When user needs do not align with system designers' visions, new technology implementation becomes a complex process as users appropriate the new technology to meet their needs. Prior studies recognize this complexity, but focus on the complex implementation of simple systems in which user groups are well defined and the IT artifact is the primary change. We extend the research lens by examining the implementation of the Brazilian correspondent banking system, a complex system involving multiple actors, system elements, and settings intended to address the social problem of financial exclusion. Our case study comparison of two settingsÑretail stores and post officesÑreveals that actors' appropriations extended beyond the IT artifact to include technical, role, usage, social, and policy appropriations. The intended users (poor clients in remote and underserved areas) barely interacted with the IT artifact or other system elements; instead, they relied upon remote bankers (correspondents) to appropriate the system on their behalf. Because rewards, incentives, and constraints differed by setting, correspondents' appropriations differed by setting. We call the resulting mix of appropriations across multiple elements by multiple actors in multiple settings multiplex appropriation. Complex societal challenges often involve multiple users in multiple settings with varied needs and few technology skills; thus, designing systems to meet user requirements may prove impossible. Instead, allowing multiplex appropriation might foster system success because, rather than forcing a global alignment among system elements or trying to ascertain multiple user needs, it allows for multiple local alignments of system elements that fit local settings.
Studies have shown the knowledge transfer problems that arise when communication and storage technologies are employed to accomplish work across time and space. Much less is known about knowledge transfer problems associated with transformational technologies, which afford the creation, modification, and manipulation of digital artifacts. Yet, these technologies play a critical role in offshoring by allowing the distribution of work at the task level, what we call task-based offshoring. For example, computer-aided engineering applications transform input like physical dimensions, location coordinates, and material properties into computational models that can be shared electronically among engineers around the world as they work together on analysis tasks. Digital artifacts created via transformational technologies often embody implicit knowledge that must be correctly interpreted to successfully act upon the artifacts. To explore what problems might arise in interpreting this implicit knowledge across time and space, and how individuals might remedy these problems, we studied a firm that sent engineering tasks from home sites in Mexico and the United States to an offshore site in India. Despite having proper formal education and ample tool skills, the Indian engineers had difficulty interpreting the implicit knowledge embodied in artifacts sent to them from Mexico and the United States. To resolve and prevent the problems that subsequently arose, individuals from the home sites developed five new work practices to transfer occupational knowledge to the offshore site. The five practices were defining requirements, monitoring progress, fixing returns, routing tasks strategically, and filtering quality. The extent to which sending engineers in our study were free from having to enact these new work practices because on-site coordinators acted on their behalf predicted their perceptions of the effectiveness of the offshoring arrangement, but Indian engineers preferred learning from sending engineers, not on-site coordinators. Our study contributes to theories of knowledge transfer and has practical implications for managing task-based offshoring arrangements.