Core Insights - The article discusses the launch of "AI Qingfu," an intelligent knowledge engine developed by Ping An Fund, aimed at transforming the investment research ecosystem through AI technology [1][2]. Group 1: AI Technology and Investment Research - The "AI Qingfu" platform significantly reduces the time required for report analysis from hours to minutes, enhancing efficiency in investment research processes [1][3]. - The platform utilizes RAG technology to create a closed loop of "data-knowledge-application," employing natural language processing (NLP) to analyze unstructured data from various sources [2][3]. - The system allows for rapid retrieval of relevant knowledge segments by generating multi-dimensional search keywords and optimizing result relevance through a ranking model [3]. Group 2: Broader Impact on Company Operations - Beyond investment research, "AI Qingfu" extends its capabilities to human resources, financial compliance, and administrative management, improving cross-departmental collaboration [4]. - The platform has led to the development of intelligent assistants like "HR Xiaoling," which provide instant responses for standard services, thereby reducing management costs [4]. Group 3: Safety and Reliability Measures - To address potential issues with large models, Ping An Fund has implemented a dual control mechanism combining technical constraints and human verification to ensure the reliability of generated content [5][6]. - A three-tiered protection system has been established to safeguard data security and privacy, including input filtering and output purification [6]. Group 4: Future Development and Goals - The "AI Qingfu" platform aims to expand its application across the company and engage in industry exchanges to share digital transformation experiences [6]. - The technology team plans to continuously upgrade the platform by incorporating more unstructured data sources and exploring advanced technologies for automated analysis [6].
平安基金发布“AI青蚨”:破解数据“沉睡”困局 构建智能投研新范式
Quan Jing Wang·2026-02-06 02:23