Reflective journal My account of the SDF work, written in February 2026 about events in 2025.
Technical materials The code, processing logic and diagrams of what I built, showing how it worked.
Dated messages Messages from the time, used only to confirm dates and the order of events.
Participant account What an interviewee described. Not observed or checked independently.
Interpretation dashed outline Conclusions drawn from the evidence.
Stage of use Whether a change was in use, being tested, explored or proposed. These labels describe a stage, not whether the change succeeded.
Master of Science (Research) in Computing Sciences · University of Technology Sydney
Sociotechnical friction, workaround formation and less visible work
When does GenAI become workable in university work?
In the settings I studied, a capable GenAI tool was not enough. Whether the work could be done dependably also depended on the information, the task, how the work was organised, how people responded to problems, and who stayed responsible for the results.
Thesis Sections 7.6 and 7.9
The problem
Part of my job was preparing statements of each staff member's Staff Development Fund (SDF) spending against their professional development allocation. A statement, draft or summary can look complete on screen. What it does not show is everything that made it usable.
Staff statement
Example layout · no real dataWhat you see
A total, ready to send.
What it does not show:
RQ1
What made the work difficult
Matching records by hand each cycle; finding the right IT environment and permissions.
RQ2
How I responded
Building an arrangement that kept records linked and figures traceable.
RQ3
What work remained
Checking figures, reconciling them with the funding context, explaining them to staff.
These became the three research questions, asked of my own workflow and of other university work. They are not fixed steps, and not every one arose in every setting. Adapted from Thesis Figure 1.2.1.
In November 2025 a staff member replied to their statement. A recent large claim had been paid from a separate award, not the development fund, and they said they were not trained to read financial spreadsheets. The exchange showed why a figure needed to be understood in relation to the funding source covering the expenditure.
Paraphrased from my reflective journal, with the date confirmed by the messages · Thesis Section 4.6.2 · one exchange, not a measure of how often this happened
Three research questions
RQ1 · Friction
I call this sociotechnical friction when the poor fit caused delay, correction, extra checking or unclear responsibility, not just inconvenience.
Found: friction mattered through its effects on the task. Some of it came from other systems, not from GenAI.
Follow RQ1 →RQ2 · Responses
Participants described changes already in use, alongside testing, exploration and proposals. I use workaround formation for the case where an alternative became part of the regular work of completing a task.
Found: my SDF reporting arrangement showed workaround formation. The interviews provided less detail about how their changes developed.
Follow RQ2 →RQ3 · Continuing work
Preparing, checking and explaining often sat behind the output, so it was less visible. Calling it invisible labour would also need evidence that it lacked enough time, support or recognition.
Found: less visible work appeared in all three settings. The evidence did not establish invisible labour.
Follow RQ3 →How the three ideas connect
Friction can prompt a response. A response can move work elsewhere or bring hidden work into view. That continuing work can reveal new friction. These are possible links, not fixed steps.
Adapted from Thesis Figure 2.9.4.1
What I did
I studied my own reporting workflow from the inside, using a reflective journal written in February 2026 about events in 2025, alongside the technical materials I built. I examined all three settings through the same research questions. My workflow provided a detailed account of how one arrangement developed, while interviews showed how other staff described their work.
1
My SDF reporting, studied through Studying a setting you took part in through your own recorded experience, and interpreting it with research concepts. My main record is a reflective journal written in February 2026 about events in 2025.Thesis Sections 3.2.1 and 3.5.1. It can show how one arrangement developed and became part of regular work.
The SDF workflow →3
People whose work involves university data, systems, AI development, governance or support (IP01–IP03). They described review, testing and support around their own projects.
Institutional support →8
Professional and academic staff in teaching, research and administration (SP01–SP08). They described using, adapting and checking GenAI in their own tasks.
Everyday staff →The 11 interviewees volunteered through the university's Microsoft 365 Copilot Teams channel, so they do not represent all staff. The three settings are compared side by side, not as stages of one process. How the study was designed (Thesis Sections 3.3.1 and 3.6.3).
Scope
Thesis Sections 3.7.2, 7.7 and 7.8