Automated Testing for Insurance Estimation Software of a Worldwide Insurance Company

Summary

A worldwide insurance company was building a mission-critical, microservices-based insurance estimation system integrated with many external solutions, and needed it to run flawlessly while still shipping fast. The Client handed the entire testing and QA process to INNERLUXES for the two-year development project. Combining automated and manual testing on Protractor and building continuous testing into the delivery pipeline, INNERLUXES reached around 95% test coverage and cut test design costs by 20% — and pre-release acceptance testing surfaced only two bugs.

About the Client

The Client is a global provider of insurance solutions and insurance-related services with worldwide interests. It covers casualty and accident insurance and delivers related services to businesses in areas such as engineering, logistics and inventory, aviation, and energy.

A Mission-Critical Estimation System Needed Flawless, Fast Releases

The Client's core insurance estimation software was under development as a mission-critical, microservices-based system integrated with multiple external solutions, and it had to run smoothly, without errors or disruptions. The Client needed to accelerate time-to-value while preserving high quality, and chose to hand the complete testing and QA process to INNERLUXES for the duration of the two-year development project.

Combined Automated and Manual Testing Built Into a Continuous Pipeline

The project ran on Scrum, with testing in parallel with development. To meet the need for fast, high-quality releases, the INNERLUXES test automation engineers used a combined approach that integrates automated and manual testing: most testing was automated on a Protractor-based environment, while manual tests gave developers prompt feedback after hotfixes. To cover the application thoroughly, the team performed:

• Functional testing of new features and regressions, covering the functional requirements with more than 750 test cases.
• Integration testing of the microservices APIs and integrations with external systems.
• Localization testing to verify worldwide operation against legislation, currency variations, and more.

Tests ran at both the UI and API levels; testing at the API level shifted verification earlier in the cycle and suited the fast Scrum releases. Working with the DevOps team, INNERLUXES designed and implemented a continuous integration pipeline and built continuous testing into it, so new software could be verified quickly without compromising quality. To test efficiently against large datasets, the team used the pairwise approach to test design — which, applied early, saves up to 20% of test design costs while raising coverage by 20-30%.

Despite the complexity of the insurance calculations, the team achieved 90-95% test coverage. After two years of development, pre-release user acceptance testing revealed only two bugs.

Near-95% Coverage, Lower Test-Design Costs, and Only Two Bugs at Acceptance

• Despite tight time frames, the Client obtained a smoothly operating, thoroughly tested solution.
• Test coverage reached 90-95%, with test design costs cut by around 20% through early pairwise test design.
• Pre-release user acceptance testing surfaced only two bugs after two years of development.
• The Client also received a comprehensive package of reusable automated tests for future regression testing of modifications and updates.

Tools

Protractor, TypeScript, JavaScript.