Working papers › Externalising tacit institutional knowledge · 2026
An analytic autoethnography of an AI-augmented compliance reporting workflow
Scholarship on generative AI in higher education has concentrated on student-facing applications, with limited attention to AI-augmented professional work in specialised compliance domains. This paper examines that neglected territory from the inside: a report governance system developed for multi-jurisdictional government compliance reporting at an Australian public university, and what its construction reveals about how AI mediation can help a practitioner render tacit institutional knowledge explicit, durable, auditable and transferable.
From the paper: a note of proportion is warranted, as the third column documents the System as established, in advance of independent deployment.
From the paper: the case is a report governance system designed and documented within a discrete organisational unit at an Australian public university, managing multi-jurisdictional compliance reporting to the Commonwealth Government and two state government departments for a suite of funded university coursework programs. The inherited context is recognisable to anyone who has held such work: obligations dispersed across approximately fifteen instruments with no single consolidated statement of operative requirements; each reporting cycle a near-bespoke effort; structural and language conventions held tacitly; calendar awareness and submission channels held by individuals; and high institutional exposure to staff transition.
The paper treats the case as both a design account and an analytic window onto AI-augmented institutional knowledge work, asking what design choices the system deployed across the dimensions of multi-jurisdictional compliance reporting, and what the case reveals about how AI-augmented work reshapes the codification, coordination and visibility of institutional knowledge.
Each dimension is mapped across the inherited reporting context, the AI capabilities deployed, and the governance design developed. Reading across the layers shows where tacit institutional knowledge was rendered explicit. Select a dimension to see all three layers.
A consistent pattern of partial externalisation runs through each of the five dimensions: aspects of institutional knowledge previously held substantially tacitly across multiple staff and dispersed files are rendered explicit through AI-mediated codification. A recurrent two-surface workflow pattern appears across dimensions, with the embedded data dictionary as the bridge, naming data points consistently so that the AI on each surface can engage without intervening translation. And the system exhibits design under conditions of human and AI capability change, with the model-agnostic principle operating as a commitment to operability across platforms.
Scholarship on generative artificial intelligence (AI) in higher education has concentrated on student-facing applications and prominent corporate services functions, with limited attention to AI-augmented professional work in specialised compliance domains. This paper addresses that gap through an analytic autoethnography of a report governance system developed for multi-jurisdictional government compliance reporting at an Australian public university. Drawing on a documentary audit trail, the analysis maps five dimensions of the System across three analytical layers: the inherited reporting context, the AI capabilities deployed and the governance design developed. The paper is grounded in organisational knowledge codification, boundary objects, infrastructure and articulation work.
The case shows how the requirements of AI-augmented work prompted tacit institutional knowledge to be articulated, structured and verified in explicit governance artefacts. It identifies partial AI-mediated externalisation as the central theoretical contribution and presents the two-surface AI workflow and AI model-agnostic design as potentially transferable design logics. The analysis also demonstrates how conventional outputs obscure the sophisticated design work supporting them, with implications for institutional recognition, professional development, system ownership and maintenance. The findings are bounded by a single-practitioner, single-institution case, with successor operation and accountability infrastructure identified as priorities for further inquiry.
The conversion of tacit into explicit knowledge (Nonaka & Takeuchi; Polanyi) supplies the paper’s central lens: the requirements of AI-augmented work — the need to state definitions, conventions and criteria so tools can act on them — operate as a codification forcing function on knowledge previously held in individual and corporate memory.
Artefacts that hold meaning across different communities of practice (Star & Griesemer): the master timeframes document, the templates and the data dictionary coordinate between internal compliance functions, successive role-holders and external funders without demanding shared context.
Infrastructure typically supports work invisibly until disruption or failure reveals it (Star & Ruhleder; Bowker & Star). The system’s folder conventions, archives and guides are infrastructure in exactly this sense, and inherit its central problem: what sustains them when their builder moves on.
The often invisible coordination labour that makes formal work processes actually function (Strauss; Suchman). The paper uses it to name the design labour behind conventional-looking outputs, and to ask what institutions owe to the labour sustaining their reporting.
The central theoretical contribution: AI mediation supported the practitioner to partially externalise tacit institutional knowledge into governance infrastructure designed for durability, auditability and transfer. Partial, deliberately: the paper does not claim complete codification, and the third analytical layer documents the system as established, in advance of independent deployment.
A transferable design logic separating structured data extraction (the spreadsheet surface) from generative composition (the LLM surface), with practitioner verification between them and the embedded data dictionary bridging the two.
The second transferable logic: standard formats, markdown-compatible content, placeholder and appendix conventions readable by any capable model, and no platform-specific dependencies — a design commitment to operability across technological change.
The system’s outputs are visually indistinguishable from artefacts produced manually, so the design work behind them goes largely unseen — with implications the paper draws out for institutional recognition, professional development, system ownership and maintenance.
The artefacts are visible while the design work behind them largely is not: outputs indistinguishable from manually produced documents conceal the sophisticated governance design supporting them.
From Section 4 of the paper. The findings are bounded by a single-practitioner, single-institution case, and the paper names its own open questions: successor operation, and the accountability infrastructure such systems will need.
From the paper: the study draws on the practitioner reflection tradition, which uses reflective and autoethnographic methods to examine knowledge developed through sustained engagement with generative AI, and extends that tradition from individual creative and academic practice into institutional governance work. Reflective accounts of AI-augmented governance system design are not yet a developed body of scholarship: practitioners positioned to produce such accounts may have limited outlets or incentives for documented reflection, while the necessary combination of compliance expertise, AI capability literacy and reflective discipline is unevenly distributed across institutional roles. The analysis is grounded in a documentary audit trail, and the case is anonymised throughout, as it is in the paper itself.
Adamakis, M., & Rachiotis, T. (2025). Artificial intelligence in higher education: A state-of-the-art overview of pedagogical integrity, artificial intelligence literacy, and policy integration. Encyclopedia, 5(4), 180. https://doi.org/10.3390/encyclopedia5040180
Almatrafi, O., Johri, A., & Lee, H. (2024). A systematic review of AI literacy conceptualization, constructs, and implementation and assessment efforts (2019-2023). Computers and Education Open, 6, 100173. https://doi.org/10.1016/j.caeo.2024.100173
Anderson, L. (2006). Analytic autoethnography. Journal of Contemporary Ethnography, 35(4), 373-395. https://doi.org/10.1177/0891241605280449
Bittle, K., & El-Gayar, O. (2025). Generative AI and academic integrity in higher education: A systematic review and research agenda. Information, 16(4), 296. https://doi.org/10.3390/info16040296
Blanco, C. (2026, May 20). Will AI deliver freedom from admin or greater mental load? University World News. https://www.universityworldnews.com/post.php?story=20260519152453263 Bourke, B. (2014). Positionality: Reflecting on the research process. The Qualitative Report, 19(33), 1-9. https://doi.org/10.46743/2160-3715/2014.1026
Bowker, G. C., & Star, S. L. (1999). Sorting things out: Classification and its consequences. MIT Press.
Buame, J. A., Tiika, B. J., & Lotsu, S. A. (2025). The role of AI in shaping the discharge of duties of university administrators: A systematic review. Pan-African Journal of Education and Social Sciences, 6(2), 13-31. https://doi.org/10.56893/pajes2025v06i02.02
Creely, E., & Blannin, J. (2025). Creative partnerships with generative AI. Possibilities for education and beyond. Thinking Skills and Creativity, 56, 101727. https://doi.org/10.1016/j.tsc.2024.101727
DeLong, D. W. (2004). Lost knowledge: Confronting the threat of an aging workforce. Oxford University Press.
Edwards, P. N. (2010). A vast machine: Computer models, climate data, and the politics of global warming. MIT Press.
Faraj, S., Perez-Torrents, J., Mantere, S., & Bhardwaj, A. (2026). A time for monsters: Organizational knowing after large language models. Strategic Organization, 24(2), 343-356. https://doi.org/10.1177/14761270251410675
Freeman, B. (2026). I’ll be your mirror: Disaggregating technological affordance from developed competence in AIaugmented practice [Working paper]. SSRN. https://doi.org/10.2139/ssrn.6527560
Hofer-Alfeis, J. (2008). Knowledge management solutions for the leaving expert issue. Journal of Knowledge Management, 12(4), 44-54. https://doi.org/10.1108/13673270810884246
Holmes, A. G. D. (2020). Researcher positionality: A consideration of its influence and place in qualitative research. A new researcher guide. Shanlax International Journal of Education, 8(4), 1-10. https://doi.org/10.34293/education.v8i4.3232
Jackson, E. (2013). Choosing a methodology: Philosophical underpinning. Practitioner Research in Higher Education, 7(1), 49-62. Le Gallais, T. (2008). Wherever I go there I am: Reflections on reflexivity and the research stance. Reflective Practice, 9(2), 145-155. https://doi.org/10.1080/14623940802005475
Liu, X., Guo, B., He, W., & Hu, X. (2025). Effects of generative artificial intelligence on K-12 and higher education students’ learning outcomes: A meta-analysis. Journal of Educational Computing Research, 63(5), 1249-1291. https://doi.org/10.1177/07356331251329185
Lodge, J. M., & Loble, L. (2026). Artificial intelligence, cognitive offloading and implications for education. University of Technology Sydney. https://doi.org/10.71741/4pyxmbnjaq.31302475
Ncube, P. D. N., Dzvapatsva, G. P., Matobobo, C., & Ranga, M. M. (2026). Redefining student assessment in AI-infused learning environments: A systematic review of challenges and strategies for academic integrity. AI and Ethics, 6, Article 68. https://doi.org/10.1007/s43681-025-00871-w
Nonaka, I., & Takeuchi, H. (1995). The knowledge-creating company: How Japanese companies create the dynamics of innovation. Oxford University Press.
Nonaka, I., Toyama, R., & Konno, N. (2000). SECI, ba and leadership: A unified model of dynamic knowledge creation. Long Range Planning, 33(1), 5-34. https://doi.org/10.1016/S0024-6301(99)00115-6
Panke, S. (2025). How can (A)I research this? An autoethnographic exploration of generative AI in research, teaching and instructional design. Journal of Teacher Education, 76(3), 230-244. https://doi.org/10.1177/00224871251325065
Polanyi, M. (1966). The tacit dimension. Doubleday.
Savin-Baden, M., & Major, C. H. (2013). Qualitative research: The essential guide to theory and practice. Routledge.
Schön, D. A. (1983). The reflective practitioner: How professionals think in action. Basic Books.
Schutz, A. (1976). Collected papers II: Studies in social theory. Martinus Nijhoff.
Star, S. L. (2010). This is not a boundary object: Reflections on the origin of a concept. Science, Technology, & Human Values, 35(5), 601-617. https://doi.org/10.1177/0162243910377624
Star, S. L., & Griesemer, J. R. (1989). Institutional ecology, ‘translations’ and boundary objects: Amateurs and professionals in Berkeley’s Museum of Vertebrate Zoology, 1907-39. Social Studies of Science, 19(3), 387-420. https://doi.org/10.1177/030631289019003001
Star, S. L., & Ruhleder, K. (1996). Steps toward an ecology of infrastructure: Design and access for large information spaces. Information Systems Research, 7(1), 111-134. https://doi.org/10.1287/isre.7.1.111
Strauss, A. (1985). Work and the division of labor. The Sociological Quarterly, 26(1), 1-19. https://doi.org/10.1111/j.1533-8525.1985.tb00212.x
Strauss, A. (1988). The articulation of project work: An organizational process. The Sociological Quarterly, 29(2), 163-178. https://doi.org/10.1111/j.1533-8525.1988.tb01249.x
Suchman, L. (1995). Making work visible. Communications of the ACM, 38(9), 56-64. https://doi.org/10.1145/223248.223263
Walsh, J. P., & Ungson, G. R. (1991). Organizational memory. The Academy of Management Review, 16(1), 57–91. https://doi.org/10.5465/amr.1991.4278992
Woelert, P., Chesters, J., Martinussen, M., & Gannaway, J. (2026). Administrative burden in Australian universities: Insights into dimensions and drivers from a nationwide survey. Science and Public Policy, 53(2), 171–183. https://doi.org/10.1093/scipol/scaf029
Further papers will be added to this list as they are released. Each paper is published as its own page, with the full paper available for download as a PDF.