Agile Transformation of a Traditional Bank Software Development Team: A Case Study

The Agricultural Bank's Software Development Center Department 2 began an agile transformation at the end of 2017, selecting the "Spring Project" as a pilot to develop an agile development model suitable for a traditional bank.

As an external agile coach, I participated in the entire transformation, witnessing the team evolve from having no knowledge of agile to becoming self‑organizing and self‑improving. Kanban stand‑up meetings served as a crucial entry point for the team's agile transition.

The transformation progressed through three main stages:

1. Using Kanban stand‑ups to expose problems and establishing a continuous‑improvement foundation with Scrum. This stage focused on visualizing work and, through strict daily stand‑ups, built the basis for ongoing improvement. The board integrated requirement analysis, prototype design, story splitting, and the progress of three product groups into a single visual board for the Spring Project. By rigorously applying Scrum’s four events, especially retrospectives, the team quickly entered a continuous‑improvement rhythm.

2. Changing the testing model by introducing Git Pull Request + CI/CD workflow to accelerate development and testing iterations. Although the pilot project still had to pass the bank’s traditional testing‑center process before production, the team sought more frequent deliveries and feedback. They enabled Pull Requests in TFS and linked them to a CI/CD pipeline, providing faster, reliable, and traceable environment provisioning. These TFS capabilities supported organizing development and testing around user stories.

3. Connecting the full continuous‑delivery chain. After adopting Git Pull Request + CI/CD, the team linked story/task cards on the Kanban board with Git branches (Pull Requests), achieving end‑to‑end traceability from requirement to deployment. This integration prompted the team to consider using TFS work items/e‑boards for more direct tracking of the continuous‑delivery process.

The entire process took roughly six months, resulting in a progressively mature team. The transformation illustrates how engineering tools can drive changes in work patterns, and how management support—providing error‑tolerant policies, dedicated Scrum Masters, external coaches, and continuous tracking—facilitates success.

I offered the team recommendations that are often hard for traditional bank IT to accept, such as abandoning long‑standing configuration‑management processes, switching to distributed Git, adopting Pull Request‑based workflows, and even splitting repositories. The team carefully validated and iterated on these practices before fully embracing them.

We hope the team continues to grow and discover even better agile practices.

Related Articles

Agricultural Bank's Agile Transformation Practice Based on TFS

How to Play Agile: Scrum Management Practice Based on TFS Electronic Kanban

Agile and Waterfall Must Coexist Peacefully