A Decision Support Tool to Select Candidate Business Processes in Robotic Process Automation (RPA): An Empirical Study

Abstract

AbstractRobotic process automation (RPA) is the automation of business processes (BPs) using software robots. RPA robots automate repetitive, non-value-adding human work. The extent that a BP can be transformed into a software robot and its utility depends on several factors such as task type, complexity, repeated use, and regulatory compliance. The estimated RPA project failure rates are relatively high, and transforming the wrong BP is attributed as one of the critical reasons for this. Therefore, given a candidate set of BPs, it is imperative to identify only the suitable ones for RPA transformation. In this paper, a decision support tool (DST) is presented to select candidate BPs for RPA. First, 25 factors are identified from the literature that captures the characteristics of a BP. The list is then reduced to 16 factors based on a set of interviews with RPA subject matter experts (SMEs). Then an online survey with snowball sampling was conducted to measure the relevance of those factors in predicting the outcome of RPA transformation of a candidate BP. Finally, the two-class decision forest classification model was used to develop the DST. The utility of the proposed DST was validated by applying it to three RPA projects of a global IT consulting services company.KeywordsBusiness process automationDecision forest algorithmDecision support toolRobotic process automation