Programme launch offer - save 15% on tickets

AutomationSTAR

Test Automation Conference Europe

Track Talk W4

When AI Fails, QA Saves. Real Stories.

Wendy Erdheim

15:55-16:40 CET, Wednesday 4th November

AI ships code at lightning speed. QA stops the disasters. This is your origin story.

Right now, developers are using Cursor, ChatGPT, and other AI assistants to generate entire features in minutes. Velocity is through the roof. Test coverage looks amazing. And production is quietly breaking in ways your green builds never catch.

Here’s what’s happening: AI writes code with subtle flaws. Then AI writes tests that validate those same flaws. Your CI says ship it. Your users say it’s broken. And you’re the only one who can spot the difference.

I’ve caught authentication bugs that locked out thousands of users after every automated test passed. I’ve found race conditions AI couldn’t see. I’ve stopped security vulnerabilities that looked perfect in code review. Not because I’m special. Because I knew what questions AI never asks.

This talk arms you with that same superpower.

You’ll see real production failures where AI-generated tests gave false confidence and human QA saved the day. I’ll show you the exact bugs, the tests that missed them, and the Playwright patterns that caught them. You’ll learn to spot AI’s blindspots before they hit production.

You’ll walk away knowing which tests only you can write. The business logic validations AI doesn’t understand. The edge cases it never considers. The integration failures it can’t predict. I’ll give you the audit framework that turns “looks good to me” into “I know this actually works.”

The industry narrative says AI replaces QA. The reality? AI creates more bugs that slip past automation, making experienced QA engineers the most valuable people in the room. You’re not being replaced. You’re being promoted to the one role AI can’t do: thinking critically about what actually matters.

Stop letting AI lie to your CI. Start being the quality guardian your team desperately needs.

BACK TO PROGRAMME

What you will Learn

  1. Pattern Recognition - Spot the common AI test failures before they hit production: tests that check code runs instead of validating behavior, async blindspots, and business logic violations AI can't understand.
  2. The Audit Framework - Simple questions to ask about any AI-generated test: Can this pass when the feature is broken? Does AI know our business rules? Which tests to delegate to AI versus which require human judgment.
  3. Positioning Strategy - Language for explaining this work to developers and managers so you're seen as essential rather than a bottleneck, plus the Monday morning action plan for implementing it immediately.

Session Details

  • Introductory
  • 30 minutes
  • 15mins Q&A
  • Real-World Impact & Storytelling
Buy Conference Ticket

Session Speaker

Wendy Erdheim

Elementor, Israel

Wendy Erdheim‑Poch is a QA professional with a strong focus on both manual testing and automation development. At Elementor, she works to ensure quality across complex digital experiences by combining sharp exploratory instincts with maintainable, scalable automation. She’s passionate about accessibility and thoughtful testing strategies, and she’s always focused on making products more reliable, usable, and inclusive. Wendy brings creativity, curiosity, and a collaborative mindset to every project, always thinking ahead to catch issues before users ever see them.

part of the