The Predictive Evolution of UX Through Cognitive Architectures from Intuitive Craft to Predictive Science

Traditional user experience design has long relied on intuition, aesthetic judgment, and retrospective usability testing to identify and resolve interface problems. However, such reactive approaches often neglect the cognitive mechanisms that underlie human perception, attention, and decision-making. This paper introduces a cognitive-science-driven framework for proactive, predictive UX design grounded in cognitive architectures—computational models such as ACT-R, Soar, and CLARION that simulate how the human mind processes information, makes decisions, and recovers from errors. Through a comparative analysis of these architectures, the paper demonstrates how evidence-based UX design can shift from intuition-led artistry to an empirically validated science. It outlines how cognitive modelling enables designers to anticipate user errors, quantify cognitive workload, and optimise interface flows before implementation. The study also provides a glossary of key cognitive terms, model templates for ACT-R and EPIC, and pseudo-code examples for applied HCI contexts. The resulting framework offers a replicable methodology for embedding cognitive predictions into design evaluation, promoting safer, more efficient, and cognitively aligned interfaces across domains such as healthcare, e-commerce, and productivity applications.