Testing an application shortly before release is no longer sufficient to ensure the quality of modern software, according to an article published on the JetBrains blog by Kerry Beetge. The central idea is to distribute quality checks throughout the software development life cycle, so that defects emerge closer to the moment they are introduced rather than being discovered after additional changes and complexities have accumulated.
The article links this shift to increased reliance on AI-generated code. Citing the 2025 Stack Overflow survey, it states that 84% of developers use AI tools or plan to use them in the development process. JetBrains believes that the increased volume of code produced by these tools makes frequent, scalable reviews more important, while human verification remains essential because errors may be transferred into the automated output.
From the Pre-Release Gateway to Continuous Assurance
The article distinguishes software quality assurance (SQA), which tracks compliance with operational, reliability, security, and standards requirements throughout the development cycle, from traditional quality control (QC), which often focuses on the final product and handles defects reactively. The continuous approach makes it possible to detect problems while code is being written or integrated, when fixing them is less costly and more closely connected to the original context of the change.
The article notes that release cycles in modern development environments have shifted from months to days, while CI/CD pipelines allow code to move from commit to production with limited human intervention. Therefore, relying on a single tool is not enough; each testing layer reveals a different type of problem.
What Do the Tools Cover in Practice?
- Static analysis: Examines code without running it to detect defects, vulnerabilities, and violations of coding standards. Qodana is one example.
- Unit testing: Verifies that functions and components work individually, with examples including JUnit, Jest, PyTest, and NUnit.
- Integration testing: Examines the interaction of services, APIs, and data flows. Its tools include Postman and Soap UI.
- Functional and interface testing: Simulates user journeys through a browser using tools such as Playwright, Cypress, and Selenium.
- Performance testing: Measures behavior under load and helps reveal bottlenecks, memory leaks, and slow queries, using tools such as JMeter, LoadRunner, and k6.
- Security testing: Combines SAST, SCA, dependency scanning, DAST, and the detection of secrets or API keys accidentally included in the code.
How Are Tools Selected and Integrated into the Workflow?
The article proposes evaluating tools according to their integration with CI/CD, support for the languages and frameworks in use, ability to automate repetitive tasks, scalability as the codebase and teams grow, provision of actionable reports, and the tool’s own security practices.
At the implementation level, it recommends moving checks into the coding stage, automating repetitive tests, running them with every commit or build, and monitoring technical debt through indicators such as code complexity, duplication, and test coverage. It also stresses the need to include security scanning in the regular quality-assurance workflow rather than postponing it to a separate pre-release review.
certi.news’s Assessment
The actual change proposed by the article is not the addition of a new test, but the redistribution of quality responsibility throughout the development cycle. This approach benefits teams that work with frequent releases or rely on distributed architectures and external dependencies, but it does not eliminate manual testing or engineering judgment. Automation, particularly AI-based automation, may accelerate problem detection without guaranteeing complete coverage or correct results on its own. The source also provides general guidance and examples of tools, but it does not establish a quantitative framework for comparing them or present independent comparative performance results.