Beijing, 16 July 2026 — Artificial intelligence (AI) has become a central strategic priority for China's banking industry, but the debate is evolving rapidly. Over the past 18 months, banks have focused on building large language model capabilities, improving productivity and expanding practical applications. Finance China 2026, organised by The Asian Banker, demonstrated that the industry's attention is now shifting towards a broader set of questions: how AI will reshape banking institutions, influence governance, redefine customer relationships and create sustainable business value.

The discussions built on themes first explored at The Asian Banker International AI Finance Summit in Shanghai in March 2026. That event examined how AI infrastructure, economics and deployment models were evolving as banks sought to scale adoption. Since then, advances in reasoning models, agentic AI and open-source platforms have accelerated adoption across the industry, while reinforcing the need for stronger governance, clearer accountability and disciplined implementation. The conversation in Beijing focused less on whether AI could be deployed at scale and more on how banks should integrate AI responsibly into their business models while maintaining customer trust, governance and commercial discipline.

Opening the conference, Cindy Yu, director and head of Greater China at TAB Global, the holding company of The Asian Banker and related digital media platforms, said discussions over the previous 18 months had focused on scaling AI capabilities, while the next phase would examine how AI reshapes banks' revenue models, operating structures and governance. Yu highlighted the National Financial Regulatory Administration's (NFRA) recently issued Guiding Opinions on the Safe Development and Application of Artificial Intelligence in the Banking and Insurance Sectors, describing it as an important milestone for the industry. She noted that one of its central principles, placing responsibility on the financial institution deploying AI, would fundamentally influence how banks approach governance, procurement and accountability.

That regulatory backdrop framed many of the discussions throughout the conference. The NFRA guidance establishes expectations across governance, technology development, data management, risk management and oversight, while requiring institutions to maintain meaningful human supervision over high-risk AI applications. It also encourages larger financial institutions to share AI infrastructure and capabilities with smaller organisations, signalling that AI development is increasingly being viewed as an industry capability rather than solely an institutional one.

Providing the broader economic context, John Gong, professor of economics at the University of International Business and Economics, argued that the global economy is being shaped by two powerful forces: geopolitical conflict and technological advancement, particularly artificial intelligence. While geopolitical uncertainty continues to affect trade and investment, he said sustained spending on AI infrastructure by China, the United States and other major economies is likely to remain one of the defining features of global economic development.

"The world today is being shaped by two major forces—war and technological development, particularly AI," Gong said. "Different economies are affected differently depending on their exposure to these two forces."

Gong argued that China is comparatively well positioned because of sustained investment in advanced technologies, rapid AI development and continued industrial upgrading. As China's economy matures, he suggested, productivity growth and technological capability will become increasingly important drivers of long-term competitiveness, while evolving trade relationships and geopolitical tensions will continue to shape the operating environment for Chinese financial institutions.

Building on that context, Emmanuel Daniel, founder and chairman of TAB Global, drew on themes from his latest book, Building the AI Bank, to argue that banks should avoid treating AI simply as the next stage of digital banking. While digital banking largely digitised existing processes, Daniel said AI fundamentally changes the institution itself by redistributing decision-making, reshaping customer interaction and requiring new governance structures. As AI agents become capable of initiating transactions, making decisions and interacting autonomously, banks will need to rethink organisational responsibilities, payment architectures and accountability frameworks rather than simply adding AI to existing operating models.

Daniel also argued that customer relationships are entering a new phase in which intelligent agents increasingly act on behalf of customers, reducing dependence on traditional banking channels and requiring institutions to integrate with broader digital ecosystems. Developments such as programmable money, tokenised deposits and AI-native financial institutions, he suggested, illustrate how banking infrastructure itself is likely to evolve alongside artificial intelligence.

Evidence presented by TABInsights, the research, benchmarking and consulting arm of TAB Global, suggested that many Chinese banks have already begun this transition. Presenting the firm's annual research, Hugh Zeng, research manager at TAB Global, said submissions for the TAB Global 2026 Excellence in Retail Finance and Financial Technology Innovation Awards showed AI deployment expanding rapidly beyond software development and employee productivity into customer servicing, compliance, operations, risk management and enterprise decision-making. Agentic AI featured in more than 60% of AI-related submissions from Chinese institutions, reflecting how they are moving beyond experimentation into operational deployment. Zeng said the research suggested that technology is no longer the principal constraint on AI adoption. While many institutions have established the computing infrastructure and model capabilities needed for enterprise deployment, the research found that AI-ready data remains uneven across the industry. Leading institutions have achieved significantly higher levels of data readiness, but many banks continue to strengthen data governance, integration and quality before AI can deploy consistently across business functions.

Together, the keynote presentations, research briefing and leadership roundtable illustrated how the industry's priorities have evolved within a matter of months. The debate has moved beyond experimentation towards the organisational, governance and customer questions that will determine how AI reshapes banking over the coming decad.

AI banking enters an operational phase

If the first phase of AI adoption focused on proving the technology, the discussion among banking leaders showed that the industry's priorities have shifted towards embedding AI into core business processes while establishing the governance needed to manage increasingly autonomous systems.

The discussion began with retail banking, where Chen Shenlong, director of digital technology, retail banking at China Minsheng Bank, argued that AI's greatest opportunity lies in enabling genuinely personalised banking rather than automating existing products and processes. Traditional retail banking has largely relied on standardised products because institutions lacked the ability to analyse large volumes of contextual customer information. AI, particularly reasoning models, allows banks to incorporate a broader range of customer information into credit assessment, product recommendations and service delivery.

Chen said the same capabilities are reshaping risk management. While conventional statistical models remain indispensable because they are explainable and well understood by regulators, reasoning models provide banks with additional tools to interpret rapidly changing market conditions and customer behaviour. He said AI is already demonstrating value in fraud detection and anti-financial crime applications, while creating opportunities to improve credit decisions through richer analysis rather than larger datasets alone.

Looking beyond individual use cases, Liu Zhan, deputy general manager of the retail credit department at Shanghai Pudong Development Bank (SPDB), said AI is forcing banks to rethink how retail financial services are designed and delivered. Personalisation requires far more interaction between customers and banks than traditional standardised products allow, making AI an important enabler of customer engagement rather than simply a productivity tool. He suggested that the next generation of retail banking will increasingly depend on AI's ability to understand customer needs, analyse diverse information sources and support more individualised financial services.

For Eason Wang, assistant general manager for digital finance development and AI architecture expert at WeBank, the more immediate challenge is preparing institutions themselves for AI. He argued that banks should build flexible engineering architectures capable of adapting as AI technologies evolve, rather than optimising around today's interfaces or models. WeBank has focused on secure deployment through AI sandboxes, penetration testing and controlled production environments, recognising that increasingly capable AI agents also introduce new forms of cybersecurity and operational risk.

The discussion suggested that AI is no longer simply another technology project. As institutions expand deployment from software development into customer servicing, operations, compliance and decision support, AI is beginning to reshape operating models across the front, middle and back office. The challenge for banks is becoming less about deploying models and more about redesigning organisations so that AI can be deployed safely and consistently at enterprise scale.

Governance emerged as the central theme of the roundtable. Jin Haimin, deputy general manager of the fintech department at ICBC, said the NFRA's Guiding Opinions on the Safe Development and Application of Artificial Intelligence in the Banking and Insurance Sectors establish an important principle by integrating AI governance into existing enterprise risk management rather than creating a separate governance framework. He argued that responsibility should remain aligned with business ownership, while technology teams remain accountable for model infrastructure, data quality and platform resilience. He said the principle that institutions remain responsible for AI outcomes provides a clear foundation for accountability as deployment accelerates.

Jin also emphasised that regulatory expectations remain particularly high for customer-facing and other high-risk applications. Human oversight, continuous monitoring and clearly defined intervention mechanisms remain essential where AI contributes to significant business decisions. Rather than replacing existing governance, he argued, AI requires institutions to strengthen governance across the entire lifecycle of model development, deployment and operation.

Ding Yan, deputy general manager of the intelligent operations center at China Everbright Bank, emphasised that governance frameworks should enable institutions to continue experimenting with AI rather than discourage deployment. He argued that institutions must balance prudent risk management with business experimentation if AI is to move beyond isolated pilot projects. In his view, organisational commitment and executive sponsorship are as important as technology, because successful AI deployment requires changes to operating models, responsibilities and decision-making processes across the institution.

Offering the perspective of an international bank, Kevin Feng, chief technology officer of E.SUN Bank (China), said explainability remains a critical consideration for customer-facing AI. His bank has therefore focused AI deployment on internal operations before extending its use to customer interactions, ensuring that decisions affecting customers remain understandable and subject to appropriate human judgement. While AI is improving efficiency across banking operations, he argued that governance frameworks must mature alongside the technology if institutions are to maintain regulatory compliance.

The discussion also highlighted opportunities for industry collaboration. Zhang Yanming, general manager of the trade finance Department at Jiangsu Su Merchants Bank, suggested that AI could strengthen corporate banking through broader data sharing and greater access to customer-authorised data. Better information, he argued, would improve enterprise financing and strengthen risk assessment, especially for smaller businesses.

Jin proposed that collaboration should also extend to AI safety, including common testing methodologies, evaluation benchmarks and security standards. Wang added that banks could benefit from greater information sharing on cybersecurity threats, model vulnerabilities and emerging attack techniques, particularly as increasingly autonomous AI systems interact across institutions and digital ecosystems.

Gong added that AI should also simplify banks' own internal processes. Many customer interactions, he suggested, remain unnecessarily complex despite years of digital transformation. AI offers an opportunity not only to improve operational efficiency but also to shorten processing times, simplify procedures and deliver more responsive banking services.

Taken together, the discussion suggested that competitive advantage will depend less on who deploys AI first than on who integrates it most effectively into institutional governance, operating models and customer experience. As Chinese banks move from experimentation towards enterprise deployment, success will increasingly be measured by execution, accountability and the ability to translate AI capability into sustainable business outcomes.

From experimentation to competitive advantage

While many questions surrounding artificial intelligence remain unresolved, the discussions at Finance China 2026 suggested that China's banking industry has entered a distinctly different phase of adoption. The conversation is no longer centred on whether AI works or how many applications banks have deployed. Instead, attention is turning to how institutions embed AI into their operating models, strengthen governance, prepare enterprise data and redesign organisations to support increasingly intelligent systems.

The discussions also reflected a broader shift in thinking that has emerged over the past few months. At The Asian Banker International AI Finance Summit in Shanghai, the industry's focus was on whether AI could be scaled economically as infrastructure investment, energy consumption and inference costs escalated. By the time banking leaders gathered in Beijing, rapid advances in reasoning models and deployment techniques had moved the debate forward. The more pressing challenge has become how institutions can deploy AI responsibly, deliver measurable business value and maintain customer trust as AI becomes integral to banking operations.

The conference also highlighted the increasingly close relationship between technological innovation and regulatory development. China's AI banking agenda is evolving alongside a regulatory framework that seeks to encourage innovation while reinforcing institutional accountability, human oversight and sound risk management. That combination provides banks with greater clarity as they move from pilot programmes to enterprise deployment, while recognising that governance will remain central as AI capabilities continue to advance.

For TAB Global, the discussions in Beijing form part of a continuing industry dialogue rather than a standalone event. The themes explored at the conference will continue through its research, benchmarking programmes and leadership dialogues, culminating in the AIX World Awards and the AIX World Summit later this year. Together, these platforms will continue to examine how financial institutions are translating AI investment into stronger customer outcomes, more resilient operating models and sustainable competitive advantage as the industry enters the next phase of AI-enabled banking.