AI Tools: Fantastic or Frustrating?
We’ve firmly entered the age of AI. Tools like ChatGPT can help you draft emails, summarise documents, brainstorm ideas, and tackle all kinds of everyday work more quickly.
But getting a useful response isn’t always as simple as typing a question and pressing Enter. Sometimes the result is exactly what you need. Other times, it’s too vague, too long, off-target, or oddly focused on the wrong thing.
How can you improve your results when using AI tools? A few simple changes to your prompts (text requests) can make a big difference.
In this article, we’ll look at six practical prompting techniques that help you get clearer, more relevant, and more useful responses from AI chat tools. Whether you're using ChatGPT, Microsoft Copilot, Google Gemini, Claude, or another AI assistant, the core principles are broadly the same.
Whether you’re drafting an email or asking for help with a recipe, these tips will help you get relevant responses from the start.
Be Careful What You Wish for…
Think of using AI as a bit like talking to a genie — your wish is their command, but only if you phrase it clearly. Say, "I wish for a million bucks!" and you might end up with a neighbourhood full of male deer happily chewing through every patch of grass.
The problem is not the genie. It’s the wording.
When using AI, a vague prompt can lead to an answer that misses the mark. A clear, specific request is much more likely to produce a useful, on-target response.
Today’s AI tools are much better at inferring meaning from context than they were even a couple of years ago. But they still perform best when you give them less guesswork to do. The clearer your prompt, the better your result is likely to be.
6 Key Elements of a Quality Prompt
A strong prompt does more than ask a question. It defines the task, removes ambiguity, narrows the scope, and shapes the tone and format of the response.
These six prompt elements are useful building blocks for doing that:
- TASK
- CONTEXT
- EXEMPLARS
- PERSONA
- FORMAT
- TONE
Let's break down each element and understand its role in effective prompt engineering:
1) TASK: Start with an Action Verb
- Purpose: Define the main action you want the AI to perform.
- Examples:
- "Draft a follow-up email to a client who hasn't responded in two weeks."
- "Compare the pros and cons of these two vendor proposals."
- "Summarize the key points from this meeting transcript in five bullet points."
Tip: Begin with a clear action verb to set the AI's focus. The more specific the action, the more useful the response — "draft," "summarize," "compare," and "explain" will get you further than "tell me about" or "help me with."
Chain tasks together in a single prompt: "First summarize this document, then draft a follow-up email based on that summary." Modern AI tools handle multi-step instructions well.
2) CONTEXT: Give the AI Your Real Situation
- Purpose: Provide background that grounds the AI's response in your specific needs rather than a generic answer.
- Examples:
- "I'm a project manager at a mid-sized IT firm. Summarize the key challenges our industry faces with remote onboarding."
- "My audience is non-technical small business owners. Explain what two-factor authentication is and why they need it."
Tip: Don't be afraid to give real context — a job description, a document excerpt, or a brief situation summary. The more relevant detail you provide, the less the AI has to guess.
3) EXEMPLARS: Show the AI What You're Looking For
- Purpose: Give the AI a concrete example of the output style, tone, or format you want — so it matches your vision rather than defaulting to a generic response.
- Examples:
- "Write a project update email in the style of this example: [paste example]."
- "Generate a blog post outline following this structure: Introduction, Main Points, Conclusion."
- "Here's a report I wrote last quarter: [paste excerpt]. Match this tone and format for the new one."
Tip: Rather than describing what you want in words, show it. Paste in a sample of your own writing, a previous AI response you liked, or a template you already use. A real example is worth a hundred adjectives.
4) PERSONA: Set the Perspective
- Purpose: Tell the AI whose perspective to write from, so it can match the response to a particular audience or explore more than one point of view.
- Examples:
- "Explain this as if you're talking to someone without a tech background.”
- "Write this from the two perspectives: 1) a customer 2) a salesperson."
- “Rewrite this for an audience of busy small business owners.”
Tip: Assigning a persona can improve voice and relevance, but it does not make the AI a genuine expert. Important facts and recommendations should still be checked.
5) FORMAT: Tell the AI Exactly What the Output Should Look Like
- Purpose: Specify the desired structure or format of the output.
- Examples:
- "Give me the answer as a table with three columns: Option, Pros, and Cons."
- "Write this as a FAQ — ten questions and concise answers."
- "Summarize this in three sentences, then follow it with five bullet points."
- "Keep the entire response under 200 words."
Tip: Think about where the output is going to land — an email, a slide, a report, a social post — and describe that format explicitly. If you don't specify, the AI will default to flowing paragraphs, which often isn't what you need.
6) TONE: Define the Emotional or Stylistic Feeling
- Purpose: Set the tone or mood of the output to match your objectives.
- Examples:
- "Write this in a warm, reassuring tone — the reader may be anxious about the change."
- "Use a confident, no-nonsense tone. Avoid corporate jargon."
- "This is for a general consumer audience — keep it friendly and conversational, not technical."
Tip: Name the feeling you're going for — formal, casual, empathetic, direct, enthusiastic. If the tone really matters (a sensitive client email, an HR message, a public statement), you can add: "Before you write, confirm your understanding of the tone I'm asking for." It takes one extra exchange but can save several rounds of revision.
Putting the Elements Together
Now that we have the basic anatomy of a quality prompt note the following example:
Why Your Prompts Might Be Ineffective
When an AI response misses the mark, the problem is often the prompt. Three common issues are:
-
Lack of specificity — the prompt is too broad and doesn’t give the AI enough direction.
-
Ambiguity — the prompt uses language that could be interpreted in more than one way.
-
Unclear expectations — the prompt doesn’t explain what kind of response is needed.
Example 1: Lack of Specificity
Ineffective prompt:
“Tell me about elephants.”
Issue:
This prompt is too broad and gives no indication of what kind of information is desired and for what purpose.
Improved prompt:
“Tell me about the social behaviour of elephants in the wild for a Grade 9 class presentation. The presenter and audience are both at Grade 9 level.”
Why it’s better:
It narrows the topic, makes the request more specific, and gives the AI a clearer sense of the knowledge level it should aim for.
Example 2: Ambiguity
Ineffective prompt:
“What is the best way to get ahead?”
Issue:
“Get ahead” could mean advancing in a career, succeeding in school, or even winning a race.
Improved prompt:
“List the most effective strategies for getting ahead in a corporate career.”
Why it’s better:
It clarifies what “get ahead” means in this context.
Example 3: Unclear Expectations
Ineffective prompt:
“Write an email about the server outage.”
Issue:
This prompt doesn’t say who the email is for, how long it should be, what tone to use, or what details to include.
Improved prompt:
“Write a 150-word client-facing email in a calm, professional tone explaining that our servers experienced unexpected downtime this morning from 9 to 11 am, that the issue is now resolved, and that we apologise for any disruption. The audience is non-technical business owners.”
Why it’s better:
It sets clear expectations for the audience, tone, length, purpose, and key facts
When Your Prompt Still Isn’t Working
Even well-constructed prompts sometimes miss the mark. When that happens, try refining the conversation instead of starting over:
- Ask the AI to ask questions first: “Before you respond, ask me any clarifying questions you need.” This is a simple but often overlooked way to get a more useful answer.
- Tell it what missed the mark: “That response was too formal and too long. Try again in a more conversational tone and keep it under 150 words.” You don’t need to start a new chat to improve the result.
- Break the task into steps: Complex requests often work better as a back-and-forth exchange than as one large prompt.
- Try a different tool: Different AI assistants have different strengths. If one gives a flat or unhelpful answer, another may approach the same task in a more useful way.
Become a Prompt Genius
And there you have it — your guide to taming the AI genie.
Good prompting is less about perfection and more about clarity. When you define the task, provide the right context, and set clear expectations, AI tools are likely to give you useful results on the first try. And if the first response misses the mark, you can refine it without starting from scratch.
Whether you’re drafting an email, summarising a meeting, creating a report, or brainstorming ideas, better prompts will help you get faster, more relevant, and more useful answers from any AI chat tool.