What Actually Happens When AI Self Heals a Test

What happens when AI self-heals a test? I asked Karim Jouini, CEO and co-founder of @Thunders-AI to explain. Thunders tests are tied to intent, not selectors. When the UI changes, the AI adapts at runtime instead of breaking the test. But that raises real questions: Can you see what the AI changed? Can you review it? And how do you know it didn't hide a real bug? Karim's answer is auditable self-healing: - Every self-heal is flagged, and the test owner gets an alert explaining what changed and why - You see how the step looked last run and how it looks now, plus a screenshot to check the visual - The run passes, but it's marked as "successful with something to look for" in run history and analytics - If it's a real bug, you override it and file a bug - Slow environments get the same treatment: the AI waits, but flags that it had to wait longer than usual "It's not completely autopilot. It's autopilot with flags.” Full episode here: https://testgld.link/the-mobile-testing-workshop/replay About Thunders Thunders' AI test agents write, run, and maintain your automated tests in plain language. No scripts, no maintenance. https://www.thunders.ai #SelfHealingTests #TestAutomation #QA