reports, intelligent text analysis and other functions, focusing on improving the quality and efficiency of the key links of audit "check, write, analyze", realizing the full process intelligence of internal audit application, solving the pain points of low audit analysis efficiency under massive data, inconsistent qualitative standards of problems in complex violation scenarios, and low efficiency of report writing. The big model application is deeply embedded in the daily inspection of supervisory personnel, promoting the upgrade of the "AI+audit" mode, and comprehensively improving the efficiency and quality of audit work.