A powerful way to get better answers from AI is to name a person, a framework or a mental model in your prompt. This guide shows how to pick the right one for more ideas, deeper answers or a strategic view.
A powerful way to get better answers from AI is to name a person, a framework or a mental model in your prompt. This guide shows how to pick the right one for more ideas, deeper answers or a strategic view.
One question, three lenses, three different answers.
TL;DR
Naming a person, a framework or a mental model in your prompt makes AI answer with the questions, words and structure that belong to it.
Pick the lens by what you need. How might we and SCAMPER give more ideas, first principles and five whys go deeper, and SWOT and a pre-mortem give the strategic view.
Start by asking AI to suggest a lens for your challenge, then let it use the one it recommends.
Most prompts ask a plain question. You will get an answer, but it is the
average of all the knowledge the model has trained on. It sounds like everyone
else's.
Name a person, a framework or a mental model in the prompt. The answer then
gets more ideas, more depth or a wider view. The prompts work in
ChatGPT, Claude, Gemini and Copilot.
The examples come from a fictional company we call Company A. It has 400
people in Norway and is deciding whether to expand to Sweden next year.
How a lens steers the answer
AI has read about almost everything, from recipes to tax rules. Most of it is
of no use for your question.
When you name a role such as a CFO, the model brings forward what belongs to
that role. It works the same with the ideas of a person. Charlie Munger was
Warren Buffett's business partner for decades and was known for asking who
gains from each choice.
Naming a role or a person's ideas brings the matching questions and habits to the front.
A framework such as SWOT gives the answer a structure. A mental model is a
rule of thumb for thinking, and it steers the answer too. A lens does not
make the model know more. It changes the angle, the words and the shape of the
answer.
People as lenses
Borrow the view of a person who would see the question differently. It could
be a CFO, a sceptical customer or a well-known thinker. Each one brings new
questions to the same decision.
Three people at the same decision, each with the question they would ask first.
Ask for the ideas a person is known for, not for what they would say. The
answer is the model's reading of those ideas. Treat it as a lens and not as a
quote.
You
Look at [decision] through three people, a CFO, a sceptical customer and the ideas Charlie Munger is known for. For each, give the two questions they would ask first and what they would want to see before saying yes.
Pick a lens for your goal
Start from what you want from the answer. The figure shows five lenses for
each of four goals. This guide covers the first two for ideas, depth and
strategy.
Start from what you need, then pick the lens.
You do not need to know the frameworks. Ask AI to suggest some and let it use
the one it thinks fits best.
You
Here is my challenge, [challenge]. Suggest five frameworks, mental models or people whose view would help, say in one line what each would show me, then use the one you recommend.
If you are already in a conversation, you can also ask
You
Suggest 5 frameworks and well-known people I could benefit from, based on what we are talking about.
How might we, for more ideas
"How might we" turns a problem into open questions that invite many answers.
Min Basadur brought it
into use at Procter & Gamble in the 1970s. The design company IDEO later made
it part of design thinking.
Each question leads to different ideas, so you do not settle on the first one.
One yes-or-no question becomes five open How might we questions.
You
Act as a workshop facilitator. Our challenge is [challenge]. Turn it into five "How might we" questions, from small and practical to bold. Pick the one that opens up the most new ideas, and give three ideas for it.
The question Company A picked led to a plan to follow three Norwegian
clients that already have offices in Sweden.
SCAMPER, for new versions of an idea
SCAMPER gives seven ways to change an idea you already have. The letters
stand for substitute, combine, adapt, modify, put to another use, eliminate and
reverse. Bob Eberle published it in 1971 as part of the creative problem solving
method of Alex Osborn, the advertising executive who coined the word brainstorming.
SCAMPER on the Sweden plan, one change per letter.
You
Use SCAMPER on [idea]. Give one concrete change for each of the seven letters, then mark the two that are cheapest to test.
First principles, for deeper answers
First principles means breaking a problem down to what you know is true.
Then you build the answer up again from there instead of copying how others do
it.
Company A first assumed that Sweden needs a local office and a local sales
team. Broken down to what is true, the problem came to four facts about how
Swedish clients buy.
Break it down to what is true, then build it up again.
You
Use first principles thinking on [challenge]. List what we take for granted, break the problem down into what we know is true, and build three new solutions from those facts.
Five whys, for the real reason
Toyota's production system, made known by Taiichi Ohno, asks
why again and again until you
find the real reason. Behind the Sweden plan, the real reason was that
Company A needs new clients in Norway or Sweden.
Ask why until you reach the real reason. Here it turned into a different question.
You
Ask me why we want to [plan], then ask why again, up to five times. Stop when you reach the real reason, and tell me if it points to a different question than the one I started with.
SWOT, with the boxes connected
A SWOT sorts a plan into strengths, weaknesses, opportunities and threats.
On its own it gives four lists that are hard to act on. The useful part is to
connect the boxes and match each strength with an opportunity it can make use
of. The management professor Heinz Weihrich called this a TOWS matrix
in 1982.
Connect a strength to an opportunity, and the SWOT ends in a move.
You
Do a SWOT analysis of [plan] for [company], four points per box. Then connect the boxes. For each strength, name the opportunity it can make use of, and for each threat, the weakness that makes it worse. End with the three moves you would make first.
A pre-mortem, before you decide
The pre-mortem is a
method from the psychologist Gary Klein. Before you decide, you imagine the
plan has already failed and ask why.
You
Take the role of a sceptical customer. Imagine our [plan] failed after six months, and tell us the five most likely reasons. Then say which one we can do something about now.
For Company A, the reason it could act on now was simple. No one had asked
the three Norwegian clients in Sweden whether they would buy there.
Three lenses in one prompt
You can name a person, a framework and a mental model in the same prompt.
Try a CFO, a SWOT and turning the plan upside down to see what would make it
fail. Each one adds its own questions.
Sweden next year
+ New chat
Chats
Projects
Artifacts
Recents
Sweden next year
Take the role of a CFO. Do a SWOT of expanding to Sweden next year, then turn it upside down. What would make it fail within a year?
Our references fit Swedish companies that are like our clients.We fail if Swedish clients see us as a Norwegian firm that does not speak their language.Cap the first year at one hire, and stop if fewer than five clients sign.
Reply to Claude
+Tools ▾↑
A role, a framework and a mental model in one prompt.
Next time you get an answer that sounds like everyone else's, name a lens and
ask again. Once a lens prompt works for you, save it as a
skill so you do not have to type it again.
