Qux360: A Validation Framework for Improving Reliability and Transparency of AI-Supported Qualitative Analysis

Driven by fast-paced product cycles and advancements in generative technology, the use of generative AI (GenAI) for qualitative analysis has become increasingly common in user research. However, user research requires robust qualitative methodologies, and the integration of GenAI creates challenges in maintaining rigor. This work introduces Qux360, a framework that implements validation as a first-class concept in AI-assisted qualitative analysis, based on pain points identified in exploratory interviews with user researchers. We demo an application that showcases the analysis and validation capabilities of Qux360, including validators that check for different aspects of quality in automated participant identification, topic extraction, and thematic analysis. By facilitating transparent analysis without compromising efficiency, Qux360 enables researchers to deliver actionable results while maintaining high standards of research quality and integrity.