Luo Aihua: With simplicity to perfection, low code empowers the digital transformation of Agricultural Bank of China

It is predicted that by 2024, more than 65% of application development activities will be carried out in a low-code mode, which will burst out of great development potential in promoting full innovation, ecological integration, and digital twins. This paper takes the digital transformation of commercial banks enabled by low code as the starting point, summarizes the current situation of low code application in combination with the case study of Agricultural Bank of China, and finally looks forward to the prospect of low code application of commercial banks and makes suggestions.

Luo Aihua, Deputy General Manager of the Network Finance Department of the Agricultural Bank of China

Low code powers banks' digital transformation

As a modern application development model, low-code development has the characteristics of model-driven, visual guidance, self-service development, and open integration. Its emergence provides commercial banks in the tide of digital transformation with "fast, accurate, provincial and easy" curve overtaking tools.

1. Fast - achieve market agile response

At present, "only fast is not broken, only change is not broken" has become a killer of competition in the same industry, but it is difficult for headquarters-style research and development and complex system coupling to fully and quickly meet the various personalized needs of front-line operating institutions. The low-code development model supports the agile collaborative development of the total and branch through reusable transaction models, functional components, and business processes, greatly shortening the project delivery cycle and enabling rapid response to the market.

2. Quasi--promote the deep integration of industry and technology

Under the background of digital transformation, how to quickly put into production of new products with "what you think is what you get, what you see is what you get" has become an urgent problem for business personnel to solve. The low-code platform visualizes abstract requirements through structured data and visualization processes to reduce the communication cost between business and technology, greatly improves the fit between software systems and business needs, and enables product functions to more accurately meet market customer demands.

3. Province - effectively reduce the cost of innovation

The surge in demand has led to an increase in R&D and labor costs, and the complexity of the system and the mobility of personnel have caused a serious shortage of technical personnel's skills depth and breadth, which has further increased the cost of innovation. Through the encapsulation of high-quality code components, the workload of code development can be minimized, the collaborative efficiency can be improved, and the operation and maintenance costs caused by code quality can be reduced while reducing labor costs and R&D costs.

4. Easy - to help the development of open banking

Open banking has now become the main channel for scene acquisition and ecological integration. However, due to the uneven level of technology development of first-line branches and partners, coupled with the relatively complex business logic of financial products, the docking cycle is long and the threshold for cooperation is high. The use of low-code services to package components and customize templates according to scenarios can help branches and partners achieve rapid assembly and docking, and greatly improve the open integration capabilities of commercial banks.

Practice exploration of low-code business scenarios

With the acceleration of the digital transformation process, ABC fully recognizes the innovation advantages and development potential brought by the low-code development model, and has successively infiltrated the low-code concept into business statistical monitoring, intelligent model construction, product customization assembly, open banking docking and other related fields, and continuously explored the agile innovation model of low-code empowerment industry technology integration through total and sub-linkage.

1. Monitoring of business statistics

The application of the ABC's low-code development model in the field of statistical monitoring is the earliest and most mature, and it has basically reached the low-code implementation of the whole process from data import, parameter configuration to front-end interaction. In 2009, the ABC Credit Management System Group Report Statistical Tool- Credit Data Express, has introduced the concept of low-code development. Data maintenance personnel load Cubes with interactive forms, configure data models, business personnel based on Report Studio and Query Studio drag and drop combination of fields, flexible customization of reports, one-click rapid release, effectively meet the bank's credit risk monitoring, business decisions and other daily management needs, which is also the earliest practice of ABC in the field of low code.

In recent years, in order to achieve flexible statistical analysis covering all businesses, ABC has released a low-code platform for self-service query and analysis, a business intelligence data service platform. The platform further simplifies the data configuration process and enables system administrators to quickly import low-level wide table data and interactive parameter configuration. The front-end of the platform is based on the VUE framework to greatly improve the interactive experience, provide business personnel with more intelligent and friendly flexible query, report customization and Kanban assembly services, and support the rapid release of cross-system reports and Kanban boards. The platform now supports multi-dimensional query and report customization for retail banking, scenario finance, online payment and other services.

2. Intelligent model building

With the rapid development of financial technology, the demand for big data and artificial intelligence in various lines within the Agricultural Bank of China has shown an upward trend. In order to reduce the threshold of technology application, ABC explores the construction of intelligent model laboratory in accordance with the concept of low-code development, supports data analysts to drag and drop components on the canvas, configure processes, and initially realize interactive modeling such as data exploration, feature processing, feature engineering, model construction, model evaluation, and one-click release of models. The platform takes into account the needs of business personnel and R&D personnel, supports the drag-and-drop combination workflow of the basic meta-operation node, and supports the free programming of the command line to design the workflow. Since its launch, the platform has assisted the general and branch data analysts to explore and build a number of intelligent models in the fields of smart marketing, smart risk control, and smart operation, accelerated the process of "data businessization and business data", and promoted the efficient empowerment of data to innovate and develop business.

3. Customized assembly of products

In order to reduce O&M costs and accelerate the efficiency of product innovation, ABC develops a comprehensive application platform flowchart mode, which adopts the low-code method of "component + process configuration" for transaction logic development for background transactions and product micro-innovation, quickly assembles transactions with flowchart drag, and supports process nesting to build complex transaction logic. This model not only speeds up the development of new products, but also reduces the error rate of complex transaction coding. Taking the digital RMB two-dimensional code payment product as an example, the Agricultural Bank of China uses the relevant functional components such as two-dimensional code generation and account processing precipitated by the original RMB two-dimensional code application platform system to assemble the process and configure the parameters through the flow chart mode, reuse the complex accounting processing logic packaged by the components, and quickly and at low cost in the form of low code to realize the production and launch of the digital RMB two-dimensional code payment module, effectively helping to seize the market opportunity of the digital renminbi payment scene.

4. Open scene docking

In the early stage, in the field of open banking, ABC mainly relied on branch expansion scenarios and head offices for system docking, so there was a process bottleneck. In order to enhance the technical research and development capabilities of branches, ABC launched a low-code development platform based on the enterprise-level architecture- light cloud platform, which supports branch developers to complete business logic and process construction through a graphical interface by dragging and dropping components and configuration parameters, effectively promoting agile innovation and open customer development on the mobile terminal of branches. Taking a branch as an example, in order to solve the customized needs of financing products in various cities under its jurisdiction, the branch applied low-code development technology to encapsulate the characteristic SDK components of the component line such as identity recognition and quota measurement, and used template technology to standardize the pre-processing of the documents required for each process, realizing the rapid innovation and combination of new products, and greatly improving the docking efficiency with the government affairs platform.

Future application outlook and recommendations

Overall, at this stage, the more mature low-code development practices are mainly concentrated in the internal management fields such as data analysis, and the application in the core areas of product innovation is still in its infancy, and there is still great potential for development in the future. At the same time, in the application process, how to deal with the three balances of versatility and flexibility, business and technology, agility and stability is also a difficult problem that we must solve.

1. Development opportunities

The first is to achieve technological democracy and promote innovation for all. The low-code research and development model has greatly reduced the threshold of product innovation, truly democratized technology, and has great application potential for large state-owned commercial banks with rich personnel. Business personnel can easily learn to assemble products by themselves, and quickly meet customer customization needs with micro-innovation, so as to create a more powerful innovation-driven engine through full-staff co-creation. The second is to achieve Bank 4.0 and accelerate ecological integration. Different from Amazon, Ali and other Internet giants to output low-code services, the bank's low-code platform is still mainly used to empower internal branch personnel, and the follow-up can provide open services to partners through low-code platforms, supporting partners to facilitate operations such as scene template import and graphic form customization, which will effectively break through the bottleneck of partners' technical capabilities. The third is to achieve digital twins and support accurate decision-making. Based on the low-code data application platform, it can effectively solve the time-lag bottleneck of digital twins. According to the needs of real-time risk monitoring, business statistical analysis, and user intelligent operation, the data analyst of the front desk department can quickly assemble business indicators and customer profiles in real time, optimize them with one click, and carry out high-real-time digital dynamic simulation mapping for services and customers, providing a comprehensive and accurate decision-making basis for business decisions.

2. Challenges and suggestions

First, the balance between versatility and flexibility. The construction of the low-code platform is based on component packaging and process cutting, and the definition of component functions and process links will directly affect the flexibility and feasibility of R&D and self-service integration of business personnel. It is recommended that based on market and business needs, business experts and technicians work together to define and encapsulate functional components, decouple basic functions and business rules, and support the necessary parameter customization through the rule engine library.

Second, there is a balance between business and technology. At this stage, some low-code platforms have the problem of unclear definition of user portraits, resulting in the technical side suspecting that the packaging is too versatile and flexible, the business side suspects that the configuration logic is too much and not easy to understand, and both the industry and technology are not used. It is suggested that when planning and construction, commercial banks should clarify the positioning and goals of low-code platforms, minimize learning costs with code-free operations for business application platforms, and support configuration to improve flexibility for collaborative development platforms.

Third, the balance between agility and stability. For commercial banks, due to the lack of practical testing, the stability and security impact of the low-code development model on the core key financial systems with high concurrency (such as palm banking, online banking, etc.) still need to be further evaluated. It is recommended that in the actual promotion, start from the business areas with small transaction volume, small concurrency volume and few systems, and control the impact of the pilot with mechanisms such as grayscale release, and steadily expand the scope of application.

(Column Editor: Han Weimi)