Prompting from a Position
Prompting is often presented as a skill of wording.
But effective prompting is not primarily about wording.
It is about thinking.
Specifically, it is about starting from a position.
The Core Shift
Most prompting advice focuses on:
- how to phrase questions
- how to structure instructions
- how to use keywords
These can help.
But they are secondary.
The primary driver of prompt quality is this:
What you bring before you write the prompt.
Prompting Without a Position
When no position is defined, prompts tend to be vague.
Examples:
- “Explain machine learning”
- “Write a business plan”
- “Analyze this dataset”
- “Give me ideas”
These prompts:
- lack direction
- produce generic outputs
- do not reflect your thinking
Even if the wording is improved, the result often remains shallow.
Prompting From a Position
A stronger approach is to start with:
- a problem
- a direction
- or a hypothesis
Then use AI to extend that thinking.
Examples:
- “I want to understand which model fits this problem and why.”
- “I am building a small consulting service and need a realistic structure.”
- “I want to test whether variable X is associated with outcome Y.”
- “I have three possible directions. Help me compare them.”
These prompts:
- carry intent
- guide the response
- produce more useful outputs
The Structure of a Strong Prompt
A strong prompt often contains:
Context What are you working on?
Position What do you currently think or want?
Task What should AI help you do?
Example:
“I am building a guide on human-first AI use.
My goal is to make it practical and defensible.
Help me structure the key sections clearly.”
Why This Works
AI responds to signals in the prompt.
A position provides strong signals:
- what matters
- what direction to follow
- what level of detail is expected
Without these signals, AI defaults to general patterns.
With them, AI becomes more targeted.
Improving a Weak Prompt
Weak
“Help me write about AI.”
Improved
“I am writing a guide on responsible AI use and want to explain how to start with a position. Help me structure this clearly.”
The improvement comes from:
- adding context
- stating intent
- defining direction
Prompting as a Thinking Process
Prompting should not be a one-step action.
It is part of the Human → AI → Human loop.
You may:
- Write a prompt from your position
- Review the output
- Refine your prompt
- Ask a more focused question
Each step improves clarity.
Common Mistake: Over-Engineering Prompts
Some users focus too much on:
- complex instructions
- long templates
- rigid formats
These can help in specific cases.
But without a clear position, they do not solve the core problem.
A simple prompt with a strong position is often more effective than a complex prompt without one.
Practical Pattern
You can use this simple pattern:
“I am working on [context].
I think / I want [position].
Help me [task].”
Example:
“I am analyzing customer data.
I think churn may be driven by usage patterns.
Help me identify how to test this.”
Prompting for Different Goals
Exploration
“I am exploring possible causes of this pattern. What should I consider?”
Refinement
“Here is my current explanation. What is unclear or weak?”
Structuring
“Help me organize this into a clear workflow.”
Challenge
“Challenge this reasoning and identify gaps.”
Ownership Still Applies
Even with strong prompts, the responsibility remains yours.
AI can:
- suggest
- structure
- expand
But it does not:
- verify reality
- take responsibility
- make final decisions
That is your role.
Key Insight
Prompting is not about asking better questions alone.
It is about starting from better thinking.
Better prompts begin with a position.
Takeaway
- Do not start with empty requests
- Start with a clear position
- Use prompts to extend your thinking
This leads to:
- more relevant outputs
- clearer reasoning
- stronger decisions