The impact of Generative Artificial Intelligence (AI; GenAI) on organisational dynamics remains understudied. Scholars emphasise that institutional change in a Public Administration (PA) context is shaped less by the technology itself and more by the behaviours and shared practices of public servants (e.g. Grimmelikhuijsen & Meijer, 2022). To examine these processes, there is a need for further empirical research to investigate how technological advancements affect bureaucratic conduct more broadly, and how the work of bureaucrats unfolds across multiple organisational settings (ibid; Raviola and Norback, 2013).
This study seeks to answer these calls by examining the micro-level dynamics of AI adoption within Poland’s public procurement institution. Specifically, the concept of institutional work (IW) is applied to achieve a better understanding of the work involved in AI introduction and to identify the actors engaged in these processes. The objective is to uncover how organisational dynamics are gradually shaped through everyday practices using the following research questions:
RQ1: What kind of IW is being done and by whom when AI is introduced into the PA context?
RQ2: How and why is IW triggered at the introduction of AI at the intersection of different organisational domains within the public sector?
To answer these research questions, the study employs an abductive longitudinal case study design focused on two interrelated public procurement organisations in Poland: a regulatory body and a judicial body, both in their early stages of GenAI adoption.
Theoretically, the study is grounded within the tradition of Organisational Institutionalism (OI), drawing on the perspective of IW and complemented by insights from Actor-Network Theory (ANT), particularly Callon’s (1984) sociology of translation. By combining these frameworks, the study integrates materiality into the analysis of organisational dynamics through the concepts of non-human actants and translation. AI introduction is therefore conceptualised as an emergent process of negotiation, where both relational and political aspects are considered.
The study will be predominantly based on semi-structured interviews, which will allow for a better understanding of the participants’ interpretation of IW, uncover their deeper motivations and practices, and navigate the translation process. Additionally, participant observations and document analysis will complement the findings, allowing for data triangulation. Data collection will take place in three different stages over approximately 6-12 months. The stages are interrelated with the phases of AI introduction, followed by the two institutional bodies. Data analysis follows the methodology outlined by Gioia et al. (2013); a process which has been tailored to the study’s specific aims.
The initial findings, in line with the scholarship, revealed AI to act as a trigger initiating IW in both of the studied organisations. After the first phase of AI introduction, actors made efforts to understand AI-driven changes, attempting to align them with existing norms, routines, and professional identities. Bureaucrats also initiated skill development to improve AI performance, especially where AI replaced repetitive tasks. However, when supporting complex cases, AI’s role has remained contested: legal professionals across both organisations view it as a research tool, while the IT department frames it as decision support. In later stages, I anticipate a shift in institutional discourse, specifically in the context of AI introduced to support more complex cases, where actors will either integrate AI into existing structures, beginning early institutionalisation or engage in IW to preserve the status quo. In both cases, AI’s institutional role will be redefined through translation processes. Nevertheless, although the socio-technical context of the institution will evolve, I expect both of the organisations tol retain their core functions. Regardless of AI’s success, the study will reveal who engages in IW and why. I anticipate actors at different levels of the organisations to engage in distinct forms of IW, where frontline bureaucrats will engage in forms of adaptation to adjust AI to their working routines, while managers are likely to take on the role of institutional entrepreneurs, aiming to legitimise the role of AI in the PA context.
Empirically, this study enhances understanding of how technological advancements affect bureaucratic conduct and, to the author's knowledge, is among the first to apply IW to AI adoption in PA in Central and Eastern Europe (CEE). Applying IW to this novel context contributes to the field by demonstrating its theoretical transferability and potentially expanding IW classifications. Theoretically, the study responds to recent calls (e.g., Baptista et al., 2020) by offering a more progressive framework to capture the relational and political dynamics between humans and technology, emphasising the ongoing work required to sustain these interactions.
Possible challenges include securing long-term access to both institutions and managing the time-intensive nature of the method; challenges that are addressed through an official contract, careful planning of visits, and continuous documentation. Variations in participants' role visibility and ethical considerations are mitigated by prioritising interviews and offering reflective opportunities to ensure mutual benefit. All non-public data are anonymised to protect identities. To track the inherently non-linear nature of IW, the study employs a three-stage longitudinal design. The researcher also maintains self-conscious awareness and continuous self-monitoring to minimise personal bias.
Baptista, J., Stein, M.-K., Klein, S., Watson-Manheim, M. B., & Lee, J. (2020). Digital work and organisational transformation: Emergent digital/human work configurations in modern organisations. Journal of Strategic Information Systems, 29(2), 101618.
Callon, M. (1984). Some elements of a sociology of translation: Domestication of the scallops and the fishermen of St. Brieuc Bay. The Sociological Review, 32(1_suppl), 196–233.
Gioia, D. A., Corley, K. G., & Hamilton, A. L. (2013). Seeking qualitative rigor in inductive research: Notes on the Gioia methodology. Organizational Research Methods, 16(1), 15–31.
Grimmelikhuijsen, S., & Meijer, A. (2022). Transparency and trust: An experimental study of online disclosure. Public Administration Review, 82(1), 110–124.
Raviola, E., & Norbäck, M. (2013). Bringing technology and meaning into institutional work: Making news at an Italian business newspaper. Organization Studies, 34(8), 1171–1194.
