Quality Sense Conf 2025

Over the years, I have watched software testing get rewritten. Not by testers. Not by teams. But by narratives crafted to sell tools, promise speed, and cut costs. From "manual testing is dead" to "AI will replace QA," we have been sold stories that sound smart but reduce the craft into checklists, dashboards, and demos.

To be honest, I have had enough.

This keynote is not a rant against innovation. It is a reality check on how narratives, clever, convincing, and commercially driven, have hijacked the way we think about testing. We did not evolve organically. We got pitched. First by automation platforms that promised speed. Then by low code tools that promised inclusivity. And now by AI models that promise autonomy. And through each wave, something essential was lost. Critical thinking. Human judgment. The courage to ask what really matters.

I will walk you through how testing got boxed into the tools it was supposed to use, how testers got reduced to tool operators, and how business leaders began chasing dashboards instead of risk awareness. This is not theory. It is drawn from real experiences. Fintechs that mistook coverage for confidence. Retail giants that handed testing to citizen users and paid the price. Teams that trusted AI too early and missed what mattered most.

But this talk is not just about what we lost. It is about what we can still take back. Because behind every narrative is a choice. And we still have one.

If you are a tester tired of defending your worth, a leader questioning your tooling ROI, or a team wondering why your velocity is up but your quality is down, this keynote is your mirror and your map. It will help you reconnect with testing as a craft of inquiry, not compliance. As a practice of thinking, not ticking boxes.

Because testing was never supposed to be a product. It was always meant to be a responsibility.

Testing AI models is not like testing traditional software, there is no single correct answer and its behavior changes depending on the context.

In this talk I will show how QA evolves facing generative systems, with real strategies like using validation datasets for controlled tests and automatic validations with semantic criteria. We will also see why metrics like faithfulness, answer relevancy and semantic similarity are key to ensuring quality in these products.