Event Insight

ExpoQA 2026 - Our Madrid Takeaways

Last week, the Refleqt Spain team made our way to Madrid for ExpoQA 2026, one of the biggest events in the QA world. Beyond the sessions, it was a fantastic chance to spend time together as a team. We strengthened some bonds, had a lot of laughs, and made the most of the city (even if the heat was something else entirely!).

Expo2026 Madrid - Refleqt Team

Refleqt

June 4, 2026 • 4 min leestijd

Deel deze insight

AI-driven test tools in the spotlight

AI-driven testing tools were one of the main themes at ExpoQA, with frequent references to Playwright (and its powerful MCP), Cypress, Browser Use, Cursor, GitHub Copilot, and emerging autonomous QA agents such as LangGraph. Below is an overview of the talks that stood out to us the most.

Testing AI Systems: A New QA Paradigm

This one was a bit of a wake-up call. The core message was testing AI isn't like testing traditional software. You can't just check if the answer is "right or wrong" anymore. AI outputs can vary, so testers need to think more about whether responses fall within an acceptable range rather than matching a fixed expected value.

On top of that, there's a whole new checklist to worry about: Is the test data up-to-date? Is the model actually suited for the job? Is it producing biased or harmful outputs? The skill set testers need is expanding fast: think prompt engineering, ethics awareness, and a lot more collaboration across teams. Exciting, but also a clear signal that the QA role is evolving quickly.

Asimov's Zeroth Law of Robotics: Observability for AI

One of the standout sessions of the whole conference. The big takeaway here was simple but powerful: if you can't see what your AI is doing in real life, you can't really trust it.

Since AI systems are non-deterministic, thus they don't behave the same way twice, running tests before launch just isn't enough anymore. You need to keep an eye on things once they're live. Watching out for issues like the model drifting over time, unexpected behaviour, or the AI being manipulated through clever prompts. The session walked through a practical setup using open-source tools to do exactly that. This one was one of the most directly useful things we brought back.

The Multi-Agent Revolution That Is Redefining Testing

This one was genuinely exciting. The session took us on a journey from the early chatbot days of 2022 to where we are now: fully automated pipelines where multiple AI agents work together, each with a specific role: planning, building, implementing, and even judging the quality of the output.

One of those agents acts as an autonomous quality gate, rejecting code that doesn't meet the bar before it even gets committed. Testing becomes baked into the process rather than bolted on at the end. A live demo brought it all to life, and it was honestly one of the most forward-looking things we've seen in a while.

The through line across all of it, believe it or not, was the same: AI is here, it's changing how we build and test software, and the conversations around it are finally getting practical. We came back from Madrid inspired, a little sun-kissed, and with plenty to put into action.

Keep on learning, keep on testing!

Software quality never stands still, and neither do we.

That is why we attend conferences such as ExpoQA, but the real work only starts afterwards, when we apply those insights with our clients.  

Interested in sparring about AI in testing, test automation or observability? Get in touch! We are happy to think along with you.

Deel deze insight

milan

Milan Meuleman

Business development & sales

Contact Refleqt today

Would you like more control over software quality, test automation, or performance? We are happy to explore together how we can support your team with an approach that works in practice.

Plan a meeting Plan a meeting Fill in contact Fill in contact