Like many other financial institutions, DevOps was a key initiative at this leading financial services organization and the development teams were under pressure to accelerate software delivery. They had hundreds of Java microservices, many containing legacy code that had little code coverage. Evidence indicated that changes made to the services with lower code coverage correlated to significantly higher defect rates.
Developer productivity was another important initiative. Poor coverage was creating an alarming amount of late-cycle rework. Furthermore, developers were spending too much time trying to manually create unit tests to increase coverage. These challenges were creating unpredictability in both quality and delivery timelines. Software delivery leaders were looking for a way to help the developers work more efficiently to increase productivity without sacrificing quality.
The organization determined that they needed to find a solution that would:
- Measure existing code coverage levels for both automated and manual testing efforts.
- Automate the generation of unit tests to maximize coverage.
- Support the following tools and technologies.
- JUnit testing frameworks 4 and 5
- IntelliJ and Eclipse IDEs
- Maven and Gradle build systems
- Jenkins and GitHub CI systems
- Spring, Spring Boot, and Lombok
The company moved forward with a proof of concept with Jtest where Parasoft provided
- Excellent guidance and support for implementing the Java testing solution to run unit tests
- Visibility and actionable metrics from an automated collection of existing code coverage levels
- One-click generation of unit tests that increased coverage by three times in just a few minutes
Using AI to automate unit test generation, Jtest offered some clearly differentiated capabilities that made it easy to build the business case.
- Unit Test Assistant. Parasoft Jtest’s AI-enabled Unit Test Assistant is an IDE plugin that helps guide users through the unit testing practice with easy one-click actions for creating, scaling, and maintaining unit tests.
- Test impact analysis. Jtest leverages AI to selectively target only changed code and identify the right subset of tests to validate those changes.
- Spring and Spring Boot support. Specialized support for these frameworks reduces a huge amount of effort for developers for MVC tests, mocking, and more.
- Aggregated coverage. Developers can view code coverage markers within the code for combined code coverage based on multiple testing practices, like unit testing, manual testing, and functional testing. The holistic view focuses testing efforts, which increases efficiency.
- Static application security testing (SAST). A huge bonus with Jtest is that it can also perform automated static analysis testing for code with support for CWE, OWASP, and more. While the company wasn’t looking for SAST, they realized Parasoft offered a solution that met their need to test for security vulnerabilities.
After implementing Jtest, the development team at this financial organization was able to generate comprehensive test suites in a matter of hours. Within a few weeks, they reached their goal of 85% code coverage on modified code for their most business-critical microservices.
Using Parasoft’s AI-enabled Jtest, the team was able to create better unit tests and drive higher code coverage.
“Since we implemented Parasoft Jtest, we have successfully reduced the amount of time it takes to create and maintain unit tests by more than 50%.”
—Director of development at the financial services organization
The solution delivered positive business outcomes for their strategic initiatives.
- Increased developer productivity. The Java developer productivity gain was estimated at 4-10 hours a week of time per developer that could be shifted from unit testing to innovation. That adds up to more than 20,000 hours annually for their team of 100 developers.
- Improved code quality. Defect rates dropped by more than 20%, resulting in:
- Accelerated delivery
- Reduced rework
- Improved confidence in software deliveries
The financial firm was able to create better unit tests and drive higher code coverage with AI-powered unit testing. They achieved significant results that enabled them to rapidly deliver high-quality software.
- 100% faster unit test creation. One-click generation of unit tests for legacy code and services improved developer productivity. They built a comprehensive and meaningful suite of JUnit test cases two times faster.
- 85% code coverage in a few weeks. The company increased code coverage from 20% to 85% coverage on average for their most business-critical microservices.
- 90% reduction in regression test execution time in their CI/CD pipeline. Jtest leverages AI to detect code changes and only executes the tests that are impacted by change.
In addition to increasing the development team’s productivity, this financial organization reduced the overall cost of testing. The ROI of implementing the tool was achieved in less than three months. They delivered quality code faster with fewer defects at a lower cost.
Take the Next Step
Download our whitepaper to learn how your team can benefit from AI and ML in software testing. Find real examples of how it’s being used to increase unit testing coverage and remove redundant work.