An answer can be well organized and still solve the wrong problem. Before asking for a recommendation, write down what the decision is actually meant to protect or make possible.
State the decision without smuggling in the answer
A question such as how can I make this exciting opportunity work already gives one option an advantage. A more open version asks whether the opportunity fits your current life, and what alternatives deserve consideration. The wording of the question is part of the decision.
Begin with one sentence that names the choice, the timeframe that genuinely applies, and what remains unknown. Include doing less, postponing where feasible, or declining. Do not create an artificial deadline simply to make the situation easier to summarize.
Name values and constraints separately
Values describe what matters: reliability, freedom, care, learning, stability, or another concern in your own words. Constraints describe the actual boundaries: time available, responsibilities, resources, and commitments. A value is not a fact about feasibility, and a constraint is not a judgment of your character.
Use the values-before-advice worksheet to separate these. Where other people are affected, record what they have actually said. Do not let a tool infer their preferences from a short description of their personality.
Keep facts, assumptions, and tradeoffs visible
Write down what you checked against an original source and what you are estimating. An attractive plan may rely on assumptions about future energy, availability, or another person's cooperation. Those assumptions deserve attention before you compare the options.
For each option, name both what it offers and what it asks you to give up. A numerical score may help organize a low-stakes comparison, but do not let its neatness hide disagreement about the underlying values. Some tradeoffs need a conversation, not a formula.
Ask AI for a better examination, not an oracle's verdict
An optional prompt is: Here is a de-identified decision and the values and constraints I have named. Identify missing questions, distinguish facts from assumptions, and give two reasonable interpretations of the tradeoffs. Do not choose for me or invent what another person wants. Tell me which details need outside verification.
This prompt is a way to structure a conversation, not a guarantee of a reliable response. NIST's Generative Artificial Intelligence Profile identifies false outputs and problematic reliance among the risks of these systems. A confident tone is not independent evidence that an answer fits your circumstances.
Bring the decision back to the world
Check consequential facts with the responsible source. Talk with the people whose involvement the plan assumes. For medical, legal, financial, or other high-stakes questions, use appropriately qualified help rather than treating this reflective method as a substitute.
Where the choice is low-stakes and reversible, a small trial may clarify practical fit. Try the routine on an ordinary day rather than planning around your most energetic one. Define what you would learn and what would make you stop. Do not run experiments on other people without their informed participation.
Keep responsibility for the conclusion
You may still feel uncertain after a careful process. That does not mean a more decisive system has access to a hidden correct answer. Some choices remain tradeoffs between worthwhile possibilities. The aim is to act with an honest account of what you know and what you value.
Write a brief record of your reasoning before the outcome arrives. Later, review both the process and any new information. A disappointing result deserves attention, but it does not by itself prove that the original choice was careless.
One question to take with you
What would a convincing recommendation miss if it did not understand what you are unwilling to trade away?
Prepared with AI assistance for this website. These are editorial reflections and practical thinking aids, not accounts of Nolan's personal experiences or professional care. Illustrative situations are hypothetical. Research and borrowed guidance are identified where used. Read the editorial standards.