A planning question for educators
Keep the original
question on the page.
When a pilot ends, make it easy to see what you were trying to learn.
Return to the question.
The October 16 Field Notes asked what would count as progress. The planning conversation behind that edition sharpened a goal by making its evidence explicit. Here is the next practical step: keep that original wording available when you come back to review the work.
A finished activity can be useful without answering the question that started it. That distinction matters when a school team is deciding whether to keep using a tool.
A hypothetical pilot
Imagine a team trying an AI tool to help draft meeting agendas. The question is whether it reduces preparation time while preserving the information participants need. Creating an agenda establishes that the tool produced a document. The team still needs to inspect its usefulness and count the human revision time.
Keep the scope small.
Use a fictional meeting and made-up details for the first rehearsal. Follow local policy and approved-tool rules. There is no need to put identifiable student work or private staff information into this exercise.
Give the review a small, readable record. Keep the original question, the evidence and the next decision together. This is a suggested pilot-review practice, not an additional formal component of Objectives and Key Results.
Before
What are we checking?
Write down the intended change, the evidence you will collect and who will review it. Include the time spent checking and correcting the output.
After
What happened?
Compare the result with that same question. Note what the evidence supports, what remains uncertain and any unexpected work.
Next
What will we do?
The responsible human team decides whether to continue, adjust or stop. Record a reason someone outside the pilot could understand.
Leave room
to learn.
In the agenda example, suppose the first draft takes less time but the corrections take longer. That would be worth knowing. So would finding that the agenda is useful only for a narrow kind of meeting. Neither finding needs to disappear behind a cheerful completion report.
If the team discovers a better question, write it down as a revision. Keep the first one, too. The record should show how the team's thinking changed.
Keep responsibility human.
An AI-generated summary can help organize the evidence. The people responsible for the work still decide whether the result is adequate and whether the next use is appropriate. A small pilot does not establish schoolwide readiness.
Make the review
easy to follow.
Choose one small, approved experiment. Put the original question at the top of the review page and keep the result underneath it.
The question
What would you need to see before deciding to use the tool again?