How can you test a system that you don't understand?

Every tester hits that Monday-morning moment: an urgent ticket lands, nobody's free to explain it, and you know nothing about the feature. Drawing on their path from customer service through a coding bootcamp into QA, Kim reframes "I don't understand this" as "I don't understand this yet." They explain how reading a ticket for what's missing, not just what's there, surfaces unhandled edge cases and error paths, and points you to the people who have the answers.

Kim then gets practical about AI. They used Claude to design a learning experiment that revealed they'd been skimming rather than reading, which changed how they review code. They also used ChatGPT to stress-test a test plan, where it caught obvious risks they'd stopped checking because those always passed. 

Kim's takeaway: AI is a useful tool once you have the fundamentals to validate its output, it can't replace human feedback when learning from scratch, and specific prompts beat vague ones every time.