How Commercial Banks Can Leverage Low‑Code in the AI Large‑Model Era: Pathways and Key Considerations

1. Digital‑Intelligence Shift: New Challenges for Core Banking Systems

Rapid AI model innovation (e.g., DeepSeek, ChatGPT, Manus AI) and accelerating digital‑economy growth demand banks to reshape IT architecture, achieve agile development, and deploy new applications quickly. Six concrete challenges are identified:

Limited technology resources hinder rapid, diversified business‑innovation.

Outdated development and operation models cause high labor costs, long cycles, poor component reuse, and rigid architectures.

AI‑assisted development hurdles such as code‑generation quality, security, explainability, and IP concerns.

Insufficient data‑governance leads to data silos, low quality, and duplicated data.

Information‑security pressures from AI, cloud, and big‑data integration create new leakage and API‑security risks.

Technology‑compatibility and ecosystem challenges under the "Xinchuang" (domestic‑innovation) push, especially legacy system integration.

2. Innovation Path: Building an Integrated Digital Foundation with Low‑Code

Low‑code platforms, with visual, model‑driven, open‑integration, and self‑service features, enable banks—especially smaller regional ones—to improve agility and innovation efficiency. The Chinese low‑code market, launched in 2019, has entered a high‑speed growth phase. The 2024 Low‑Code Platform Research Report predicts the market will reach CNY 186.5 billion by 2028, a 36.8% CAGR over five years.

According to the same report, domestic vendors have rapidly closed the gap with overseas leaders in both technology iteration and market adoption. The report’s quadrant diagram (Figure 2) shows local players such as Puyuan Information and Yonyou occupying the “Leader” quadrant, while NetEase‑Shufan and Tencent‑Weida sit in the “Innovator” quadrant. International platforms like OutSystems and Mendix remain in the “Innovator” quadrant due to localization and cost constraints.

3. Eight Key Considerations for Low‑Code Adoption in Banks

1. Fuse professional coding with low‑code – platforms must support script extensions and hard‑code integration to handle complex scenarios while retaining visual, low‑code efficiency for routine tasks.

2. Implement data‑driven development – model‑based design turns business logic into visual data structures, enabling rapid data‑centric app creation and deeper data value extraction.

3. Strengthen component‑based asset governance – encapsulate reusable business processes as components, govern their lifecycle, and avoid duplicate development.

4. Build a digital enterprise‑process hub – a unified process center integrates low‑code, high‑code, workflow engines, and integration platforms to achieve end‑to‑end, compliant, and efficient business flows.

5. Create an application‑connection center – ensure seamless service‑and‑data integration across existing IT systems, cloud services, and third‑party apps.

6. Establish full‑stack development‑to‑delivery governance – visual tools and pre‑built modules support continuous management of development, testing, deployment, quality, and security.

7. Deploy a one‑stop application‑governance hub – consolidate authentication, permission, service, data, and component management to improve governance efficiency and security.

8. Enable a self‑controlled full‑stack platform – support front‑end, back‑end, database, API, and cloud‑native capabilities, and provide a Xinchuang‑compatible stack for secure, innovative, and ecosystem‑aligned development.

4. Insight: Low‑Code Boosts Banking R&D Efficiency

The report presents a case study of a joint‑stock bank that, with Puyuan Information’s low‑code platform, built a highly extensible “empowering branch” solution. By modeling business processes as modular PBC “blocks”, the bank rapidly assembled new scenarios.

Results:

South‑China branch created >100 specialty apps, cutting work time by dozens of times.

North‑China branch launched 42 apps, saving ~800 person‑months annually and reducing response cycles from months to days.

North‑West branch delivered 236 functional modules with 2 staff in 4 months, lowering development cost by 85%.

5. Outlook

As AI large‑models continue to evolve, "AI + low‑code" will become indispensable for banks’ digital transformation. Domestic leaders such as Puyuan Information are expected to drive continuous platform upgrades, supporting banks in reshaping IT architecture, accelerating innovation, and solidifying a robust digital foundation.