In times of accelerated digital transformation, many organizations still struggle to put enterprise systems to effective use quickly. While prior work suggests either system or task complexity as a source for these difficulties, this case study of a major system implementation at a European bank reveals the most important source to be the complexity arising from co-dependency between the system and the task. We conceptualize this co-dependency as inherent in system-enabled tasks by proposing system dependency (the extent to which a task is supported by a system) and semantic dependency (the degree to which semantic understanding is required for task completion). Together, these dependencies create representational complexity, which constrains users from achieving effective use in system-enabled tasks and can explain differences in achieving effective use through variations in learning effort. The concepts and insights emerging from this study provide researchers and practitioners with a deeper understanding of what complexity means and why, in some contexts, learning how to use systems effectively takes longer.