Read the evidence at the right level
Edgar, Biloslavo and Rizzo's paper on AI and change management is listed by Glasgow Caledonian University as accepted/in press from 12 October 2026. Its narrative review of 47 sources considers AI both as an aid to change and as a continuing source of disruption. Cultural and ethical questions are part of that account.
A related conceptual paper by Biloslavo and colleagues connects AI and strategic planning in uncertain environments through the Cynefin framework. These contributions develop ideas and research directions. They do not provide a controlled demonstration that introducing an AI tool will improve a particular organisation's results.
Look beyond the demonstration
Consider a fictional team testing a tool that drafts internal reports. A convincing demonstration may show a fluent report in seconds. The team still needs to learn who checks the output, how errors are corrected and whether the draft helps the eventual reader. This example is an editorial teaching device, not a case drawn from the papers.
The same tool might move work between colleagues. Someone may spend less time drafting and more time verifying, while another person becomes responsible for maintaining source material. A useful discussion would ask each participant to describe the work before and after the change, including tasks that remain difficult to see.
Make the learning visible
For a small exploratory trial, define the task, comparison and review period before interpreting the results. Record both useful outputs and corrections. Invite the people doing the work to explain what the measurements miss. A faster draft and a more useful report are separate outcomes and may need different evidence.
Treat the trial as a question about a working arrangement, including roles, skills and decision authority. Any conclusion should identify the setting and its limits. A team can then decide which assumption deserves another test instead of allowing one successful demonstration to carry a much larger claim.
Questions to carry forward
- Which tasks move to a different person when the tool is introduced?
- What would count as a useful result for the reader of the output?
- Which observation would lead the team to revise its initial plan?