Purpose
Use this lesson to convert discussion notes into accountable follow-up work. The aim is to produce work that a colleague can review and act on, not merely text that sounds convincing.
Lesson Explanation
Scenario: Notes mention a campaign delay, an owner, and a due date.
Better instruction: Extract action items into a table: action, owner, due date, dependency, and unresolved question. Use “Unassigned” when no owner is stated.
Meeting notes mix decisions, opinions, and follow-ups. Give a schema that separates them and a rule for missing ownership. This prevents the model from silently assigning work to the wrong person. Every action item has the five fields, with unstated owners marked Unassigned.
Work in a short cycle: provide the relevant input, state the result and limits, inspect the draft against the acceptance criteria, then revise the instruction when a requirement is missing. Keep the final human decision with the person responsible for the work.
Practice Questions
-
A prompt asks an AI to “summarize the meeting” without specifying how to structure follow-up tasks. What does this lesson say meeting notes typically mix together?
View Answer
Decisions, opinions, and follow-ups, all blended together, without a clear separation.
-
What five fields does this lesson recommend using to extract action items from meeting notes?
View Answer
Action, owner, due date, dependency, and unresolved question.
-
Meeting notes mention a task with no named owner. What does this lesson recommend doing with this?
View Answer
Marking it “Unassigned” rather than the model silently assigning it to the wrong person.
-
Why does this lesson warn against letting the AI silently assign ownership when notes don’t name one?
View Answer
This can result in the model silently assigning work to the wrong person, creating confusion or missed responsibility.
-
Write a prompt instruction for extracting action items from raw meeting notes, following this lesson’s five-field structure.
View Answer
Something like: “Extract all action items into a table with columns: Action, Owner, Due Date, Dependency, and Unresolved Question. If no owner is stated, write ‘Unassigned.'”
-
A set of extracted action items includes one with an owner that wasn’t actually named in the original notes. What does this suggest went wrong, per this lesson?
View Answer
The AI likely guessed at an owner rather than marking the task “Unassigned” as instructed – or the prompt failed to include this explicit rule.
-
What problem does this lesson’s five-field schema (action, owner, due date, dependency, unresolved question) solve, compared to a plain paragraph summary?
View Answer
It separates decisions, opinions, and follow-ups into a clear, consistent structure rather than leaving them mixed together in unstructured text.
-
Meeting notes mention that a task depends on legal approval, but this dependency isn’t explicitly extracted in the AI’s action item table. What element of this lesson’s schema was missed?
View Answer
Dependency.
-
A team member notices an “unresolved question” from the meeting was never flagged in the extracted action items. Why might this lesson consider this a meaningful gap?
View Answer
The unresolved-question field is part of the recommended schema specifically because open questions matter for follow-up, and omitting them loses important information from the notes.
-
A prompt for extracting action items doesn’t specify what to do when no owner is named. What specific fix does this lesson recommend?
View Answer
Adding an explicit rule: use “Unassigned” when no owner is stated.
-
Why might mixing decisions, opinions, and follow-ups together (without separation) make meeting notes less useful for tracking work, according to this lesson?
View Answer
It becomes harder to identify concrete, actionable items versus general discussion, making it easy to lose track of what actually needs to happen next.
-
A completed extraction shows five action items, each with an action, owner (or “Unassigned”), due date, dependency, and unresolved question filled in. What does this reflect?
View Answer
A complete, well-structured extraction following the five-field schema this lesson recommends.