End-user understanding of automated process-related risk in a South African financial technology organisation

Abstract

This qualitative study aimed to explore how end-users understand automation-related risks in a South African financial technology organisation. Research on human–automation interactions suggests that there are operational risks when end-users adopt or use automated systems. For example, organisations do not dedicate enough time to ensure that end-users understand automated, process-related, built-in risks presented to them. Furthermore, documented information about how end-users understand automated process-related risks in financial technology organisations is scarce in the academic literature and also in the particular organisation in this study. Unlike the more usual top-down research approach where managers are first involved, a bottom-up approach was taken to address this research problem, focusing on end-users facilitating client-related transactions to consumers on behalf of various financial organisations. The author collected data through semi-structured, one-on-one interviews with 13 end-users of automated processes in the business. The sample had end-users from two organisational functions, handling client-related transaction processing and reconciliations, respectively. Reflexive thematic analysis was used to arrive at the themes presented in the study. Overall, the findings indicate that trust in automation is needed for end-users to understand automated process-related risks. Once trust in automated systems is present, end-users will notify developers of inefficiencies in automated processes, especially when they perceive those risks to reflect on their work responsibilities and roles. Participants cited personal time-wasting due to the unavailability of specific system functionality and increases in personal risk accountability due to overall system unavailability as contributing risk factors. Of special interest to the risk culture of the organisation in the study, a lack of knowledge sharing and inefficiencies related to power distance were found to be contributing factors.