What motivates use of an expert system? Recent studies have found that the anticipated performance benefits of using an expert system -- such as increases in decision quality, consistency, and speed of decision making -- can lead to increases in expected usage. But is motivation limited to performance benefits? Findings in job design theory suggest that other factors -- such as increasing a user's sense of control over a task or making a task less routine -- might also have an impact. If so, understanding these factors could be extremely valuable to managers seeking to build expert systems that will be readily accepted by users. This paper synthesizes findings from expert systems, information systems, and job design research to model how the task change experienced by an expert systems user during adoption can affect that user's motivation to continue using the system. Using existing task constructs from the job design literature, a simplified version of the model is operationalized and tested on a data set of expert systems (all constructed in the early and mid-1980s) for which extensive quantitative and qualitative task change data was available, as well as data on systems usage. The findings suggest significant relationships between the nature of the task changes associated with adoption and long-term usage of the systems, all consistent with the predictions of the job design literature. The study, therefore, concludes that a job design perspective of expert systems adoption can be a valuable tool in predicting user acceptance and, ultimately, systems usage.