This session demonstrates how UX researchers can use AI as an “expert misinterpreter” to refine interview protocols, usability testing scripts, and survey questions. By applying prompt engineering through the lens of UX heuristics, participants will learn to identify hidden ambiguities, challenge assumptions, and improve the clarity and reliability of their research instruments. The session covers AI-driven thinking-aloud simulations, policy- and mandate-based scenario testing, and iterative refinement of questions before engaging human participants. Attendees will walk away with practical methods to pressure-test their protocols, improve reproducibility, and build trust in AI-assisted research processes. This approach ensures that UX research is more resilient, ethical, and evidence-based, while maintaining the human-centered rigor that drives actionable insights.