QA Automation · Fintech · Business Banking
UI Automation Architecture
for Business Banking
Architecting a scalable, low-maintenance UI automation framework to replace brittle manual regression testing and accelerate release cycles for a high-stakes banking application.
✓ auth.login.spec 12ms
✓ payments.checkout.spec 8ms
✓ account.transfer.spec 15ms
✓ defect.regression.spec 6ms
✓ api.users.spec 4ms
The Challenge
Manual regression was a production risk.
In the business banking sector, UI regressions don't just cause poor user experiences — they result in financial discrepancies and broken trust. As the product scaled, manual regression cycles became a severe bottleneck. The existing testing approach was causing environment bloat and slowing down CI/CD pipelines due to inefficient test data management and inconsistent execution environments.
The Solution & Architecture
A modular framework built to survive scale.
Optimised Test Data Generation
A major bottleneck in UI automation is the time spent creating test states. I re-architected the automation script logic to prevent redundant data generation. Instead of generating new mock financial entities for every endpoint call and test run, I restricted data generation to two strictly defined scenarios — cutting environment bloat and drastically reducing overall suite execution time.
Hardened Execution Policies
To ensure zero flakiness across different environments, I standardised the script execution policies and Cypress configurations at the project-folder level. This guaranteed that tests behaved identically whether run locally by a developer or triggered remotely in the pipeline.
Integrated Defect Management
Automation is only as good as the bug resolution it drives. I tightly integrated testing outcomes with GitLab, successfully managing over 200 defects. By providing developers with reproducible Cypress test steps and clear error logs, we maintained an exceptionally high bug resolution rate.
The Impact
Quality shipped faster, with confidence.
Accelerated CI/CD
Refined data-generation logic and optimised Cypress configurations resulted in highly reliable, fast-executing pipelines.
Increased Coverage & Confidence
Replaced hundreds of hours of manual regression testing with deterministic, automated UI checks — zero critical regressions in production.
Developer Collaboration
Bridged the gap between QA and development through transparent GitLab defect tracking and reproducible automated scenarios.
Technologies