Experience Report: My aqua KI-Copilot Journey –Surprises, Learnings and Practical Insights

Can AI really support software testers? QA consultant Gillian Trombke took the new version of aqua AI Copilot for a spin — and in his experience report, he shares what surprised him, what impressed him, and where AI truly makes a difference. Read his full review in our new article!

As I had the opportunity to try out the updated version of aqua's AI Copilot as part of the official andagon course I was excited. I was keen to see how much the software had improved since my previous experience of it. In this blog post, I share my key findings, including the positive surprises and practical benefits I discovered, as well as areas where the AI could still be improved upon. Can it really save time? Which features are worth highlighting? Why do I recommend it, even though I can't use it in my current project? Read on to find out!

Why did I choose the “aqua AI-Copilot” course?

As a consultant at andagon, I regularly work on test projects where aqua is a core tool. I know the platform well and have built up solid experience with it over the years—especially in the area of test automation. So, when I heard this course was up for review, I was immediately curious. I had actually tested the AI-powered assistant about two years ago during a project but ended up setting it aside. Back then, it simply wasn’t mature enough to meet our complex and specific requirements.

What had changed since then? That’s exactly what I wanted to find out. The timing of the course was perfect—it gave me a chance to see how the Copilot had evolved and to really put the new version to the test.

Between “Aha” Moments and Eye-Rolling – What the Course Can Really Do

At first, I was a bit skeptical about how the course was structured. The section on writing a “good prompt” only came after the practical exercises. I initially thought: why not cover this earlier? But in hindsight, it made sense. Trying out the AI Copilot without guidance first made it clear how crucial prompt quality is. That hands-on experience really reinforced the learning — especially the importance of phrasing queries properly to get useful results. And that’s coming from someone already familiar with prompt engineering.

What really surprised me this time was the huge time savings: Copilot was able to generate a large number of test cases, requirements, and even test data in no time — with impressively solid quality. That’s a real game changer, especially for early project phases or repetitive tasks. Instead of spending hours writing everything manually, I could produce complete sets of tests or requirements with just a few targeted prompts. It not only saves time but also frees up mental space for analysis and strategy.

I found it particularly helpful that the tool could generate not just functional test cases, but also positive and negative scenarios, BDD formats (Given-When-Then), and PRD structures. The automatic creation of test data, test steps, or entire requirement packages shows how far the tool has come. There are even new features like boundary value analysis and equivalence class partitioning — so new, in fact, they aren’t yet included in the course. That didn’t stop me from exploring them on my own, and I was impressed by what I saw.

But what was the issue two years ago, and how has Copilot improved since then? Back then, it wasn’t much help in our complex projects—it lacked contextual awareness and often “guessed in the dark.” Unfortunately, that’s still a limitation today. The content it generates is solid, but not precise enough for highly project-specific needs. While you can add context through detailed prompts, it’s time-consuming and not practical for complex or niche projects. That said, aqua has announced plans to integrate project-level context support. I’m curious to see how well that will work in practice.

Conclusion: A Worthwhile Journey – Even If I Can’t (Yet) Apply It Directly

I found the “aqua AI Copilot” course genuinely valuable — so much so that it’s a pity I can’t apply what I learned in my current project, simply because aqua isn’t being used there. That makes me all the more excited for my next aqua project, where I can put these new features into practice.

I’d recommend the course to anyone working with aqua — or anyone curious about how AI can support the testing process. It not only showcases what Copilot is already capable of but also offers plenty of inspiration for applying it in real-world scenarios. What makes it even more valuable is that user feedback directly influences the tool’s further development.

I’ll definitely share the course with former project colleagues and friends in software development. It’s a great way to get to know aqua better while contributing directly to its improvement. And let’s not forget the bonus: the certificate at the end, especially while the course is still free.

In short:

A well-structured, hands-on course that doesn’t just teach the aqua AI Copilot’s features—it gives you the chance to explore and test them directly. It’s especially worthwhile for anyone using aqua who wants to unlock more of AI’s potential in testing. The course is also clear proof of how far Copilot has come in just two years. Not just a marketing pitch—real progress.

My tip: Take advantage of the course while it’s still free!

You can simply register here.