Identify the decisions in a task I should not hand to AI
Pinpoints the decisions, ambiguities, and exceptions in a task that a person must own, and why AI should not decide them.
The prompt
Help me find where human judgment matters most in a task I want to use AI for.
The task or workflow: [DESCRIBE IT, or PASTE the steps]
Who is affected by the outcome: [customers, employees, partners, the public]
What is at stake: [money, reputation, legal exposure, relationships, safety]
How I plan to use AI: [e.g. drafting, analysis, recommendations, sending automatically]
Review every step and flag where a person must stay in charge because of:
- Consequences - errors are costly, public, or hard to reverse.
- Context over information - the right answer depends on history, relationships, politics, or culture that is not in the data.
- Ambiguity - the goal, the facts, or the trade-offs are unclear and need interpretation.
- Stakeholders - the decision needs understanding of people's interests, buy-in, or trust.
- Ethics and fairness - it affects people's jobs, pay, access, or treatment.
- Exceptions - unusual cases that fall outside the normal pattern.
- Accountability - someone has to be able to explain and defend the decision.
Return:
1. A table: Step | Judgment needed? (yes / partly / no) | Reason (from the list above) | What AI can still do to help | What the human decides.
2. The decisions I should not delegate to AI at all, ranked by consequence.
3. A list of exception triggers - signals that a case needs a person, written so a team could spot them.
4. A one-paragraph rule of thumb for this task that I can share with my team.
Be concrete to my task. Where a step is a close call, say so and explain both sides.
Highlighted [brackets] are blanks to replace with your own details.
Fill in the blanks and use this promptMore Working with AI prompts
What to fill in
- DESCRIBE IT, or PASTE the steps
- customers, employees, partners, the public
- money, reputation, legal exposure, relationships, safety
How this prompt is built
This prompt follows the RTF structure (Role, Task, Format). RTF tells the model who to be, what to do and what shape the answer should take. It is the quickest structure that still gives noticeably better results than a bare request.
Related prompts
- Check whether a task is appropriate for AI before I startA go/no-go screen that says, step by step, whether AI should do the work, assist you, or stay out, and why.
- Choose between a chatbot, copilot, automation, or AI agent for a taskRecommends the right kind of AI setup for a task, with the trade-offs, risks, and what you need to make it work.
- Create a quality-control checklist for AI output on a recurring taskA reusable checklist your team runs on every AI output for one recurring task, with pass/fail criteria and spot-check rules.
- Spot where AI is creating more work instead of more valueFinds where AI use is adding rework, review load, content volume, or noise, and what to change so it creates real value.