This keynote discusses the move away from siloed QA towards shared quality across all stakeholders, to enable faster delivery and higher quality. By exposing the flaws of siloed QA (bottlenecks, late bugs) and highlighting the benefits of democratized quality (speed, better products, stronger teams), it provides a path to implement this crucial shift and recover a once-fundamental testing skill.

Democratized quality, a core testing principle, offers benefits such as faster delivery, higher quality, and stronger teams. The testing industry is rediscovering this essential skill, and the keynote will explain how to implement this crucial shift

De nombreux testeurs se demandent : « Que peut faire l’IA pour améliorer mes pratiques de QA ? », ou encore : « Va-t-elle me remplacer ? ». Nous pensons que la véritable question est ailleurs : Comment se préparer à tester efficacement une technologie déjà bien ancrée, en pleine expansion et toujours plus présente dans nos systèmes ?
Les tests d’IA sont différents des tests traditionnels, car les systèmes d’IA ne sont pas statiques, mais dynamiques et adaptatifs. Ils apprennent à partir des données et des retours, et modifient leur comportement en conséquence. Cela signifie que les tests d’IA doivent prendre en compte la variabilité et l’imprévisibilité de ces systèmes, et veiller non seulement à leur bon fonctionnement, mais aussi à leur fiabilité, leur éthique, et leur conformité aux exigences spécifiques de chaque secteur. C’est là que le rôle de testeur prend toute son importance.

Dans ce keynote, David explorera les nouveaux défis que l’IA impose à la QA, les opportunités qu’elle ouvre, et les pratiques clés à adopter. À travers des exemples concrets, nous verrons pourquoi les tests d’IA ne peuvent pas être abordés comme des tests traditionnels et comment adapter nos méthodes et nos outils du quotidien pour garantir qualité, fiabilité et conformité.
L’IA n’attend pas. Le train est lancé — ne restez pas sur le quai.

The test pyramid. It’s almost 15 years old now. Some say it’s passed its due date. Well, in my experience most people actually do not understand it and don’t apply it correctly!

The software development industry seems focused on doing everything as End to End tests, which the pyramid states should be minimal. Why is there such a disconnect?

I want to properly explain how you should layer your tests. Let’s be practical – I will walk you through a simple (but thorough!) example of testing a web page and show you where your tests should go in the pyramid.

• Unit tests? Well of course, my developers do those ALL THE TIME, wink wink!
• Tons of API tests done at integration level? Check!
• UI tests not done as End to End tests? Wait, what? I know, I was there! I made the mistake of equating those things in the past! But not anymore!

Learn how to properly setup your tests (and test automation) to test things at the lowest level of the test pyramid, to be as efficient as possible.

• Be reminded that sorting and/or filtering a table are mostly API tests, not UI tests
• Introduction to Selenium Webdriver BiDi as a way to confirm that they UI is making the right API calls, (something that was previously only possible in Cypress and Playwright)
• Quick intro on how to do integration testing at UI level with modern test frameworks

how artificial intelligence complements testing. Types of solutions that can be implemented for different platforms. Horizon can be seen in testing with AI. Myth or reality – will we lose our jobs?

How can we provide solutions in CRM and ERP applications where acceptance criteria can vary? What challenges might arise? How can we rely on AI? What can be achieved and what are the limitations?

Expert in trustworthy software and AI with dual faculty roles in Canada and Ireland. Canada Research Chair at uOttawa and leader at Lero. Over 30 years’ global experience in software engineering, industry collaboration, and innovation. IEEE and ACM Fellow. Passionate about impactful research and teaching.