Software Testing, Automation, and AI – Are New Tools Really the Answer? Link to heading
In recent years, we’ve seen a surge of new tools and theories. Many of them talk about test automation, while some even suggest that software engineers — whether developers, system designers, or testers — could soon be replaced by AI. But that raises an important question: are all these new tools really the most important thing in software testing?
Automation Is a Tool, Not the Goal Link to heading
From my perspective, automation is incredibly useful, but only when used wisely. Well-designed automated tests at various levels can help us catch issues faster and reduce mental workload. On the other hand, poor automation can create a false sense of security and end up harming both the project and the people using the system.
A lot of the so-called “magic solutions” being promoted today remind me of a quote from Plato’s Apology:
“Each of them, because they were skilled in one area, believed they knew everything — but when questioned, it turned out they didn’t know what they thought they did.”
That feels very relevant to today’s specialists, who sometimes apply knowledge from one area to another without fully understanding the new context. I’ve caught myself doing that too. And context is key — understanding it takes time, experience, and thought.
Automation Isn’t Everything Link to heading
I’m not going to argue over definitions. We know that test automation usually means a set of scripts checking whether something moves from state X to state Y. But that’s just one part of testing — and often not the most important one. Compared to hands-on exploration by an experienced tester, automation can feel very narrow, especially in real-world, user-facing systems.
The effectiveness of automation depends heavily on the company’s culture and how software is designed. If end-to-end tests are created separately from the development process and based only on user scenarios, things start to break down quickly in complex systems. Preparing test users, repeating unnecessary steps — these efforts often bring little real value. And even then, tests still tend to fail because of unstable environments or external integrations.
What Actually Makes Testing Work Link to heading
Instead of obsessing over tools, we should focus on how controllable our system is. That means asking some important questions before we even start building:
- How can we easily get the app into the right state for testing?
- Can an internal API help us with that?
- What can we safely test in isolation?
- What can we simulate or skip to reduce risk and speed things up?
- How do we manage dependencies on external services?
Thinking this way lets us design systems that support both automated and exploratory testing — efficiently and without unnecessary pain.
Where AI Fits In Link to heading
AI tools can be a real boost when it comes to writing automated tests, especially if your framework is solid and your testing approach is consistent. In those cases, LLM based tools like Copilot, Cursor, or MCP servers for popular testing frameworks like Playwright’s MCP can help cut down the time needed to write and set up new tests.
But the more thought you put into making your app testable, the more you’ll get out of AI tools. That includes clear components, stable environments, and solid testing foundations. Still, we shouldn’t blindly trust the results. If your test doesn’t fail when it should, something’s wrong — and that needs attention.
In the End… Link to heading
Software is built for people. Tools can help us, and it’s great to gain experience with them. We live in a time full of exciting opportunities, but we also tend to chase shortcuts. And as the old saying goes:
“Take too many shortcuts, and you may never reach your destination.”
If we take quality seriously — using the definition from Weinberg, Bolton, and Bach: “quality is value to someone who matters” — then people will always play a key role in testing. In fact, with today’s pace of development and the temptation to take shortcuts, thoughtful testers are needed now more than ever. Because if we let robots take over our thinking about quality, we may soon find ourselves in a world where robots build for robots.