How CITIC Bank Built a Digital Testing Quality Control Platform for End-to-End Risk Management

1. Overview of the Quality‑Control Process

Traditional testing is a black‑box approach that relies heavily on individual tester expertise, lacks quantifiable risk metrics, and can hide delivery quality issues. CITIC Bank’s digital testing quality‑control “control cabin” decomposes complex financial transactions, providing multi‑dimensional, end‑to‑end visibility of single transactions, cross‑system flows, full business cycles, and customer journeys, while also mapping vertical data and performance risk links to enable visual, analyzable, and trustworthy data throughout the process.

Key Practices

Business rule design diagrams: Using 16 categories and over 2,000 accounting products, the bank extracts business enumeration values and creates simulated sequence diagrams, linking 70,000+ end‑to‑end transaction paths with standardized customers and products, achieving 100% coverage and a “high‑definition map” of the financial industry.

Data‑link visualization: Database tables are classified by business, establishing multi‑dimensional relationships between UI fields and data entities, enabling 24/7 unattended automated testing and data‑driven impact analysis.

Execution‑result pyramid: A four‑layer view (global + transaction + scenario + code) drills down test results for comprehensive quality assessment and release gating.

Performance‑risk network: Leveraging 78 performance data models, the platform provides precise, all‑round system risk evaluation and full‑coverage health checks.

2. Building the Digital Testing Quality‑Control “Control Cabin”

Based on a three‑dimensional business‑asset view, a standardized customer‑scenario map, five‑dimensional transparent performance factors, and unified data extraction, the bank constructed a standardized, digital testing quality‑control middle platform that shifts testing from manual reliance to full machine control.

Left‑shift risk prevention: Following TMMI Level 5 standards, intelligent analysis systems formulate realistic test plans.

Real‑time end‑to‑end monitoring: Automatic data point collection, AI‑driven deviation calculations, and timely supervision of test progress and risks.

Precise post‑execution review: A three‑state data loop (plan, execution, result) closes the quality‑control cycle.

3. Strengthening Digital Construction

The control cabin adopts a top‑down quality‑control strategy, mapping the entire test lifecycle, identifying core decision points, and building a data‑driven indicator system. Automated data collection, multi‑source integration, and indicator‑based monitoring enable transparent, quantitative quality management, AI‑assisted insight, and proactive risk alerts.

Data transparency: Consolidated multi‑source data creates a live, interconnected digital world.

Indicator system: Multi‑dimensional quality metrics improve control precision.

Automation of control: Data‑driven checks and automated gatekeeping enforce machine‑controlled releases.

Intelligent decision‑making: A “data + technology” engine supports problem insight, assisted decisions, and feedback loops for smart warnings.

Forum Details: The 2023 China International Service Trade Fair “Enterprise Digital Transformation Forum” was held in Beijing, featuring announcements of the 2023 “Digital Influence” case‑collection results, with 10 leading enterprises, 5 leading experts, 5 outstanding experts, and 58 innovative cases selected.

Key speaker: Gao Rui, master’s graduate of Jilin University, senior testing manager at CITIC Bank Software Development Center, holder of the Ministry of Industry and Information Technology’s senior communications engineer certification, and contributor to multiple national standards on DevOps and continuous testing.