Pixel-by-pixel UI testing is too brittle for modern release cycles, yet LLMs introduce a dangerous lack of determinism. In this practical roundtable, you will learn how to engineer a visual testing strategy that combines the flexibility of AI with the strict repeatability required for continuous deployment.

Join QA veterans John Howland, Jack Lukomski, and Taylor Fouts as they strip away the AI hype and look at real-world architectures that mimic human vision without sacrificing test stability.

What we’ll dissect:

• The Math Behind the Maintenance: Why pixel comparisons fail and how to drastically reduce the "self-healing" script burden.


• The Non-Deterministic AI Trap: Where LLMs break down in regression pipelines and how to enforce predictable outcomes.


• Hybrid Automation Architecture: How to successfully merge image-based and object-based testing into a single framework.


• Live Architectural Walkthrough: A first look at the algorithmic approach Qt is using to bridge the visual testing gap (and how to apply these concepts to your own stack).

This is a conversation, not a presentation. Bring your questions for a live Q&A with the panel.