At the 8th China Fintech Forum, held in Beijing on September 9th as part of the China International Fair for Trade in Services, industry leaders convened for a critical discussion. The session, themed "Quality Enhancement, Transformation, and Ecosystem Reconstruction in the Financial Sector Amid the Wave of Digital Intelligence," brought together key figures from banking, insurance, and securities. Tian Hanyong, Chief Information Officer of Weihai Blue Ocean Bank; Sun Chenyu, Head of AI Applications for the Audit Department at China Export & Credit Insurance Corporation; and Zhu Zhaochen, General Manager of the Digital Transformation Management Office at East Asia Qianhai Securities, engaged in a deep-dive conversation moderated by Ma Jie, Deputy General Manager of the Intelligent Platform Department at Bank of Jilin.
In his opening remarks, Ma Jie noted that new-generation information technologies like AI, big data, and cloud computing are increasingly integrated into financial operations, driving the sector's digital and intelligent transformation. He emphasized that digital intelligence has become a key driver for financial institutions to enhance development quality and service capabilities. The dialogue aimed to exchange practical experiences, key challenges, and future directions in digital-intelligent transformation across banking, insurance, and securities sectors.
Kicking off the introductions, Tian Hanyong explained that as a private bank with essentially no physical branches, his institution operates entirely online, particularly for C-end business. He shared that while they have performed well over the past eight years in consumer credit and big data risk control, they face significant pressure from new policies and market changes, prompting a strategic shift towards B-end and corporate business transformation. He expressed eagerness to learn from peers in this transition.
Sun Chenyu then introduced himself as the AI Applications lead for the audit department, framing his perspective on digital risk control. He emphasized that their focus lies in innovating AI-augmented audit practices, compliance governance, and maintaining a stable, robust industry ecosystem, all while providing independent, objective supervision and value-added assurance.
Zhu Zhaochen offered a unique perspective, admitting he comes from an 18-year finance background, not technology. He deconstructed the theme into two actions: digitization and transformation. He posed thought-provoking questions about which parts of the business need digitization, which need transformation, and who should be responsible for these actions, setting the stage for a discussion on how the industry's ecosystem and collaborative structures need to evolve.
Ma Jie then directed the first substantive question to Tian Hanyong regarding how smaller banks, with their resource and data constraints, can leverage AI and big data for smart risk control and customer operations. Tian Hanyong proudly cited a digital bill discounting product that achieved transaction volumes in the trillions annually, ranking among the best nationally. He argued that "tech equality" and "AI equality" empower small banks to innovate by focusing on customer pain points. He acknowledged the inherent disadvantages in tech investment compared to large banks but expressed optimism that newer, more efficient open-source models are lowering cost barriers, allowing small banks to cultivate AI awareness and build competencies incrementally rather than attempting whole-business-process overhauls initially.
Tian Hanyong was then asked how smaller banks can build a collaborative fintech ecosystem with tech companies and peers to overcome issues like data silos and compliance balance. He responded candidly that they are still "groping in the dark" for answers, particularly regarding their business model shift. However, he believes that by actively exchanging ideas with tech firms and other platforms, and by learning from each other's experiences, they can adapt new insights and grow together through mutual support and trust.
Turning to Sun Chenyu, Ma Jie asked about applying AI in credit insurance and audit scenarios, especially balancing AI empowerment with prudent risk control in cross-border operations. Sun Chenyu articulated a clear philosophy: "AI assists, humans decide." He detailed a three-tier approach. First, AI handles "full-scale initial screening" using RAG technology and rule-based text analysis to automate rule matching and anomaly detection, reducing work that once took days down to minutes. However, he stressed AI only outputs "audit clues and evidence chains," never conclusions. Second, professional staff perform "professional judgment and skill cultivation," reviewing the clues and distilling their methodologies into reusable 'skills' that feed back into the system, creating a closed loop of "AI screening, human refinement, rule solidification, and AI re-application." Third, the "decision-making power always stays with business personnel," with AI acting as a "senior staff officer" providing data support and logical references, but never crossing the line to become the decision-maker.
Following up, Ma Jie asked why it's crucial for business personnel to learn AI and basic digital thinking. Sun Chenyu passionately argued for shifting from "AI replacing people" to "AI arming people." He explained that by breaking down data silos and creating accessible data assets, institutions can give "digital employees" the fuel they need. He advocated for building business-savvy AI models with clear human-machine boundary design, where AI handles initial vetting and prompts, but qualitative decisions remain with humans. The ultimate goal, as he put it, is a shift in the human role from "executor" to "judge and trainer," allowing each auditor to command a "digital team" of tools. The endgame of digital-intelligent transformation, he concluded, is not "machine replacing human" but "human-machine symbiosis."
Ma Jie then posed a question to Zhu Zhaochen about the practical pathways for AI across investment research, operations, and risk control. Acknowledging the typical resource constraints of smaller brokerages, Zhu Zhaochen shared his company's gritty, reality-driven approach. He emphasized a foundational principle: "cultural leadership" is paramount. By creating a data-driven corporate culture where decisions are based on shared data, the responsibility for digital transformation rests with the business, not just the tech department. All his firm's BI analysts and RPA developers come from business functions. Their current AI strategy is hyper-pragmatic, leveraging free tools and public cloud APIs for experimentation.
He shared a vivid example of an AI-assisted archiving tool developed by business staff using natural language prompts, reducing a project's compliance file archiving time from 2-3 hours to 10 minutes. He also explained how they use AI for investment performance attribution analysis on anonymized, non-sensitive data to help colleagues validate investment logic. Zhu Zhaochen passionately argued for a more collaborative ecosystem. From upstream partners, he hopes for a "trusted financial cloud" - a shared, dedicated space with pooled computing resources so small banks and brokerages don't have to individually invest in expensive infrastructure. Downstream, he envisions a world of interconnected "agents" where his firm's AI directly connects with clients' AI through skill modules and APIs, creating a seamless service delivery layer that transcends individual institutions.
Agreeing with Zhu Zhaochen's points on low-cost AI application, moderator Ma Jie added that limited computational resources shouldn't mean missing the AI opportunity. He stressed the importance of finding application paths that match an institution's own resource endowment and shared his interest in exploring trusted external computing power resources. In a supplementary point, Zhu Zhaochen reinforced this by explaining that very capable, compact language models (around 30B parameters) can perform specific tasks like "intelligent data querying" far more efficiently and accurately than older NL2SQL approaches, and can run on a single additional GPU server, representing a major breakthrough for small firms.
In his capacity as a city commercial banker, Ma Jie shared insights on two fronts. On breaking down data silos, he detailed his bank's long-term commitment to a dedicated data management mechanism and the construction of a data middle platform which has been key to unifying data resources across departments. He stressed that without high-quality "data," reliable "intelligence" is impossible, and shared their practice of using AI to assist in data governance itself. Regarding avoiding "platform-building theater," he emphasized starting from real scenes and real pain points; the goal is solving problems, not applying AI for its own sake. Their "Jizhi Project" was launched based on this principle, evolving from a few key pain-point scenarios into a structured, systemized program.
On ecosystem positioning, Ma Jie stressed a strategy focused on clear cost-benefit analysis, prioritizing high-value, essential-use cases, and not being afraid to choose simpler, mature technologies like RPA when they are the most effective tool. He argued for a "lightweight" approach to model selection, preferring smaller, fit-for-purpose models based on cost and efficiency rather than always chasing the largest parameters. He also voiced a common aspiration among small-to-mid-sized institutions for large financial institutions to share mature experience, common capabilities, and industry knowledge, reducing trial costs and improving transformation efficiency for the whole sector under proper compliance and risk control.
In his final contribution, Zhu Zhaochen delivered a powerful closing thought. He pointed out that "digital" and "intelligence" had been discussed extensively, but "transformation" is often overlooked. He asserted that "all transformation is business transformation." For a company to be successful, business departments must assume the main responsibility for transformation; if only IT is transforming, there is no hope. This sentiment was strongly endorsed by Ma Jie, who concluded that true "business-technology integration" is not driven unilaterally but through business departments' active participation and tech lines' collaborative support, moving the transformation forward together.
Ma Jie concluded the session by thanking all the guests, noting that the financial sector's digital-intelligent transformation is more than an upgrade of technology tools; it's a systematic elevation involving business capabilities, operational models, risk management, and ecological collaboration. He affirmed that financial institutions need to continue exploring deeper integration of digital tech with business scenarios, focusing on quality improvement, risk prevention, serving the real economy, and open cooperation. The 8th Fintech Exchange concluded successfully, with everyone looking forward to reconvening next year.