" 你会关心一个电饭锅能做多少种不同的饭菜，而不是单纯关注炉 子的好坏。 " 作者丨周蕾 编辑丨包永刚 阻碍金融机构把Agent从演示PPT推向核心业务场景的，究竟是什么？是算力成本，是模型能力，抑或是 一个万能的开发平台？ 在与恒生电子首席科学家白硕的深度对话中，我们得到了一个不太常见的答案：以上都不是最要紧的。 白硕早年间在中科院计算所从事前沿研究，后长期担任上海证券交易所总工程师，主导核心交易系统升 级，如今作为恒生电子首席科学家，推动AI技术落地。在经过学术前沿、行业监管核心与产业实践这一完 整路径之后，他对当下最热门的Agent话题，给出了具有历史纵深感的、颇具穿透力的洞察。 他指出，缺乏足够"厚度"的业务接口——这里并非指底层技术的API，而是指封装了业务逻辑、能"听 懂"业务人员自然语言指令的能力单元——直接导致现在许多Agent项目陷入"读不懂"真实业务需求当中的 复杂意图，无法解读有业务语义的自然语言的指令，或者只能对原有系统做简单粗暴的封装。他风趣地提 到：你会关心一个电饭锅能支持多少种花式菜谱，至于底下加热组件好不好用，会是你关注的重点吗？ 而目前通用型Agent平台的价值，其在整体解决 ...

Core Insights - The Chinese banking industry is at a turning point with the emergence of AI technology, particularly AI Agents, which are set to revolutionize core banking functions such as credit and risk management by significantly enhancing productivity and efficiency [1][3][21] - AI Agents, built on large AI models, can autonomously perform tasks, assist in decision-making, and provide personalized financial services, thereby reducing manual intervention and operational costs [1][3][4] Group 1: AI Agent Implementation and Value - AI Agents are becoming a core focus for banks, with significant investments being made to develop and implement these technologies [4][6] - The core values of AI Agents include improving efficiency through end-to-end automation, enhancing decision-making capabilities, and providing personalized customer experiences [3][21] - Major banks like ICBC and Agricultural Bank of China are leading in financial technology investments, with ICBC planning to spend approximately 28.518 billion yuan in 2024 [6][8] Group 2: Bank-Specific Developments - Agricultural Bank of China has introduced the "Mosu Loan Scoring Card" AI Agent, which can generate credit reports in 30 seconds, significantly speeding up the due diligence process [8] - Postal Savings Bank is rapidly advancing its AI capabilities, achieving over 87.5% automation in alarm troubleshooting through its AI Agents [9] - Other banks, including China Merchants Bank and Ping An Bank, are also developing their own AI Agents to enhance data analysis and customer service [11][12] Group 3: Technology Partnerships - Banks are increasingly collaborating with technology companies to bridge the technological gap and enhance their AI capabilities [13][20] - Major tech players like Baidu, Alibaba, and Tencent are providing comprehensive AI solutions and infrastructure, which are crucial for the successful implementation of AI Agents in banking [14][15] - The partnership between banks and tech companies is essential for unlocking the potential of AI in the financial sector, especially for smaller banks [13][20] Group 4: Challenges and Future Outlook - Despite the rapid development of AI Agents, many banks are still focused on non-core applications, indicating a gap between potential and actual implementation [21][22] - The banking sector requires high accuracy and reliability from AI systems, which currently face challenges such as a 95% accuracy rate in leading financial models [23][24] - The transition to AI-driven banking is a long-term process that necessitates a solid AI strategy and collaboration between banks and technology providers to achieve significant ROI [30][31]