How DevOps Reshaped Agricultural Bank’s R&D Management: Key Lessons

01 Evolution of the Bank’s R&D Management System

Over the years the Agricultural Bank’s R&D Center has built an enterprise‑level management framework by adopting external standards such as CMMI and agile, creating a clear division of labor, process‑driven workflow, comprehensive tooling, and a dual‑mode operation model.

Enterprise‑level R&D management established (2008). The bank passed the CMMI Level 3 assessment and created a lifecycle covering planning, requirements, design, development, testing, deployment, trial operation, and closure.

Exploration of dual‑mode operation (2018). Agile project management was incorporated, forming a “steady‑state + agile” model with a unified agile process, two capability models, three practice sets, and four action guides.

DevOps practice rollout (2020). Tool‑chain automation and integration dramatically increased, building an end‑to‑end continuous delivery pipeline and a measurement framework.

02 How DevOps Changes R&D Management

DevOps follows lean thinking, standardizes development and technical management, and emphasizes tool empowerment to achieve tighter integration, automation, standardized operations, faster delivery, and higher collaboration efficiency.

DevOps promotes integration.

It seeks to eliminate fragmented tools, processes, and data silos, achieving tool, process, and data continuity throughout development.

DevOps stresses automation.

All code, scripts, test cases, operations, and information flow are treated as configuration items and automated to reduce manual intervention.

DevOps strengthens quality control.

Quality gates such as unit testing, code compliance checks, and vulnerability scans are embedded in the pipeline.

DevOps requires data visibility.

A measurement system visualizes efficiency, delivery quality, and tool capability, enabling problem identification and continuous improvement.

Overall, scaling DevOps raises the bar for R&D management.

Room for Lean Integration

Agile complemented the original framework, while DevOps adds tool‑centric integration, leaving space for further refinement.

Room for Engineering Process

Stability and usability of the continuous delivery toolchain, cross‑platform visibility, and deeper test automation still need improvement.

Room for Business Value

Accelerating high‑quality continuous delivery of business value remains a goal.

Room for Cultural Development

Beyond technology, DevOps requires a cultural shift, systematic planning, and a dedicated coach team.

03 Reflections on R&D Management Improvement

Management improvement is perpetual; start with easy, detailed actions to tackle big challenges.

A Comprehensive Supporting System

Given the bank’s need for stability in back‑office and innovation in front‑office, the “steady‑state + agile” modes will coexist, with DevOps and TMMi integrated to support a unified, multi‑scenario R&D system.

Integrate standards for a unified R&D system. Combine industry‑wide engineering activities with DevOps maturity models to create clear standards, unified tools, and a holistic enterprise‑level framework, also incorporating security and risk management.

Integrate scenarios to support diverse development modes. The unified process accommodates waterfall, agile, and DevOps, enabling differentiated metrics and collaborative improvements.

Integrate quality for soft landing of management. Embedding quality checks from unit testing to production monitoring automates detection and interception, simplifying governance.

Integrate data to build a one‑stop R&D platform. Platforms such as Alibaba Cloud Effect and Baidu Efficiency Cloud provide workflow orchestration, data visualization, and metric dashboards; a dedicated DevOps platform can further enhance integration and service capability.

Extending Value to Both Ends

From a technology perspective, improving quality and efficiency accelerates high‑quality business delivery; from a bank‑wide view, it translates into faster business value realization.

Establish full‑function teams. Break down silos between business and technical groups, adopt user‑story mapping, deliver MVPs in small batches, and enable rapid product delivery.

Benchmark to boost technical operations. Embed measurable feedback loops in product design to continuously enhance business value.

Building a Coaching Culture

Develop a DevOps coach ecosystem at both organizational and team levels, leveraging external expertise to deepen delivery, agile, operations, micro‑services, and security competencies.

Foster an internal DevOps community through external exchanges and knowledge sharing.

In June 2020, the GNSEC 2020 Global Next‑Generation Software Engineering Online Summit announced the seventh batch of DevOps standard continuous delivery assessment results.

Five Agricultural Bank projects passed the Level‑3 assessment of the “DevOps Capability Maturity Model” for continuous delivery, a leading domestic benchmark.

Credit Middle‑Office Project

Personal Online Banking Project

Distributed Application Interconnection Platform (AIR) Project

Value‑Added Tax Input Tax Management Project

Financial Mini‑Store Project

Achieving Level‑3 in continuous delivery signifies that these projects have reached a leading level of capability in China.