Human-automation interaction in organizations’ automation efforts

In light of the comprehensive theoretical discussion on function allocation, evidence of human-automation interaction in practice is scarce. This, however, is a prerequisite for appraising the applicability of theoretical models in practice and adjusting models if necessary. The present study explores how organizations structure and design human-automation interaction when automating. Based on data from semi-structured interviews we find a process consisting of three interlocked and iterative phases when automating: the requirements & process analysis phase, the business case phase, and the target process & task design phase. Surprisingly, companies are not consciously designing human-automation interaction, but this happens implicitly while building various scenarios in the target process & task design phase. The decision which task is automated through which scenario is taken via a business case. We conclude by proposing a shift in human-automation interaction research towards models including financial quantification as well as their interplay with additional, non-quantifiable dimensions. • Empirical evidence of human-automation interaction in practice. • Firms automate in three iterative phases: analysis, business case, and task design. • Human-automation design happens implicitly through scenario-building in task design. • Task automation decisions rely on business cases within the task design phase. • Financial and qualitative factors should be better integrated into interaction models.