Role Prompting and Prompt Length: What Actually Helps
Does telling AI to act as an expert help? How long should a prompt be, and how many questions can you ask at once? A practical guide to what works.
AI tools can answer questions, write content, analyse information, brainstorm ideas and help with everyday tasks. But sometimes the response you get isn’t quite what you expected.
You may ask AI a simple question and receive an answer that is too vague, too detailed, too technical or simply focused on the wrong thing.
Often, the problem isn’t the AI tool itself. It is the way the request was written.
Learning how to write a really good AI prompt can help you communicate your request more clearly and get more useful results from tools such as ChatGPT, Claude and Gemini.
This is a beginner’s guide. It explains what a prompt is and covers the core habits. When you are ready to apply them to business tasks, see our guide to writing effective AI prompts.
What Is an AI Prompt?
An AI prompt is the instruction, question or information you give to an artificial intelligence system to tell it what you want it to do.
A prompt could be as simple as:
“Explain how solar panels work.”
Or it could include more information:
“Explain how solar panels work in simple terms. Break the explanation into five steps and include an everyday example.”
Both are AI prompts. The difference is that the second gives the AI more direction about the response you want.
A prompt doesn’t have to be complicated. The goal is simply to communicate your intended task clearly.
What makes a good prompt?
A good AI prompt gives the AI enough information to understand your request and produce a useful response.
In most cases, an effective prompt includes some combination of:
- A clear goal
- Relevant context
- Specific instructions
- The desired format or style
- Useful limitations or requirements
You don’t always need all of these. A straightforward question may only need one sentence, while a more complicated task may benefit from considerably more detail.
The key is to provide the right information, rather than simply more information.
How to Write a Really Good AI Prompt
- Start with a clear goal
- Give relevant context
- Be specific
- Say how to present the answer
- Add useful constraints
- Use examples
- Break complex requests into steps
- Review and refine
1. Start With a Clear Goal
The first step is to tell the AI exactly what you want it to do.
A vague request can leave too much room for interpretation.
For example:
“Tell me about photography.”
This could produce almost anything: a history of photography, an explanation of cameras, tips for taking better photographs or information about famous photographers.
A clearer request would be:
“Explain the three basic camera settings that beginners should understand.”
The second prompt gives the AI a specific task and a clear scope.
Before writing your prompt, ask yourself:
What exactly do I want the AI to do?
2. Give the AI Relevant Context
The AI only knows what you tell it. If the best answer depends on your situation, say so.
Compare:
“Suggest a camera to buy.”
with:
“Suggest a camera for a complete beginner who wants to photograph family holidays and has a budget of around £500.”
The second prompt gives the AI the background it needs to choose well. It does not need your life story, only the details that affect the answer.
Before writing your prompt, ask yourself:
What does the AI need to know to do this properly?
3. Be Specific About What You Want
Specificity is one of the simplest ways to improve an AI prompt.
Compare:
“Give me dinner ideas.”
with:
“Give me five vegetarian dinner ideas that can be prepared in under 30 minutes.”
The second request gives the AI useful boundaries.
You can make a prompt more specific by mentioning:
- The number of results you want
- A particular topic
- A time period
- A level of difficulty
- A target audience
- Important requirements
- Things you want included or excluded
The more clearly you define the outcome, the less the AI has to guess about what you mean.
4. Tell AI How You Want the Answer Presented
You can also specify the format of the response.
For example:
“Explain the water cycle in five simple steps.”
Or:
“Summarise this article in five bullet points.”
Or:
“Compare these two options in a table showing their main advantages and disadvantages.”
The underlying information may be the same, but specifying the format can make the response much easier to read and use.
If you have a particular structure in mind, tell the AI.
5. Add Useful Constraints
Constraints tell the AI about boundaries or requirements for the response.
You might ask it to:
- Keep the answer under 500 words
- Use plain English
- Avoid technical terminology
- Provide three examples
- Write for beginners
- Use bullet points
- Focus only on the most important information
For example:
“Explain how inflation works in under 300 words using plain English and one everyday example.”
However, don’t add constraints simply because you think every prompt needs them.
If you ask:
“What is photosynthesis?”
you probably don’t need to specify ten different requirements.
A good prompt is not necessarily a long prompt.
6. Use Examples When You Need a Particular Result
Sometimes it is easier to show AI what you want than to describe it.
For example, if you want AI to categorise information in a particular way, you can provide an example of the desired format.
This is sometimes referred to as few-shot prompting when examples are provided to demonstrate a pattern or task.
Examples can be particularly useful when you need a specific:
- Structure
- Style
- Classification
- Pattern
- Type of response
But again, examples aren’t necessary for every prompt. Use them when they help communicate something that might otherwise be difficult to explain.
7. Break Complex Requests Into Smaller Steps
If you are asking AI to complete several related tasks at once, breaking the request into stages can make it easier to follow.
Instead of:
“Teach me everything about photography.”
Try:
“Explain the three most important camera settings for beginners. Start with aperture, then shutter speed and ISO. For each one, explain what it controls and give a simple example.”
The second request establishes a clear sequence and limits the scope.
For complicated tasks, you can also ask AI to complete one part before moving on to the next.
8. Review the Response and Refine Your Prompt
Writing a good AI prompt isn’t always a one-time process.
If the first response isn’t quite right, you can tell the AI what needs to change.
For example:
“Make the explanation simpler.”
“Give me two more examples.”
“Focus only on the most important points.”
“Rewrite this for someone with no previous knowledge of the subject.”
This is one of the most useful things to understand about AI prompting: you don’t always need to start again from scratch.
Treat the interaction as a conversation. Review the response, identify what is missing and provide another instruction.
The Most Common Beginner Mistakes
Most weak prompts share a few problems: they are too vague, they leave out important context, they ask for conflicting things or they trust the answer without checking it. Our guide to common AI prompting mistakes explains each one and how to fix it.
A Simple Formula for Better AI Prompts
If you’re unsure where to start, remember this simple framework:
Goal + Context + Specifics + Desired Output
For example:
“Explain how mortgages work for someone buying their first home. Use simple language, cover the main costs involved and finish with five key points to remember.”
The goal is to explain mortgages.
The context is that the reader is a first-time buyer.
The specifics are the main costs.
The desired output is five key points at the end.
You don’t need to use this formula every time you interact with AI. Think of it as a checklist you can use when a simple prompt isn’t producing the result you want.
Do Longer AI Prompts Give Better Results?
Not necessarily.
A longer prompt can be useful when a task requires additional context, examples or specific requirements. But adding unnecessary information doesn’t automatically make a prompt better.
The objective is to give AI the information it needs to understand the task.
Be clear and specific, not unnecessarily complicated.
Get More From AI With Maximus Mediascape
Good prompting is a useful starting point for working with AI, but businesses can also use AI for much more than generating individual responses.
Maximus Mediascape helps businesses explore practical applications of AI automation, chatbots, lead qualification, customer support and automated workflows.
If you’re looking to move from simply experimenting with AI to using it more effectively across your digital operations, Maximus Mediascape can help you identify opportunities to put AI and automation to work.
Related guides: How to Write Effective AI Prompts: A Business Checklist, 10 AI Prompting Techniques for Better Responses Every Time, 10 Common AI Prompting Mistakes and How to Fix Them, Role Prompting and Prompt Length: What Actually Helps.
