Before class
Can faculty review the material, correct assumptions, prepare accessible alternatives and explain what students will do? Check the distinction between AI drafts and faculty-approved teaching content.
Practical faculty resource
Evaluate a teaching workflow, not just a feature list.
Use this checklist when planning a demonstration or an institution-approved evaluation. It is a suggested process, not evidence that a tool improves learning or meets your institution's policies.
Start with a small, observable task
Choose a learning outcome and a fictional scenario. For example, ask a class to compare two inventory policies and explain the cost-service trade-off. Decide what would count as a good explanation before choosing a slide, activity or simulation.
During class
Can students understand the task and controls? Test joining, recovering from a connection interruption, and separating individual work from team decisions with demonstration accounts.
After class
Can faculty inspect the work behind a result and explain the evaluation criteria? Distinguish participation evidence, task performance and official marks. Missing work must not be mistaken for a completed zero.
A five-question simulation debrief
What did your team decide? What evidence did you use? Which assumption mattered most? What changed when the environment changed? What would you test next? Ask students to distinguish their own role contribution from the final team submission.
Check assessment policy before the activity
Decide whether the exercise is practice or assessed work. If assessed, explain criteria, attempt rules, permitted collaboration and the release process in advance. Do not turn ordinary practice clicks into marks after the event.
Keep the evaluation safe
Use fictional names and data in public demonstrations. Review permissions, student accessibility and export requirements with the appropriate institutional owners. Do not upload confidential records to a public enquiry form or third-party analytics tool.
Record what you actually observed
For each workflow, record the date, configuration, task, expected result, observed result and limitations. Mark untested capabilities as untested. Request a demonstration of the gaps instead of inferring them from a marketing description.