Homeinsightsartificial intelligence: real solutions for automated testing
quality assurance & testing

artificial intelligence: real solutions for automated testing

Helen Bally, Head of Quality Assurance & Testing · November 01, 2024

Software engineering leaders are constantly challenged to increase the productivity of their teams in response to market needs. They are expected to build software applications more efficiently, while maintaining quality.

AI-powered testing tools are becoming a vital resource for achieving these goals. There are a dizzying array of options and it’s important to analyse where your organisation can best profit. A recent Gartner[1] paper highlights the values of these tools for assisting in generating test scenarios, automating tests, optimising and prioritising test suites, predicting defects, analysing results, and estimating testing efforts. They reduce the need for manual intervention, enhancing test coverage, accuracy, and robustness.

In a survey conducted by Censuswide[2] this year, most organisations reported already using some form of AI-enhanced testing tools and they could also foresee increased investment (74%) in in their organization in the next year. It’s important to note the emphasis on “augmented” with AI-augmented testing: despite all the hype about autonomous testing, no-one believes that human input to the testing and validation process is going to go away any time soon. Understanding the use cases is important as you research AI testing tools and investigate capabilities. As with any technology, it’s important to think about where you will get the most benefit. Not all organisations will need all capabilities, and not all tools will offer them.

Gartner highlights five use cases where AI is improving the test process today:

1. Test Planning and Prioritisation

Testing is inherently a risk-based activity, and testers can never test everything exhaustively, especially with time constraints. AI helps minimise risk by optimising test sets, improving test coverage, and prioritising the most critical tests by looking at historical code changes, bug fixes, and other available resources. These tools can intelligently select regression tests for a release or remove redundant test cases, optimising the overall execution sequence.

2. Test Creation and Maintenance

AI is particularly effective in automating test creation and maintenance, making this one of the most innovative areas of AI-augmentation. Generative AI can generate test scenarios, tests, or specific test steps by analysing requirements or user stories, then refine them interactively with quality engineers. AI test tools can be used to automatically generate scripts for test automation tools analysing existing resources, such as manual test cases described in excel or word. While this can provide big leaps in efficiency, the benefits are inevitably going to be tied to how good your existing documentation is. Additionally, AI helps with “test self-healing”, where it can automatically update test scripts to reflect changes in the application under test, such as updates to user interfaces (UI) or application programming interfaces (API). Again, this reduces the amount of manual rework required to maintain test cases, increasing test efficiency and reliability.

3. Test Data Generation

Generating synthetic test data is another key use case where AI is being applied. AI can create large volumes of production-like data for use in test environments. Organisations are able to perform effective testing without exposing real, sensitive data. AI-powered test data generation tools typically use models trained on log files or past test data to simulate realistic testing scenarios.

4. Visual Testing

AI enhances visual testing, which is essential for ensuring that applications render correctly across various devices, browsers, and operating systems. Unlike traditional tools that require specific checks for each UI element, AI-powered tools can replicate human-like image recognition to detect any layout or content issues without predefined rules. This makes visual testing more efficient and scalable, particularly for complex user interfaces. AI also helps ensure compliance with accessibility standards like WCAG.

5. Test and Defect Analysis

AI is highly valuable for analysing test results – the tools can differentiate between flaky tests, real bugs, and false positives, helping testers focus on the most critical defects. AI is also increasingly applied in static code analysis and security testing, where it can predict defects and assist in identifying which tests should be prioritized based on code changes. By learning from previous test runs, AI can improve future tests, making testing smarter and more targeted. These capabilities help testers focus their efforts on areas that pose the highest risk to quality, leading to better overall software performance.

summary

In summary, AI-powered testing tools help streamline the software testing lifecycle by optimising test planning, automating test creation, improving test data generation, augmenting visual testing, and enhancing defect analysis. These innovations will reduce human intervention and increase the overall efficiency and reliability of testing. Organisations need to review where they can benefit most from the different capabilities available.

[1] Gartner: Market Guide for AI-Augmented Software-Testing Tools – 13 February 2024 – ID G00783848 – By Joachim Herschmann, Thomas Murphy, and 3 more
[2] AI and Software Quality: Trends and Executive Insights – 2024 – Survey conducted by Censuswide

More insights

keeping up with the times through innovation and creativity


digital solutionsgovernance & service managementquality assurance & testing

November 01, 2024

meet us at the swiss it forum(s) 2024


digital solutionsgovernance & service managementquality assurance & testing

August 26, 2024

challenges and benefits of implementing a metrology service centre


digital solutionsquality assurance & testing

May 23, 2024

Contact us