Research map

The whole study on one page

Explore how each research question connects to the evidence and findings. Select a question to focus on it.

RQ1 · The question When did a poor fit between tools, information and the way work was organised change how a task could be done?

RQ2 · The question How did people respond, and when did an alternative way of working become part of regular work?

RQ3 · The question What work remained around GenAI outputs, and when could it be called invisible labour?

RQ1 · Key idea Sociotechnical friction: a poor fit that changes how the task proceeds, not just an inconvenience.

RQ2 · Key idea Workaround formation: an alternative that became part of the regular work of a task. Local adaptation is broader and includes responses whose workaround formation was not established.

RQ3 · Key idea Less visible work sits behind an output. Invisible labour also needs evidence that the work lacked enough time, support or recognition.

RQ1 · Builds on Trist (1981) on work systems; Orlikowski (1992, 2008) on technology in practice.

RQ2 · Builds on Alter (2014) on workarounds; Azad and King (2008) on fitting them to roles and timing.

RQ3 · Builds on Strauss (1985) on articulation work; Bainbridge (1983) and Simkute et al. (2025) on work shifting to monitoring and review.

RQ1 · My SDF workflow Matching staff records, allocations and transactions by hand each cycle; IT environment and permission requests, February to May 2025.

RQ2 · My SDF workflow A connected arrangement of Python processing, prepared files, Copilot questions and my checking became part of recurring SDF reporting.

RQ3 · My SDF workflow Checking, reconciling and explaining figures, such as a November 2025 reply about a claim paid from a separate award.

RQ1 · Interviews From other systems: SP06’s Teams service interruption, SP03’s bookings spread across several platforms, SP08’s research-software errors. From GenAI output: SP03’s incorrect course codes, SP05’s invalid classification codes, SP06’s feedback carried between students. Before wider use: review, sign-off and approval requirements (IP01–IP03).

RQ2 · Interviews Reported in use: SP05’s two internal agents, SP06’s attendance workbook, SP08’s troubleshooting. Partly rolled out, in development or being tested: IP01–IP03. Explored or proposed: SP05’s external tool, SP03’s booking agent, SP07’s reporting ideas.

RQ3 · Interviews SP01 reviewing learning-design material with academic colleagues; SP03 correcting course information already used; SP06 keeping feedback specific to each student. IP03 described list preparation as “invisible time”; SP07 had no set training time; SP04 reported time savings on some tasks.

RQ1 · What it established Friction mattered through its effects: interruption, correction after use, extra checking or coordination, restricted wider use, unclear responsibility. Not all of it came from GenAI.

RQ2 · What it established Workaround formation for the SDF arrangement. In the interviews, changes in use, in testing and proposed, with too little detail to trace how they formed.

RQ3 · What it established Less visible work in all three settings. Invisible labour was not established.

RQ1 · Contribution Friction defined by its consequences for the task.

RQ2 · Contribution Workaround formation separated from broader local adaptation.

RQ3 · Contribution Less visible work separated from invisible labour.

RQ1 · What it does not show How severe each problem was, or that any problem is unique to GenAI.

RQ2 · What it does not show A measured duration or frequency of SDF use.

RQ3 · What it does not show Measured effects on time or workload, or how the work was formally allocated and recognised.

Across all threeThe detailed workflow account showed how one arrangement developed. Interviews broadened the comparison to other tasks, but offered less detail about how changes became established. Thesis Section 7.7.2

Open the RQ1 page Thesis Sections 1.4, 2.9, 4.7, 5.8, 6.9 and 7.3–7.7 · adapted from Table 1.4.1