AI Insights

In the world of QA, technologies become outdated faster than you can blink. Tools and practices once considered gold standards can quickly turn into bottlenecks for teams and businesses. We spoke with Dzmitry Mikhailouski, Lead SDET, and Volha Khudzinskaya, Head of QM, about which technologies it’s time to leave behind, why clinging to outdated practices can be risky, and which tools still deserve respect.

When you think about old technologies in QA, which ones do you feel it’s time to move on from?

Dzmitry Mikhailouski: As the industry evolves, some practices start holding us back instead of helping. For example, relying on manual regression testing without automation is simply no longer sustainable. The same goes for isolated environments that don’t use containerisation – they slow down development and make testing less reliable.

Another big issue is sticking with legacy tools that no longer have proper support or testing without any data-driven insights. Approaches like “test after build,” rather than integrating testing into CI/CD pipelines, also feel outdated. And finally, focusing only on UI testing without covering APIs or contracts leaves too many blind spots in modern systems.

Volha Khudzinskaya: Monolithic legacy systems built on outdated frameworks with zero test automation. These are not just technologies, but entire ecosystems that hinder agility, are expensive and risky to maintain, and create a “knowledge base” dependency on a few specialists. Saying goodbye to them is often a necessity for business survival, not just a technical upgrade.

Are there any tools or technologies you hope never to work with again, and why?

Dzmitry Mikhailouski: I’d really prefer not to go back to tools based on the record-and-playback approach. On the surface, they seem convenient, especially for quick starts, but in practice, they create more problems than they solve.

The main issue is their fragility – even small UI changes can break tests, leading to constant maintenance. They also don’t scale well, because it’s hard to reuse or properly structure those tests. On top of that, they lack the flexibility of code-based frameworks, making customisation unnecessarily complicated.

In the end, they often give a false sense of automation: you get quick initial results, but they don’t hold up in real CI/CD pipelines or long-term projects. Modern testing needs solutions that are stable, scalable, and maintainable – not brittle scripts that slow teams down.

Volha Khudzinskaya: Manual test management via Excel/Word documents or extremely rigid, non-integrated legacy test management tools, aka TestRail. They create immense overhead for traceability, generate unreliable metrics, and are a constant source of human error. In the era of DevOps & AI, they break the feedback loop, making QA a bottleneck rather than an integrated quality enabler.

A tool you had to give up, but didn’t want to

Dzmitry Mikhailouski: Selenium IDE is a good example. It was very convenient for quickly creating tests and prototyping ideas, especially in the early stages.

However, as projects grow, their limitations become obvious. It doesn’t scale well for complex automation frameworks, and integrating it into modern CI/CD pipelines is far from ideal. The record-and-playback approach leads to brittle, high-maintenance tests.

So while it was a great entry point for many testers, today’s requirements for reusable, stable, and scalable automation make it difficult to justify using it in real projects.

Volha Khudzinskaya: Skype. It sounds simple, but it was the first true corporate messenger for my team and me. It created a sense of a unified digital space with its familiar ringing sound and simple interface. There’s definitely some nostalgia for that era.

Is there any “old” technology that still deserves respect?

Dzmitry Mikhailouski: Absolutely – JUnit and similar unit testing frameworks still deserve a lot of respect. They played a foundational role in shaping how we approach automated testing today. What makes them stand out is their simplicity and reliability. Even now, JUnit remains lightweight and effective for unit-level testing, and it integrates smoothly with modern tools and CI/CD pipelines.

It’s also a great learning tool. For many engineers, it’s one of the first ways they understand core testing principles. Even though newer tools offer more advanced features, the ideas behind JUnit – clarity, simplicity, and maintainability – are still just as relevant.

Volha Khudzinskaya: Definitely the old-school core principles of testing, development, and delivery. Things like “test early,” “automate where it makes sense,” “understand your user.” Tools will evolve, but this foundation remains unchanged. You simply can’t build quality software without it.

How risky is relying on what worked 10 years ago?

Dzmitry Mikhailouski: Relying on what worked 10 years ago is a serious risk, both for your career and for the quality of your work. The industry has changed dramatically – we now deal with microservices, APIs, and cloud-native applications, all of which require very different testing approaches.

Automation expectations have also evolved. Today, CI/CD and DevOps demand fast, scalable, and fully integrated testing processes. At the same time, security requirements are much stricter, and older tools often don’t meet modern standards.

Volha Khudzinskaya: It’s a major risk. In our world, “it used to work” means nothing. If code isn’t continuously tested and validated in the current environment, it’s not reliable – even if it worked six months ago. Relying on outdated results is like driving a car that hasn’t been serviced in years: past performance doesn’t guarantee future stability.

What risks do companies face if they refuse to modernise?

Dzmitry Mikhailouski: When companies stick to outdated tools, the consequences manifest at multiple levels. Financially, they end up spending more and more on maintaining legacy systems, while also missing market opportunities and risking costly downtime.

From a cultural perspective, it becomes harder to attract and retain strong specialists. Teams get frustrated working with inefficient processes, and over time, this leads to stagnation and resistance to change.

Technically, the risks are just as serious. Legacy systems often have security vulnerabilities, don’t integrate well with modern ecosystems like cloud platforms or CI/CD pipelines, and struggle with scalability and performance.

Volha Khudzinskaya: The companies that evolve with technology win. If you want top talent, you need to offer modern tools and environments. Otherwise, specialists will simply choose competitors. The biggest risk is falling behind – losing your people first, and your customers next. The financial consequences are obvious.

What technologies will we say goodbye to in the next 5 years?

Dzmitry Mikhailouski: The testing landscape is evolving very quickly, and some approaches are clearly on their way out. Record-and-playback tools are already becoming obsolete due to their fragility and high maintenance.

We’re also moving beyond UI-only testing – modern systems require strong API and contract testing. Standalone test management tools that don’t integrate with CI/CD are losing relevance as well.

At the same time, proprietary scripting languages are being replaced by more flexible, widely adopted solutions such as Python, Java, and JavaScript-based frameworks. Manual regression testing is gradually giving way to automation and AI-driven approaches, and on-premise testing infrastructure is being replaced by cloud-native, containerised environments.

Volha Khudzinskaya: In such a fast-changing world, making 5-year predictions is difficult. What’s certain is that QA is here to stay. Tools and methods will change, but critical thinking, user advocacy, and the responsibility for quality will always be in demand. Tools come and go – the mindset stays.