Why Does the Way You Ask Matter?
We have now learned what ChatGPT is, how it generates an answer, and how conversation context helps it understand what we mean.
There is one more important part of the interaction:
The way you communicate your request.
Consider these two requests:
"Tell me about AI."
and:
"Explain artificial intelligence to a complete beginner in simple language, using three everyday examples."
Both requests are about artificial intelligence, but the second request provides much more information about what the user wants.
This can help the AI assistant produce a response that is more closely matched to the users goal.
Vague Questions Can Produce Broad Answers
A vague question does not necessarily produce a bad answer.
However, it may leave many possible interpretations open.
For example:
"Explain marketing."
What does the user want?
- A definition?
- A beginner explanation?
- Examples?
- A business strategy?
- A history of marketing?
- Help with a marketing plan?
The AI assistant has to choose an interpretation.
A more specific request reduces the number of possible interpretations.
Be Clear About Your Goal
One of the easiest ways to improve a request is to explain what you are trying to accomplish.
Compare:
"Write about exercise."
with:
"Write a short introduction explaining why regular exercise is important for people who are new to fitness."
The second request gives the AI assistant a clearer goal.
When an AI assistant understands the desired outcome, it has better information about what kind of response to generate.
Tell the AI Who the Answer Is For
The same subject can require very different explanations depending on the audience.
For example:
"Explain neural networks."
could produce a general explanation.
But:
"Explain neural networks to a ten-year-old using a simple everyday example."
provides information about the intended audience.
The response can then be written at a more appropriate level.
You can describe the audience using simple phrases such as:
- "for a complete beginner"
- "for a school student"
- "for a business manager"
- "for a technical audience"
Specify the Level of Detail
Another useful piece of information is how much detail you want.
For example:
"Explain machine learning."
is open-ended.
You could instead ask:
"Explain machine learning in five simple paragraphs."
Or:
"Give me a short explanation of machine learning in three bullet points."
The additional instruction helps communicate the expected size and structure of the response.
Give Useful Context
Context can be especially important when the AI assistant does not know enough about your situation.
Imagine asking:
"Write an email to my manager."
The assistant does not know what the email is about.
A clearer request might be:
"Write a polite email to my manager asking for two days of leave next week."
The second request gives the assistant useful information about the purpose of the email.
Additional details can make the response more relevant.
Examples Can Help
Sometimes the easiest way to show what you want is to provide an example.
Suppose you want an AI assistant to rewrite a sentence in a particular style.
Instead of simply saying:
"Make this professional."
you could provide an example of the style you prefer.
The example gives the system additional information about the desired output.
This idea becomes especially important when we later study prompt engineering.
Tell the AI What to Change
When asking an AI assistant to improve something, identify what you want changed.
For example:
"Improve this paragraph."
is relatively broad.
You could instead say:
"Rewrite this paragraph to make it clearer and easier for a beginner to understand, while keeping the original meaning."
Now the assistant has a better understanding of the requested changes.
Output Format Can Matter
You can also tell the AI assistant how you want the answer organized.
For example:
- "Give me five bullet points."
- "Put the information in a table."
- "Give me a short summary followed by three examples."
- "Write this as a professional email."
- "Explain the answer step by step."
These instructions provide information about the desired form of the response.
The same underlying subject can therefore produce very different outputs depending on the requested format.
Compare Two Requests
Consider these two requests:
Request A:
"Tell me about electric cars."
Request B:
"I am considering buying my first electric car. Explain the main advantages and disadvantages in simple language and give me five important things a beginner should consider."
Request B provides several useful pieces of information:
- The users situation.
- The goal of the request.
- The desired level of explanation.
- The topics to cover.
- The desired number of points.
As a result, the assistant has a much clearer target for the response.
More Detail Is Not Always Better
It might seem that the best request is always the longest request.
That is not true.
A very long request can contain unnecessary information or conflicting instructions.
The goal is not to make every request extremely detailed.
The goal is to provide the information that matters.
A short but precise request can be better than a long and confusing one.
Clear Questions Can Save Time
When a request is unclear, the first response may not match what the user wanted.
The user may then need to correct the assistant:
"No, that is not what I meant."
More specific instructions can reduce this kind of back-and-forth.
For example, instead of:
"Help me with this report."
you could say:
"Summarize this report into five bullet points for a senior manager, focusing on the main risks and recommended actions."
The second request gives the AI assistant a much clearer target.
What If You Do Not Know Exactly What You Want?
You do not always need to write a perfect request.
AI assistants can also help you discover what you need.
For example, you might say:
"I need to learn about cybersecurity, but I do not know where to start. Can you suggest a beginner learning path?"
This is a useful request even though the user does not already know the exact answer they want.
The important thing is to communicate the situation and goal clearly enough for the assistant to help.
Follow-Up Questions Can Improve the Result
You do not have to put everything into one message.
A conversation can be built step by step.
For example:
User: "Explain machine learning to me."
User: "Now give me an everyday example."
User: "Make the explanation suitable for a school student."
User: "Now summarize the main idea in three bullet points."
Conversation context allows these requests to build on each other.
This connects directly with the previous lesson about conversation context.
A Simple Formula for Better Questions
For many everyday AI requests, you can think about five useful pieces of information:
- Goal: What do you want the AI to do?
- Context: What background information does it need?
- Audience: Who is the answer for?
- Format: How should the answer be presented?
- Constraints: Are there important limits such as length, tone, or number of examples?
You do not need to include all five every time.
Use the pieces that are relevant to the task.
Example Using the Formula
Suppose you want help creating a study plan.
A simple request could be:
"Make me a study plan."
A more useful request might be:
"I am a beginner learning artificial intelligence. I can study for 30 minutes each weekday. Create a four-week study plan with one topic per day and include a short practical exercise for each topic."
This request communicates:
- The goal: create a study plan.
- The context: the learner is a beginner in AI.
- The available time: 30 minutes per weekday.
- The duration: four weeks.
- The format: one topic per day.
- The additional requirement: a practical exercise.
The assistant now has much more information about the desired result.
Good Questions Do Not Guarantee Correct Answers
This is an important warning.
A well-written request can improve the relevance and usefulness of an AI response, but it does not guarantee that the response will be factually correct.
As we learned earlier, AI-generated responses can contain errors.
Clear instructions help the model understand what you want.
They do not automatically make the information generated by the model true.
For important information, you should still evaluate and verify the response when appropriate.
The Difference Between Asking Better Questions and Prompt Engineering
In this lesson, we are learning a simple practical skill:
Give the AI assistant enough information to understand what you want.
Later, in the Prompt Engineering module, we will explore this idea much more deeply.
We will study how prompts can be deliberately designed to produce more consistent, useful, and controlled results.
For now, the most important lesson is simply to communicate clearly.
What You Should Remember
- The wording of a request can influence the usefulness and relevance of an AI response.
- Vague requests leave more possible interpretations open.
- Clear goals help an AI assistant understand the desired outcome.
- Audience information can help determine the appropriate level of explanation.
- Useful context can make a response more relevant.
- Examples can communicate the desired style or result.
- Output format instructions can make the response easier to use.
- More detail is not always better; relevant detail is what matters.
- A clear request can reduce unnecessary back-and-forth.
- Good instructions improve communication but do not guarantee factual accuracy.
- Advanced prompt design will be covered later in the Prompt Engineering module.
What Comes Next?
You now know how to make a basic AI request clearer and more useful.
But even when you ask a clear question, ChatGPT can sometimes misunderstand what you mean or make mistakes.
In the next lesson, we will examine common mistakes people make when using ChatGPT and learn how to avoid them.