Think & decide

Prepare key questions for a meeting

Prioritize questions that expose evidence, trade-offs, authority, and next-step requirements

5 min setupTested with:ChatGPTReviewed: 2026-08-28

Add your context

Your text stays in this browser. AILesson Prompts does not send it to a model or server.

Meeting outcomeKnown facts and uncertaintiesParticipants and time

Your prompt

Unfilled fields remain visible as placeholders, so you can still copy and edit the prompt

Prepare the key questions this meeting must answer.

Required outcome:
[outcome]

Known evidence, assumptions, disputes, and missing inputs:
[knowns]

Participants, expertise, authority, interests, and available time:
[participants]

Work backward from the required outcome. Produce only questions whose answers could change the decision, plan, or next action. Group and sequence them as framing, evidence, alternatives, risks, decision, and execution. For each question state why it matters, best-positioned respondent, evidence needed, and one follow-up if the answer is vague. Distinguish factual, judgment, and decision questions. Do not embed an unsupported conclusion in a leading question or assume an attendee has authority not stated. Mark questions better answered asynchronously and identify the minimum questions if time is cut in half.

Private by defaultPrompt assembly happens locally in your browser. Avoid placing confidential information into any AI service unless your organization allows it.

From input to outcome

A worked example

See how concrete context turns this recipe into a usable result

Actual input

Meeting outcome: Decide whether to extend a six-week customer-support chatbot pilot for another month, stop it, or move toward rollout; define the next evidence needed.
Known facts and uncertainties: Resolution time fell 11% for 600 pilot conversations. Escalation rate rose from 14% to 19%. Customer satisfaction has only 73 responses and is unchanged within the margin of error. Two severe answer errors were reported; root-cause review is incomplete. Costs for rollout are estimates. Support proposes extension; Sales favors rollout.
Participants and time: 30 minutes. Support lead owns operations, data analyst knows metrics, safety lead owns error review, finance partner knows estimates, VP Customer has decision authority. Sales lead attends. Detailed data table can be read beforehand.

Example output

Framing: Which choice must be made today, and may “extend” include changed safeguards? Why: bounds the decision. Respondent: VP Customer. Type: decision. Follow-up: what may remain open after today?

Evidence: Are the 11% resolution improvement and 14%→19% escalation change comparable across conversation types? Analyst; needs segmented counts and definitions. Follow-up: which segments drive each change? What do the 73 satisfaction responses permit us to conclude? Analyst; needs response rate and uncertainty. Safety: what is known and unknown about the two severe errors, and what condition must be met before wider exposure? Safety lead; needs incident review.

Alternatives and risk: What safeguards, scope, and evidence would a one-month extension add? Support and safety. What assumptions drive rollout cost ranges? Finance. What customer or commercial evidence supports immediate rollout beyond preference? Sales.

Decision: Against agreed safety, customer, operational, and cost criteria, which option is justified now? VP Customer. Execution: what owner, deadline, sample size, and stopping rule apply to missing evidence?

Async: detailed segment table and cost model. Half-time minimum: framing; severe-error threshold; segmented escalation result; option against criteria; owners and stopping rule.

Why this works

1
Outcome-linked questions reduce interesting discussion that cannot affect the meeting result.
2
Evidence and respondent fields reveal questions the room is not yet equipped to answer.