Extended reality visual language for professional contexts: a study on visual dissonance

Extended Reality (XR) is increasingly deployed in professional workflows, where adoption research has focused on hardware ergonomics and cybersickness. As these barriers are addressed, an open question concerns how such interfaces should be visually designed. Professional XR often inherits the visual register of consumer systems, with saturated pal - ettes, gamified treatments, and stylized characters that may conflict with workplace norms. We define this mismatch as visual dissonance, and professionally aligned designs as native visual language. Established acceptance models capture the social conditioning of adoption through generic normative beliefs, but carry no construct for the visual channel that signals workplace conventions in an immersive medium. This study examines whether visual dissonance undermines per - ceived credibility across three XR interface components: workspace layouts, avatar representations, and data dashboards. A within-subject survey (N = 96) compared paired aligned and dissonant stimuli with functional content held constant. Sensitivity proves component-specific rather than a single dispositional trait. When asked which variant they would trust professionally, participants selected the aligned one for both avatars and dashboards, at rates that did not differ between these two components, while for workspace layouts the choice was balanced. When each variant was rated separately, only avatars showed a reliable difference: the dashboard effect appears only when a choice is required. Graded self-report alone may underestimate visual effects; XR familiarity amplifies rather than reduces sensitivity. These findings position the visual register of professional XR as a measurable adoption factor to be evaluated component by component and through both measurement channels.