AI Test Management, AI Prompts for Playwright, Codex and More TGNS158

Have you seen the latest AI powered test management tool? What AI prompts can you use for supercharging your Playwright test? And why would you use the new Tap feature in WebDriver IO? Find out on this episode of the Test Skill News show for the week of May 18th. So grab your favorite cup of coffee or tea and let's do this. All right, I'm always on the lookout for new tools. Here's the first one that caught my attention. It's called Quality. It's a test management platform that has just introduced an AI powered agent named Hudi aimed at accelerating your test case creation and boosting test coverage. And this tool claims to help teams reach over 80% coverage while improving collaboration and delivering real time insights. Quality integrates with widely used tools in your software development lifecycle including Jira and supports custom reports, rest APIs and single sign on. And also the platform promises unlimited tests and defect tracking, AI credits and 247 support along with data migration from other tools. Product is pitched as a flexible solution that adopts to existing workflows rather than requiring teams to change the whole process to adapt the tool. And according to the platform, users can start a 14 day free trial without entering any credit card information. So rather than having to adopt to a vendor's workflow, this one actually plugs in place into your existing workflows which can get you up and running really quickly. And I think a lot of times test management is sorely missed on a lot of projects and I always say seeing is believing for yourself. So give Quality's AI agent Hudi a try and see if it really does reduce your time spent on manual test creation while easily integrating with your existing test workflows. You can try it for yourself using the special link down below. Do you want to learn more about AI prompt engineering for Playwright? If so, you're in for a treat because I found this next article and this is a new blog post by staff Quality engineer Gary Parker that outlines practical strategies for using AI assistants to enhance your your test automation with Playwright in this article published on Medium focuses on crafting effective prompts to improve output quality from AI tools, particularly in generating and debugging test code, and Gary provides some real world examples showing how testers can prompt AI to do a bunch of things. Some examples he gives is generating test stubs tailored to specific scenarios, creating page object models from application screenshots or workflows, how to set up test fixtures aligned with project structure, troubleshooting failing tests more quickly by narrowing AI's focus and a bunch more. This post really emphasizes that prompt clarity and specificity are key to leveraging AI effectively. And rather than generic request, Gary recommends structuring inputs that mirror a tester's intent, such as including test goals, context about the app under test, and preferred coding conventions. Do you want to know a way to help make your tests more reliable for mobile automation across different platforms? So WebDriver IE has just launched a new tap command designed to address long standing reliability issues with QLIK and mobile testing. And WIM mentions that the click function, although standard, has been known to behave inconsistently across Android and iOS, particularly when dealing with partially visible or off screen elements, and these inconsistencies often lead to silent failures, erratic behaviors and unreliable test outcomes. So how do you get around it? Well, WebDriver IE has just introduced Tap, which resolves these issues by mimicking actual finger taps on a mobile screen. And unlike Click, which simulates interaction in a more abstract way, Tap executes a real coordinate based tap on the center of a visible element. Wim points to a video to learn more, but he also points to documentation that goes over some key features of the new TAP functionality. And I'm not sure why more people don't use webdriver IO. So this is definitely you should definitely check out and hopefully let me know if it helps reduce flakiness and your end to end tasks across devices. Do you use Cypress? Well, here's an article that caught my attention and this is a new peer reviewed research paper titled Cypress Copilot Development of an AI Assistant for Boosting Productivity and Transforming Web Application Testing and one of its authors, Suresh, just told me. It's now been published on IEEE and this study presents a really novel approach to generating behavior driven development using Cypress along with specialized prompt engineering techniques. They go into detail on how the researchers developed Cypress Copilot, an AI assistant designed to assist developers and testers by converting natural language requirements into executable test scenarios. The key contributions of the paper is a custom prompt engineering method that enhances the accuracy and contextual alignment of the AI generated test scripts and according to the study, the proposed technique outperforms standard prompt methods, showing improved results and relevance and and execution success rate. The assistant supports iterative refinements and real time feedback for generating test code, aiming to reduce the manual efforts typically associated with test authoring. The paper also includes detailed experimental results comparing different prompt strategies and highlighting the limitations of the off the shelf approaches. Definitely a really cool read and you can find out more about it using links down below. All right, this next story is going to stop probably blowing up. It just was announced two days ago, and that's how OpenAI has just launched Codex, which is a cloud based AI software engineering agent designed to automate various coding tasks. It's already available in ChatGPT Pro, Team and enterprise users. Codex can write features, fix bugs, run test, and propose pull requests all within isolated cloud environments reloaded with the user's code base. And the article goes into detail how this is powered by the Codex 1 model, a variant of OpenAI's O3 optimized for software engineering. Codex was trained using reinforcement learning on real world coding tasks, and it aims to generate code that aligns with human coding styles and project requirements. And users can now interact with codecs through ChatGPT's interface by assigning tasks or asking questions about their code base. Each task is processed independently, with codecs capable of reading and editing files, running commands, and providing verifiable evidence of its actions through terminal logs and test outputs. And I'm sure we'll hear more about this as the weeks and months go on, so stay tuned. Do you do performance testing? Well, you're probably missing out on this forgotten metric. What is it? So In a recent LinkedIn blog post, Andre, a senior Software performance analysis, emphasized the importance of using Linux pressure stall information, otherwise known as psi, as a performance diagnostic tool that is often overlooked in software troubleshooting. It also highlights that traditional metrics such as CPU memory and I O utilization over only measure how much of a resource is being used, not how the system is currently coping with its current load. PSI fills the gap by showing whether the system is under pressure, stalling or waiting for resources, which provide a clearer picture of actual system performance. And the article goes into detail on how PSI tracks stall behavior and CPU memory in I O subsystems and indicates when tasks are delayed due to insufficient resources. And this can or hopefully expose hidden performance bottlenecks that classical utilization metrics fail to reveal. Also, as I was scrolling through LinkedIn, I found another article titled the Silent Signals the Hidden Threat of Minor Incidences, and this is by Hirinda, head of Platform Resilience, and he emphasizes this often overlooked impact of minor system inefficiencies and instances on software performance and reliability that's often overlooked. So Horrenda illustrates how seemingly trivial issues such as redundant validation logic, unoptimized front end assets, excessive logging, bloated css, and insufficient background job polling can collectively degrade user experience, strain resources, and also increase operational costs. And these minor insufficiencies while not immediately catastrophic, can accelerate and lead to significant system degradation over time. So this article explores the cultural and structural reasons why engineering teams may overlook minor incidences. And if you're actually not sure how to approach performance investigations, I found another article that takes a deep dive into a real world approach, and this is by Rico, who outlines his structured methodology for diagnosing and resolving performance issues in software systems. Rico draws from a decade of engineering experience and he breaks down the process down into three key phases problem definition, measurement, and root cause identification. Ricoh also stresses that vague or poorly framed problem statements often waste time and misdirect efforts. Rico also cautions against over fixating on CPU usage, knowing that's often not the limiting factor in real world performance bottlenecks. Want to thank this week's sponsor Zaptest AI, an AI driven platform that can help you supercharge your automation efforts. It's really cool because their intelligent copilot generates optimized code snippets while their plan studio can help you effortlessly streamline your test case management. And what's even better is you can experience the power of AI in action with their risk free 6 month proof of concept featuring a dedicated ZAP expert at no upfront costs. Unlock unparalleled efficiency and ROI in your testing process. Don't wait. Schedule a demo now and see how it can help you improve your test automation efforts using the link to down below. All right, for links of everything of value we covered in this news episode, head on over to all the links in the comments down below. Once again, I'm Joe and my mission is to help you succeed in creating end to end full stack pipeline automation awesomeness. As always, test everything and keep the good. Cheers.