Industrialising AI in banking requires both enterprise platforms and knowledge systems

Industrialising AI in banking requires both enterprise platforms and knowledge systems

The Asian Banker Shanghai International AI Finance Summit examined how financial institutions are beginning to industrialise AI across customer engagement, enterprise technology infrastructure and knowledge-driven capability systems that combine human expertise with institutional knowledge.

Artificial intelligence (AI) has been used in banking for more than a decade, but much of its early development took place through experimentation rather than large-scale operational deployment. Financial institutions initially introduced machine learning models into specific areas such as fraud detection, credit scoring and customer analytics. These early initiatives produced measurable improvements in operational efficiency, but they often remained isolated within individual departments rather than transforming the organisation as a whole.

The next phase of development may therefore involve a different challenge. Financial institutions are now examining how AI can be deployed across enterprise platforms that support multiple operational functions simultaneously. Instead of focusing on isolated use cases, institutions are exploring how AI systems can become embedded within core banking processes and decision systems.

Industrialising AI requires more than technological innovation. Banks must develop enterprise data architectures capable of integrating information from multiple internal systems while maintaining consistency, security and accessibility. They must also build computing infrastructure capable of supporting large-scale analytical workloads and advanced AI models.

Equally important is organisational alignment. AI systems can only create value when their outputs can be integrated into operational decision-making processes. This requires close collaboration between technology teams, business units and operational leadership across the institution.

These issues were explored during sessions at The Asian Banker Shanghai International AI Finance Summit, where speakers discussed how financial institutions are beginning to operationalise AI at scale. The discussion highlighted the interaction between enterprise infrastructure, customer engagement systems and organisational capability development.

AI is changing how banks engage customers

Lv Chengze, head of retail AI, research and development centre, Industrial and Commercial Bank of China (ICBC), discussed how AI may reshape customer engagement within retail banking.

Retail banking has traditionally relied on segmentation and campaign-based marketing. Customers were grouped into demographic or behavioural segments, and financial institutions launched periodic marketing campaigns designed to promote specific products. While this approach allowed banks to reach large numbers of customers, it often lacked precision and did not fully reflect the complexity of individual financial needs.

AI systems may allow banks to move beyond this model by analysing customer behaviour continuously rather than episodically. Instead of waiting for scheduled campaigns, institutions may detect patterns in customer activity and identify emerging financial needs in real time.

Lv explained that AI can analyse large volumes of behavioural data generated through digital banking platforms, mobile applications and payment transactions. These data streams provide insights into how customers interact with financial services and may help banks anticipate future needs.
Such capabilities may allow banks to move towards a model of continuous engagement. In this framework, financial institutions interact with customers through ongoing digital channels informed by behavioural data and predictive analytics rather than relying on periodic marketing campaigns.

Enterprise infrastructure enables AI at scale

Deploying AI across retail banking operations requires substantial investment in enterprise technology infrastructure. Large financial institutions must build computing environments capable of processing vast volumes of data while maintaining operational reliability and security.

Institutions such as ICBC operate at enormous scale, serving more than seven hundred million customers and processing huge volumes of financial transactions every day. Analysing such information in real time requires high-performance computing platforms capable of supporting advanced analytical models and large-scale data processing.

Lv emphasised the importance of enterprise data architecture in enabling AI deployment. Banks must integrate information from multiple operational systems, including payments platforms, customer relationship management systems and digital banking applications.

Financial institutions must also develop model engineering capabilities to ensure that AI models remain reliable over time. Analytical systems require continuous monitoring, retraining and performance evaluation to ensure that models remain aligned with changing customer behaviour and market conditions.

This transition represents a broader shift in banking technology. Rather than deploying isolated digital tools, institutions are increasingly building enterprise AI platforms capable of supporting multiple operational functions simultaneously.

Talent and knowledge systems become critical capabilities

While technology infrastructure is essential for scaling AI, financial institutions must also address organisational capability development. Deploying AI across enterprise systems requires more than technical infrastructure; it requires the ability to integrate knowledge, expertise and operational processes.

Sun Jiajia, senior director, solutions, Ping An ZhiNiao, discussed how financial institutions are approaching this challenge through AI-enabled learning and knowledge systems. His presentation emphasised that the concept of talent is evolving in the context of AI transformation.

In traditional organisations, talent primarily referred to individual employees and their professional expertise. In AI-driven organisations, however, capability increasingly includes both human expertise and institutional knowledge systems that allow organisations to capture, structure and deploy expertise at scale.

Ping An has therefore developed digital platforms that function both as learning systems and knowledge hubs. These platforms organise institutional expertise and make it accessible to employees across the organisation while also supporting AI-assisted decision systems.

This approach allows organisations to combine human expertise with structured knowledge systems and AI tools, enabling both employees and AI systems to access institutional knowledge more efficiently.

AI learning platforms support workforce transformation

AI-enabled learning systems are beginning to reshape how financial institutions approach workforce development and organisational capability building. Traditional corporate training programmes typically rely on standardised courses delivered to large groups of employees regardless of individual skill requirements.

AI learning platforms instead analyse individual learning behaviour and recommend personalised training pathways. This allows organisations to tailor training programmes to the specific needs of each employee while ensuring that workforce capabilities evolve alongside technological development.
At Ping An, digital learning systems support tens of thousands of employees and organise training programmes across more than 170 specialised learning tracks covering areas such as finance, technology, data analysis and operational management.

These platforms also function as knowledge hubs that organise institutional expertise. By structuring internal knowledge and integrating it with AI tools, financial institutions may create decision-support environments in which both employees and AI systems can access shared knowledge resources.
In practice, these systems increasingly extend beyond learning. Organisations are training large numbers of employees to develop AI-powered assistants capable of automating routine operational tasks. Across the organisation, between 30,000 and 50,000 employees have been trained to build such assistants, creating more than 40,000 AI tools that support activities such as document summarisation, information retrieval and workflow automation.

The impact of these initiatives is measured through productivity improvements rather than simply through training participation. According to Lv, these AI-enabled assistants have already generated more than 80,000 man-days of productivity gains, illustrating how AI knowledge systems may augment or partially replace manual processes.

Such deployments illustrate the emergence of what the technology industry increasingly describes as agentic AI. In these systems, software agents perform defined operational tasks on behalf of users. Emerging orchestration frameworks such as OpenClaw.ai and similar technologies demonstrate how large numbers of AI assistants may eventually be coordinated across enterprise environments.

Industrialising AI across financial institutions

The discussions at the summit suggested that the next phase of AI development in finance will focus on industrialisation rather than experimentation. Financial institutions have already demonstrated that AI can support individual operational functions, but the challenge now lies in integrating these capabilities across entire organisations.

Industrialising AI requires coordinated investment in enterprise infrastructure, data systems and organisational capability development. Institutions must build computing platforms capable of supporting large-scale analytical workloads while also ensuring that AI systems operate reliably within regulated financial environments.

Workforce capability development remains equally important. Employees must understand how AI systems operate, how analytical outputs should be interpreted and how these insights can be integrated into operational processes.

China provides one of the most advanced environments for observing this transition. Financial institutions operate at exceptional scale and are embedded within digital ecosystems that generate vast volumes of behavioural data through e-commerce platforms, payment systems and digital services.

AI may therefore reshape banking in many ways. However, the institutions most likely to succeed will be those able to combine enterprise technology infrastructure, integrated data systems and knowledge-driven organisational capability into coherent operational platforms capable of supporting AI deployment at scale.

Comments (0)
Cancel