“…Since 2012, there has been a proliferation of RPA definitions, covering three major aspects: (1) the subject of automationtasks, processes and repetitive tasks (Aguirre and Rodriguez, 2017;Fernandez and Aman, 2018;Kyher€ oinen, 2018;Lewicki et al, 2019;Morrison, 2019;Schmitz et al, 2019a;Theyssens, 2017) ; (2) the nature of automation and its effects-imitating and replacing of human work (Jalonen, 2017;K€ am€ ar€ ainen, 2018;Zhang and Liu, 2018); and (3) the nature of the RPA artefact-a software tool (Aguirre and Rodriguez, 2017;Leno et al, 2018;Morrison, 2019;Ratia et al, 2018;Suri et al, 2018;Theyssens, 2017). Two types of RPA are recognized by both practitioners and researchers: traditional RPA, as the initial form of RPA which is capable to automate processes where structured data and deterministic rules are available, and cognitive or intelligent RPA, as a more recent form of RPA, which grows hand in hand with artificial intelligence (AI) technologies to deal with unstructured data, by learning, reasoning and self-correction (Atroley et al, 2016;Lacity and Willcocks, 2018;Masood and Hashmi, 2019). Although 10 years old, there are wide expectations for rapid adoption of RPA across a wide range of sectors.…”