On March 12, the Shanghai International Financial Innovation Summit, hosted by The Asian Banker, was held in Shanghai. Multiple guests discussed the application of artificial intelligence in the financial industry, with OpenClaw, a topic that has recently garnered significant attention within the sector, becoming one of the key points of discussion. Lin Yonghua, Vice President and Chief Engineer of the Beijing Academy of Artificial Intelligence, stated that the core reason lies in OpenClaw possessing three critical characteristics. First, it understands users better. This is because such systems often operate directly within a company's business systems rather than as tools separate from business operations. By functioning in a real business environment, it can better comprehend user needs and is more likely to generate tangible value. Second, it can automate task completion. An intelligent agent is not merely a conversational tool; it can automatically invoke various tools, connect to business systems, and execute complete business processes, thereby genuinely participating in enterprise operations. Third, it offers higher efficiency. By automating a large volume of high-frequency business processes, the system can operate continuously 24/7, significantly enhancing business processing efficiency. However, she also emphasized that the security risks and computational power requirements associated with intelligent agents like OpenClaw cannot be overlooked. If an intelligent agent system is to be integrated into a company's core business environment, it must operate under an enterprise-grade security architecture because it directly handles real business operations. Furthermore, such systems typically require substantial, low-cost computational power to support the automated execution of high-frequency tasks. Regarding data, since intelligent agents need direct connections to a company's actual business systems, it is essential to ensure that data sources are trustworthy, data quality is reliable, and the system can support real-time data analysis capabilities. Otherwise, if there are issues with the data itself or high data latency, these problems can be amplified within the AI system. What role can financial institutions play in the current wave of artificial intelligence? Lin Yonghua believes that the competitive landscape for foundational large language models is rapidly consolidating. Globally, there may ultimately be only around ten major companies capable of sustaining large-scale technological iterations in large models. Competition at this level is not the core domain for banks or financial institutions. The truly critical aspect is "skills" (professional capability modules). Only through these professional modules can AI genuinely understand the business logic, knowledge systems, and application scenarios of a specific industry. While hundreds of thousands of skills have been open-sourced globally, the challenge is that verified skills capable of solving professional problems and operating stably in industry scenarios remain scarce. For the financial industry, a crucial task is building a specialized knowledge base for finance. Practice has shown that relying solely on general-purpose large models is insufficient; connecting large models to specialized knowledge bases and industry data systems is necessary to deliver real value in specific business contexts. She pointed out: Today, we are entering a phase of rapid development for intelligent agents. During this phase, one of the most important capabilities is the continuous accumulation of industry-specific skills that can be invoked by these agents. Once these capability modules are standardized and toolified, they can be repeatedly called upon by AI systems, thereby genuinely driving the implementation of industry applications.