WeTest.Athens 2026

We've long had established methods for evaluating user behavior when it comes to practical, goal-oriented applications. But what happens when we go beyond that? When we attempt to build systems, environments or tools not meant for completing tasks, but for facilitating creativity or artistic self-expression? When AI makes it easier than ever for anyone to experiment with developing these types of experiences? How do we know if they're working well? And what does "working well" in that context even mean? Through an exploration of her past academic work, and how it relates to her current work in industry, Dalia will share practical lessons for anyone looking to design user experiences that push the bounds of what can readily be measured, all while touching on why doing so is especially crucial in the age of AI.

It can be tempting to treat accessibility as just a checklist of technical fixes. But the truth is that accessibility requires a general awareness from day one - an expectation and way of thinking across your entire organization and the teams within it. Accessibility is a process, not a one-off project.
 
The talk splits into two parts:
 

• How to integrate accessibility into your organization
• How to design usable and accessible digital content

 
Designers, developers, researchers, testers, and product managers will leave with practical knowledge they can embed in their daily practices and workflows.

What if I told you that you can trick an LLM into revealing secrets, making bad choices or even acting against its own rules? AI may seem like a black box, but when you start testing it like any other system, surprising weaknesses start to appear.

Together with me, I want you to take a look into how the OWASP Top 10 (a well-known list of the most common security issues in software) can help to understand and mitigate the risks linked to LLM applications. We’ll go beyond theory and explore real-world examples both from the news and from my experience at ongoing projects and see how these vulnerabilities play out in AI-driven systems.

We’ll check how poorly crafted prompts, weak security settings, biased training data, lack of proper safeguards can lead to serious security flaws. How easily can an attacker manipulate an LLM? What happens when sensitive data leaks through unintended AI behavior? Can a seemingly harmless chatbot be turned into a security risk? I’ll answer these questions while using my own cartoons to illustrate key risks in a fun and easy-to-understand way.

The good news? You don’t need to reinvent the wheel to test AI. Strong system thinking, traditional testing techniques, critical mindset are already powerful tools for uncovering vulnerabilities. The same skills used to break and improve software (like exploratory testing, risk analysis, extensive logging and monitoring) can help make AI systems safer and more predictable.

By the end of this talk, you’ll have a fresh perspective on AI risks, practical strategies to make your LLM integrations more secure, and, of course, a few laughs along the way.

Short one-liner: How Do We Trust Software Built by AI

Statement
AI agents are generating more code than ever, but how do you ensure its reliability? Learn why traditional testing fails and how Agentic Testing guarantees trust and scalability in AI-driven systems.

Abstract
AI agents are helping teams generate more code than ever before. Even the strongest believers in vibe coding caution against placing blind trust in that code. We are living through the fastest expansion of software creation in history, not because we hired more engineers, but because machines are now building software alongside us.This raises a fundamental challenge. How do you test and trust applications created by LLMs, assembled by multiple AI agents, and deployed by automated systems? The tools we rely on today were not designed for this reality or this speed. Unit tests and code reviews inspect logic and functions, but they were never meant to validate dynamic, evolving behavior. Traditional automation struggles when code is generated and modified continuously, becoming fragile and breaking faster than teams can maintain it. As AI agents become embedded into every layer of software, we are no longer just testing code. We are testing autonomous systems themselves.In this talk, we will explore why trust in AI-generated software can no longer be assumed, and why it must be continuously proven. Agentic Testing, the use of autonomous agents to test autonomous systems, emerges as the only scalable solution. These autonomous agents not only plan, execute, and analyze tests but also adapt and challenge systems the way real users would. Agentic Testing validates behavior, reasoning, and outcomes at the system level, ensuring that AI-driven systems meet real-world expectations.Drawing from real-world enterprise deployments and lessons learned building an AgenticTesting platform, this talk will reframe quality assurance as a strategic intelligence layer rather than a release-time checkbox. In the age of AI, trust isn’t just a feature, it’s a continuously earned asset. We’ll discuss how businesses can evolve their testing strategies to stay ahead in this rapidly shifting landscape.

Key Takeaways
● Why trust in AI-generated software must be continuously proven, not assumed.
● How Agentic Testing validates AI-driven systems by testing behavior, reasoning, and outcomes.
● Real-world insights on implementing Agentic Testing for scalable, autonomous validation.

Software testing is often seen as a technical gatekeeping step—but in reality, it’s a bridge between business goals, user needs, and data-driven insights. Drawing on my experience as a Test Analyst at the European Parliament and IT Consultant at Netcompany-Intrasoft, I’ll share how combining skills in statistical data analysis, economics, and quality assurance can create a richer testing practice. This talk will cover how testers can move beyond purely technical validation to become strategic contributors—spotting risks early, understanding business context, and improving the human experience of software. We’ll explore practical examples of merging manual testing, regression strategies, and user empathy to ensure quality that’s not just functional, but meaningful. Attendees will walk away with techniques for integrating data thinking into testing, fostering better collaboration with stakeholders, and ultimately delivering products that users trust.

The classic Developer x Designer x PM trio is supposed to work like a well-oiled machine… but in reality it's often three people arguing over who broke the designs, who forgot the edge cases, and who keeps adding 'just one more feature' to the MVP. With budgets tightening and AI speeding everything up, those old role boundaries make even less sense today, yet the friction remains. How do we collaborate without stepping on each other's toes, or egos?

In this talk, we'll cut through the "AI builder" buzzword and focus on what can work inside established orgs. Morgane will share concrete examples and practical advice to help teams work together more smoothly and with less eye-rolling.