Visible but learning-weak: trace visibility and within-phase distortions in digitally supported continuous improvement

Purpose This paper aims to explain why digitally supported continuous improvement can remain visibly active while becoming learning-weak in enacted practice under trace visibility. Design/methodology/approach Using a qualitative, mechanism-oriented cross-case analysis of eight external case accounts from industrial manufacturing settings of different organizational sizes, supplemented by one reflective practitioner account used solely as contextual contrast, the study reconstructs phase-internal profiles of learning conditions, trace cues, recurrent distortions and boundary conditions. Findings Trace-producing improvement systems can sustain visible activity while enacted learning weakens under conditions of social exposure. The analysis identifies four corpus-grounded configurations: symbolic participation, scope retreat, parallel coordination channels and visibility without shared interpretation. Reporting drift is retained only as a bounded risk within the Value realization lens, while continuation without documented reconsideration is retained as a tentative risk pattern in the Renewal lens. Research limitations/implications The paper offers a bounded mechanism-based explanation rather than prevalence estimates, outcome claims or intervention-effect evidence. It supports analytic generalization about learning-weak enactment under trace visibility without proposing a maturity model. Originality/value Rather than adding another list of workplace-learning antecedents, the paper specifies how durable, attributable and reviewable digital traces can redirect established learning conditions within CI phases and synthesizes this distinct mechanism space of within-phase learning weakness into a compact Mechanism Atlas.