Digital data at work: Management consulting, Public government and Healthcare compared

Technology, organization and work co-evolve (Barley, 1996). Suggesting that we think of this in terms of socio-material assemblages, Orlikowski (2007, p. 1445) shows how “assemblage shifts over time as interests, computers, networks, choices, algorithms, websites, preferences, links, identities, and capabilities change”. The production and use of digital data through algorithms, machine learning, and artificial intelligence technologies has to date mainly been methods for inquiry rather than objects of exploration in terms of how they are put to use and what the consequences are. This is not surprising as developments in how the production and use of digital data are moving fast across professional fields, driving institutional change (Hinings, Gegenhuber, & Greenwood, 2018) in expertise-based forms of work. But how do technological changes such as AI, machine learning algorithms, and data analytics in general shape the organization and practice of professional knowledge work? How are their products, i.e. digital data, used in professional work? Based on ongoing qualitative studies in management consulting, public government, and healthcare, this paper explores how digital data is put to use in three different professional settings and ask: “How does digital data affect knowledge work?” In doing so, we will further the empirical knowledge base and theories of how use of digital data alters professional knowledge work in practice. To explore how professionals use digital data and what the consequences might be for professional knowledge and work, we draw on empirical examples from three larger studies of different professional practices respectively; management consulting, public administration case work, and healthcare. Management consulting, public government and healthcare share similarities. Professional learning, knowledge and judgment are essential in all three fields, typically supported through a case-based reasoning - what is learned in one situation can be brought to bear on the next situation. As such, it is not unproblematic to assume that professional knowledge can be separated from practice in these kinds of settings. Greenhalgh and Wieringa (2011) suggest that phronesis, situated practical wisdom, is often downplayed in favour of understandings of knowledge as objective, context-free, and easily separated from the knowledge production process - in other words data abstractions. At the same time, the ways in which professionals’ use digital data in their work might offer insights into what kinds of professional knowledge are produced hereby and how this impacts professional work more broadly, for instance in terms of education, on the job training, or management (Beane, 2019), which is our focus. Our first case take place in management consulting, which is a relatively week profession without much institutional shelter (McKenna, 2006). This case is based on ongoing ethnographic fieldwork, where we focus on a management consultancy’s attempt at advancing a digital data tool that combines machine learning and behavioral science for people analytics with their public sector and private sector clients. Indeed, consultants play an important role in driving institutional change (Abrahamson, 1996, Dimaggio and Powell, 1983), carrying and circulating (rationalized) ideas and imaginaries of new management practices. By studying this practice, we show how the digital data mediated knowledge base alter the professional practice of management consulting. Our second case is drawn from public government, a traditional case-work setting in public administration in the field of agricultural policy. Digital data practices are already significantly changing how bureaucracy is practiced and works (Clarke & Margetts, 2014) and how professional expertise and knowledge is understood and distributed. Here, public government professionals handle ‘cases’ based on complex legal rules, procedural standards for decision-making, and accumulated work experiences. Clients are farmers and the core task is to implement public agriculture policy, while providing equal high quality public service to all clients. Case work is standardized, automated to a high degree, and each case is divided into separable parts (that case workers can have or produce data on). Thus, the ‘production’ flow of cases can be monitored and managed. Digital data are produced in part by data from individual farmers and from public databases. The case is thus expected to highlight changes due to the expansion of data analytics in the field of public management (Clarke and Margetts, 2014). Our third case is drawn from healthcare, which provides a case of strong professions (Abbott, 1996). Here we discuss how digital data and machine learning algorithms are increasingly positioned as solutions to quality and safety challenges, coordination challenges, and challenges relating to prioritization and use of resources. Clinical work is both standardized and particular to each patient: the healthcare practitioner’s professional judgement and decision is still seen as essential to ensuring safety and quality of care although use of AI technologies are high on the policy agenda (Government, 2019). AI technologies that offers increased diagnostic accuracy or prediction of the next step in the patient’s pathway are especially well-positioned to alter professional knowledge and decision-making in practice, because they will perform professional knowledge work alongside healthcare professionals (Faraj, Pachidi, & Sayegh, 2018). How will professionals’ navigate in and use the growing amounts of digital data and how will this shape their knowledgebase?