At the eighth China Fintech Forum, held on September 9 during the China International Fair for Trade in Services in Beijing, Ding Zhiyong, General Manager of the Strategic Development Department at Bank of Beijing, delivered a keynote speech. He addressed the "AI Revolution and Digital Risk Control Transformation in Commercial Banking," outlining how financial institutions can leverage emerging technologies to build smarter, more agile risk frameworks. This article presents a full recap of his remarks.

Ding began by noting the global AI landscape has undergone a dramatic evolution since the generative AI breakthrough in 2022, marking the shift from experimental technology to commercial deployment at scale. Countries worldwide are accelerating their AI strategies with dedicated policies, making intelligent transformation an inevitable path across all sectors. Banking stands out as the most promising domain for AI application, given its combination of massive compliant data reserves, standardized operational processes, and mature computing infrastructure. With rapid iteration in AI agents and multimodal models, the technology now touches every facet of business workflows and management, where human-machine collaboration and AI-assisted empowerment have emerged as the dominant approaches to financial risk control. China, in particular, has intensified its AI strategy through a steady stream of supportive policies that promote deep integration between AI and the real economy, encouraging financial institutions to adopt intelligent tools for enhanced risk management.

Beyond the macro view, the financial sector's embrace of AI has matured significantly. Early-stage implementations that merely stacked computing power and expanded model parameters, often detached from practical scenarios, are no longer the norm. Today's industry logic is anchored in solving concrete business problems, prioritizing scenario relevance, and delivering measurable value. Powered by large models' self-learning abilities, multi-dimensional data fusion, and real-time dynamic analysis, these tools tackle the inherent limitations of traditional financial risk control. Years of practice have proven that AI is fundamentally an enabling instrument, not a replacement for human judgment. The long-term trajectory for financial risk control is one of human-machine symbiosis, with AI serving as an intelligent assistant. More importantly, this technology not only optimizes individual tools and workflows but is radically restructuring the collaborative models, operational logic, and risk governance systems of banks, serving as the core engine for comprehensive digital and intelligent transformation.

Turning to the weaknesses of legacy systems, Ding explained that conventional risk-control frameworks have long depended on historical financial documents and past transaction records to assess creditworthiness. This backward-looking approach, based on static data, often fails to capture the true situation in real time. It is particularly inadequate for evaluating technology-driven or asset-light enterprises, where growth potential and R&D capability hold more predictive value than historical metrics. Additionally, data within banks is frequently siloed and fragmented, with structured and unstructured information poorly integrated. When compounded by inefficiencies, subjectivity, and delayed warning signals in manual judgment, traditional human-centric risk control falls short of the refined, high-quality standards expected in today's financial environment.

AI, however, presents a comprehensive solution to these structural challenges. By incorporating technologies such as knowledge graphs and large models, the sector can rethink credit evaluation altogether. A multidimensional framework that blends financial data, transactional credibility, and forward-looking expectations enables lenders to assess both historical fundamentals and the future growth potential and risk of a borrower. In the long run, AI is poised to shift the industry's risk paradigm from the reactive "human defense" model to a proactive "intelligent defense" system driven by data and advanced algorithms. This will lead to full-spectrum upgrades in data utilization, decision-making logic, control timeliness, risk perspective, and iterative capacity, ultimately constructing an intelligent oversight system that guards the entire business lifecycle and risk cycle. This, Ding concluded, is the future direction for refined risk management in commercial banking.

Sharing his bank's practical experience, Ding detailed how the Bank of Beijing has strategically aligned with its digital transformation goals by developing its proprietary "AIB" innovation platform. This initiative consolidates data, computational power, and technical resources to create a unified AI infrastructure. The platform breaks down departmental silos, making AI tools universally accessible, universally adaptable to all business scenarios, and universally capable of empowering processes, thereby providing a solid technical foundation for intelligent risk control and operational improvements.

The AIB platform is built on a hybrid architecture that integrates "large models, small models, and scenario-based tools." It supports a comprehensive intelligent service system covering decision support, business enablement, and client engagement. Featuring sub-platforms for foundational model applications, industry-specific solutions, knowledge management, financial scenario tools, and an office assistant, it connects to advanced models from providers and formats a functional ecosystem that combines an application matrix, efficient tool clusters, and cognitive assistant networks. This setup supports a diverse array of intelligent products operating across the head office, branches, and subsidiaries. By harnessing large models' language comprehension and generalization capabilities alongside agent technologies, the platform bridges internal business, office, data, and operating systems, breaking down barriers in data, workflows, and institutional policies. In doing so, it re-engineers process flows and user interactions. Integrating AI capabilities deeply into every business stage has created a multi-dimensional, collaboratively upgraded control system. Ding highlighted that these efforts have already earned market recognition, industry-standard awards, and 21 technical patents.

Looking ahead, the bank's strategy involves a systematic upgrade of the AIB platform to cultivate an advanced intelligent technology stack. This vision embraces agile product development, semantic knowledge mapping, natural human-computer interaction, consolidated platform operations, intelligent scenario application, and extraordinary user experiences. The roadmap includes the development of five core pillars: a multi-agent collaborative platform, a rapid delivery DevOps platform, a bank-wide knowledge management hub, a business scenario ecosystem, and an AI-finance talent development program. The objective is to elevate the platform from "single-point intelligence" to "multi-agent collective intelligence." Ding concluded that this evolution is designed to augment the capabilities of all employees, empower business lines, optimize cost structures, and drive efficiency gains, injecting sustained momentum for the Bank of Beijing's digital transformation. Appreciating the audience's time, he wrapped up his remarks with a note of thanks.