On October 15, 2026, the 4th Cross-Border Digital Finance Industry Compliance and Coexistence Conference was successfully held, bringing together nearly 300 professionals from across the cross-border finance industry chain to deeply discuss hot topics such as industry compliance, new business models, and emerging markets, with AI as the technological focus.Article author and source: Mobile Payment News
The New Normal of AI: Compliance Comes First
The forum "Compliance and the New Normal," moderated by Hong Peilin, Assistant Director and Department Head of the Shenzhen E-Commerce Service Center, explored new compliance requirements in the age of AI.
Mu Chu, Chief Editor of Mobile Payment News & Outbound Journal, said in his opening remarks that, in the new normal where AI has become integrated into people’s daily lives and work, in-person events are more important than ever for fostering communication and exchange within vertical industries.
Shuai Yihua, Vice President of Products and Solutions for Mastercard Greater China, analyzed the opportunities and challenges in cross-border agent-based payments. The market expects the global agent-based commerce economy to reach $3.5 trillion by 2030. In this context, consumer behavior is rapidly shifting from traditional search engines to using AI agents to discover and purchase products. As traditional commerce evolves into agent-based commerce, merchants and PSPs face both opportunities and challenges—namely, how to enhance settlement experiences and protect their businesses while leveraging existing payment infrastructure.
Zhang Tianyi, Chief Trust and Security Officer of Ant International, shared Ant International’s security insights for the age of intelligent agents, drawing on years of hands-on experience in offensive and defensive security. In the AI era, beyond amplified traditional risks such as identity spoofing and AI-powered, customized cross-border fraud, new risks have emerged from the practical implementation of agent-based payments—including malicious exploitation of agents, unclear accountability, and disputes over transaction liability. In response, Ant International has strengthened its efforts across three areas: account protection, payment protection, and agent protection. It leverages Anti-Deepfake technology to reinforce account security; deploys its proprietary payment foundation large model, Antom 3-in-1 Transformer, to enhance risk control efficiency and accelerate the deployment of new risk scenarios; and establishes the KYA (Know Your Agent) intelligent agent trust framework to enable continuous verification and mutual recognition of certifications, while introducing a fund protection mechanism to cover losses from agent failures—comprehensively safeguarding the development of agent-based commerce.
Tang Yingxiong, Compliance Manager at PalmPay Group, presented on navigating digital finance compliance in Africa’s diverse markets. In African markets, the same product may fall under entirely different regulatory boundaries depending on the country or regional integration framework. Therefore, entering a new market requires following a six-step compliance lifecycle: defining the product and cash flows, identifying regulatory boundaries, selecting the entity and license type, completing approvals and conditions, implementing controls and training, and ongoing reporting and remediation.
Tong Ying, Strategic Lead for Risk Information Greater China at LexisNexis, analyzed payment transparency, risk management, and AI empowerment under the new phase of travel rules. Current requirements for payment transparency are evolving from merely transmitting information alongside transactions toward ensuring information quality, recipient verification, and payment security. Moreover, the application of payment transparency principles varies across markets in terms of applicable institutions, regulatory pathways, and operational requirements. Companies must determine their specific obligations by considering their market, actual business activities, and their role within the funds or value transfer chain.
Zhang Yi, Head of Product at ADVANCE.AI, introduced how to build an end-to-end identity trust system for cross-border finance from the perspective of "Know Your Activity." In cross-border financial risk prevention, KYC/KYB verification is merely the starting point—real risks persist throughout the customer’s entire lifecycle. "Know Your Activity" fills the missing unified identity trust layer between fraud and AML. Financial institutions are advised to establish a solid foundation in KYC/KYB to enable seamless integration with future transaction monitoring systems.
Liu Zhiyao, Head of Business at Songliang Technology, analyzed the construction of financial communications compliance in the globalization of Chinese financial services. For payment companies expanding overseas, communications are not a cost item but a risk control and compliance requirement, as well as a brand asset. Any communication delay at any stage of mobile payments can impact user experience and conversion rates, making verification codes a critical lifeline for business operations. In terms of specific AI innovations, AI voice agents are emerging as "digital sales teams + digital collections teams" for overseas lending services.
Sha Xuan, Director of the Dun & Bradstreet Products and Solutions Division and Business Line Lead, explained how the Trusted Business Graph is reshaping AI transformation for global payment institutions. The five core business flows of payment service providers—payment flow, enterprise flow, risk flow, compliance flow, and business outcomes—share a common set of atomic capabilities, making them naturally ideal for agentization. Without a graph, there can be no trusted identity; the business graph connects shareholders, directors, UBOs, and risk signals, enabling payment service providers to better achieve AI transformation.
Explore new business models with deep AI applications
Moderated by Cao Zhongyuan, Secretary-General of the Hong Kong FinTech Association, the discussion "AI Applications and New Business Models" explored how AI can be integrated into real-world cross-border financial scenarios.
Zhang Dewang, Vice President of LianLian Digital Technology, analyzed how AI-driven cross-border collaboration is transforming digital infrastructure. AI is no longer merely a tool—it has become a company’s new digital employee. Today, when companies expand overseas, they are not just exporting products and brands, but also their AI capabilities. For agents to operate in the real world, they require a comprehensive financial infrastructure foundation, including accounts, wallets, payments, collections, currency exchange, credit, and fund management. Building on this, LianLian Digital is developing a new type of infrastructure that enables intelligent systems to be trusted, authorized, and transacted with.
DTCPay Technology VP Tong Bin introduced how task-level data permissions enable controlled AI execution. As AI capabilities improve and internal platform support strengthens, AI agents are now capable of handling medium- to high-complexity delivery tasks. When AI agents execute business tasks, they need to access data across various business systems, which can be hindered by factors such as data fragmentation, credential access restrictions, and coarse-grained authorization. In this context, establishing a centralized permission management system to uniformly manage people, AI agents, and services, while standardizing access methods to systems, becomes critically important.
Lucas Yang, Head of Sales and Solutions at Cobo, analyzed how AI and security are reshaping the next generation of payments. On-chain AI payments have already emerged, and the key to this ecosystem's development lies in authorization. Relevant reports show that among 19 AI products, 14 operate within pre-set limits established by individuals or enterprises, indicating that restricted authorization is already common. In areas such as pay-per-use AI services, real-time cross-border payments, and AI reconciliation, digital currencies have demonstrated strong product and scenario suitability.
Li Jie, Head of Sumsub's China operations, analyzed how the cross-border financial risk control system is being restructured in the age of AI agents. In the past, the core issue addressed by digital financial risk control systems was: “Who are you?” But as AI begins to act on behalf of individuals, risk control must now focus not only on customer identity but also on the behavior and authorization relationships of agents. In this context, KYA (Know Your Agent) becomes critically important. KYA seeks to answer: Who or what is actually acting on your behalf?
Yuan Qi, Senior Engineer and Director at the China Academy of Information and Communications Technology, analyzed the core security risks of AI agents and the construction of a protection system. AI terminals are the direct vehicles for the large-scale deployment of AI, significantly differing from traditional smart terminals in service delivery, collaboration, and interaction modes, and are driving a new wave of technological innovation and industrial transformation in consumer electronics. While AI terminals greatly enhance intelligence levels, they also face novel and complex security challenges, including risks related to agent security, data security, and personal information security. To strengthen the security protection of AI terminals, it is necessary to establish a "real-time monitoring + precise interception" defense mechanism across four core layers—perception, model, orchestration and planning, and execution—to effectively improve the security posture of terminal AI agents.
Zhang Yanzhao, Head of Greater China Operations at the Global Legal Entity Identifier Foundation (GLEIF), shared international advancements and the trust foundation for agent-based payments. From user authorization and intent understanding to acquiring a digital identity, agent-based payments are gradually becoming integrated into everyday life. Currently, agent-based payments operate in three models: agents invoking merchant APIs, collaboration between agent and merchant agents, and pay-per-use APIs with on-chain settlement. This transformation driven by these three models requires all parties involved in transactions to adapt their business models and processes.
Partner Luo Yuting from the Haihua Yongtai Law Firm in Shanghai analyzed cross-border payments and financial compliance under the AI-driven transformation of global retail. Over the past two decades, global retail businesses have undergone two significant shifts: first, digitalization; and second, the current AI-driven transformation, which is redefining global retail transaction models—not merely through technological upgrades, but through the redesign of business models and transaction processes. Changes in the retail industry will ultimately impact the payments sector. Payments are not a simple post-transaction step, but a critical infrastructure connecting consumers, merchants, and the financial system, and these changes have introduced new requirements for cross-border payment compliance.
Yu Jianing, Chief Compliance Officer of MetaComp, analyzed how to build trustworthy, controllable, and interoperable institutional-grade AI Finance. The explosive growth of agent-based AI is transforming financial services, evolving from human-to-human, human-to-application, and human-to-API interactions toward human-to-agent and agent-to-agent interactions. As AI agents begin to execute transactions and make autonomous decisions on behalf of enterprises, the industry urgently needs to establish a new framework encompassing identity verification, authorization scopes, risk control, and compliance governance.
During the roundtable discussion, Zhang Dewang, Li Yilong, Chief Technology Officer of Xunhui SunRate AI, and Tong Bin explored the tangible changes AI brings to the cross-border finance industry, the opportunities and challenges of agent payments, priority levels for agent payment investments, the relationship between AI and developer responsibilities, and the future development of AI.
