top of page

How are you prompting AI?


A field note on using AI for more than speed


I use AI every day. For work, planning, thinking, even for my journaling.


Over time, that has led me into a familiar pattern. I find, save, test and tweak prompts. Sometimes I go looking for them on purpose. Sometimes I stumble across them while searching for better ways to think something through. Some are useful immediately. Some are clever but thin. Some work once and then never again. Some seem impressive until you realise they are mostly just better packaging for generic advice.



A lot of prompt-sharing online is built around output. Write this faster. Improve that sentence. Suggest a metaphor here, and there’s nothing wrong with that. But sometimes it feels like owning a Ferrari and using it to deliver groceries.


Lately, I’ve been noticing a different category: derived prompts, and I think there is something in it. By derived prompts, I mean prompts that do not begin life as “AI prompts” at all. They come from somewhere older and sturdier. They are drawn from critical thinking, scenario planning, decision science, systems thinking, reflective practice, coaching, and other disciplines that were helping people think more clearly long before language models arrived.


In other words, they are not really invented for AI. They are translated for AI.


So with the help of my usual LLM, here are eight derived prompts, not lifted from prompt blogs, but grounded in established thinking disciplines and translated into reusable prompt form. Each one works across strategy, coaching, writing, and decision-making. I’ve kept them clean so you can drop them straight into use.


1. The Skeptical Reviewer

Use for: ideas, proposals, content, strategy

“Act as a critical reviewer. Identify assumptions, weak logic, and overgeneralizations. For each, explain the risk of being wrong and how to strengthen it.”

Underlying idea: critical thinking / red teaming


2. The Premortem

Use for: projects, strategy, decisions

“Assume this failed 12 months from now. What went wrong? List the most likely causes of failure and early warning signs.”

Underlying idea: premortem (Gary Klein)


3. The Inversion Prompt

Use for: performance, habits, business strategy

“Instead of telling me how to succeed, list how this would fail. Then invert those into actions to avoid.”

Underlying idea: inversion (Charlie Munger)


4. The Future Historian

Use for: long-term thinking, leadership, positioning

“You are writing 5 years in the future. Looking back, what did we misunderstand? What turned out to matter more or less than expected?”

Underlying idea: scenario planning / back casting


5. The Constraint Amplifier

Use for: creativity, problem-solving, innovation

“Given these constraints [X], what are 5 ways to achieve the goal anyway? Prioritize unconventional but viable approaches.”

Underlying idea: creative constraint (design thinking)


6. The Trade-off Clarifier

Use for: decisions, prioritization, strategy

“What are the key trade-offs in this situation? What do I gain and what do I give up with each option?”

Underlying idea: opportunity cost / decision theory


7. The First Principles Breakdown

Use for: complex problems, learning, strategy

“Break this problem down into first principles. What do we know for certain, what are assumptions, and what follows logically?”

Underlying idea: first principles thinking


8. The Second-Order Effects Prompt

Use for: strategy, policy, leadership decisions

“If we take this action, what are the second- and third-order consequences? Who is affected and how might the system respond over time?”

Underlying idea: systems thinking


What ties these together

These prompts all:

  • slow thinking down

  • expose hidden structure

  • force better judgement


They’re less about output, more about quality of thought.

 
 
 

Recent Posts

See All
JournAIing

My reflective practice is ever evolving and this is another post on that evolution. I’ve written elsewhere about the mechanics of journaling ( see Journaling, Upgraded ), you know things like sitting

 
 
 

Comments


bottom of page