How to Write Prompts AI Can Actually Use
Who this is for
This article is for garden centre and nursery owners, general managers, and senior decision makers who read week 1 and want to move from understanding AI to using it more effectively. If last week helped explain why AI can feel vague or disappointing, this week is about learning how to give it clearer instructions. The aim is not to become technical. It is to get more useful results from the tool you have already chosen, whether that is ChatGPT, Gemini, Claude, or another similar AI assistant.
Key Takeaways
Prompting is a skill you can improve, not a trick that only experts understand.
AI responds better when your prompt includes a clear role, the right context, and a specific job.
The AIM framework, Actor, Input, Mission, gives you a simple way to build stronger prompts.
Weak prompts often miss audience, tone, detail, or a clear outcome.
A few small changes can make AI outputs far more usable.
The idea in plain English
Last week, we looked at why AI can feel disappointing when people expect it to fill in missing context on its own. This week builds on that idea.
If AI is a prediction system, then wording matters. The way you ask shapes the answer you get back.
This is where the idea of prompting becomes important. Prompting simply means writing clear, deliberate instructions for an AI tool so it has a better chance of giving you a useful result. In everyday life, people often leave things unsaid because another person can usually infer the rest. A colleague can read your tone. A member of staff already knows your customer base. A regular supplier understands the season you are talking about without much explanation.
AI is different. It works much better when you spell out what matters.
That means prompting is not about clever tricks or magic wording. It is about being clearer than you would usually be with another person. In most cases, a good prompt tells the AI three things:
who it should act as
what background or material it should use
what exact job it needs to do
That is why the AIM framework works so well, and it is worth reading on because it gives you a simple gem you can start using straight away to get better results from AI.
The AIM framework for clearer prompts
AIM stands for Actor, Input, Mission. It gives you a simple structure for writing prompts that are easier for AI to interpret.
Actor
Actor means the role you want the AI to take.
This does not need to be grand or dramatic. You are simply telling the tool what kind of help you want. That might be a customer service assistant, a product copywriter, a meeting note organiser, or a marketing helper for a UK garden centre.
Without this, the tool often defaults to broad, generic wording. With it, the output usually becomes more relevant.
Input
Input is the context, information, or material the AI needs to do the job well.
This might include your draft text, a list of product features, a rough set of handwritten notes, your audience, your opening hours, your tone of voice, or key details that must be included.
This is the part many people rush past. However, it often makes the biggest difference. If the AI has thin input, it will usually give you a thin answer.
Mission
Mission is the exact task you want completed.
Tell the tool what to produce, how long it should be, what tone to use, and what good looks like. Ask for three options if you want choice. Set a word count if you need something concise. Say whether the result is for customers, staff, or the website.
A vague mission leads to a vague result. A clear mission gives the tool something useful to aim for.
What AIM looks like in practice
It is easier to see the value of AIM when you compare a weak prompt with stronger versions of the same request.
Prompt 1 without AIM Write something for Facebook about our herb plants.
Example response Bring fresh flavour to your garden with our lovely herb plants. We have a great selection available now, perfect for adding colour and fragrance to your outdoor space. Visit us today and find the right plants for your home and garden.
This is the sort of answer many people get when they first try AI. It is not completely wrong, but it is broad, slightly bland, and could have come from almost any retailer.
Prompt 2 with AIM applied You are helping a UK garden centre write a Facebook post. We have a fresh range of herb plants in store, including basil, mint, rosemary, and thyme. The audience is everyday customers who enjoy cooking and want something easy to grow at home. Keep the tone friendly and practical. Write 2 options under 70 words.
Example response Fresh herb plants are now in store, including basil, mint, rosemary, and thyme. If you fancy something easy to grow for the kitchen windowsill, these are a lovely place to start. Pop in and have a look next time you visit.
This is already better because the AI knows the platform, the audience, the product details, and the tone.
Prompt 3 with AIM applied more fully You are helping a local UK garden centre write a Facebook post for weekend shoppers. We have a fresh range of herb plants in store, including basil, mint, rosemary, and thyme. The audience is everyday customers who enjoy cooking, want something easy to grow, and may only have a patio, doorstep, or kitchen windowsill. Keep the tone warm, simple, and practical. Mention that herb plants are a good choice for beginners and that staff can help customers choose the best spot. Write 3 options, each under 80 words, and avoid generic sales language.
Example response Looking for something simple to grow at home? Our fresh herb plants are now in store, including basil, mint, rosemary, and thyme. They are a great choice for patios, doorsteps, and kitchen windowsills, especially if you are just getting started. If you are unsure which one would suit your space, our team will be happy to help.
This version is much stronger because AIM gives the tool a clearer role, richer context, and a more precise mission. The wording feels more natural, more useful, and more like something a real garden centre could post with only light editing.
Why human-style prompts often fall flat
People often write prompts as if they are speaking to another person who already knows the setting.
That is understandable. It is how we communicate every day. We use shorthand, jump between ideas, and assume the other person will fill in the blanks.
With AI, that usually leads to weaker output.
Take a prompt like this:
Write something about our new roses.
There are several gaps straight away. Is it for a Facebook post, an email, a website page, or an in store sign. Are the roses bare root or potted. Is the customer a keen gardener or someone buying a gift. Should the tone sound practical, warm, or promotional. How long should the answer be.
The tool has to guess all of that. Sometimes the guess will be passable. Quite often, it will feel average.
A clearer version would be:
You are helping a UK garden centre write a short email section for customers interested in rose planting. We have a new range of potted roses in store, including climbing, shrub, and patio varieties. Keep the tone friendly and practical. Write 3 versions under 90 words and mention that staff can help customers choose the right spot and variety.
That prompt gives the AI a role, useful context, and a clear outcome. It also makes the audience clear. As a result, the output is more likely to sound relevant and usable.
What makes a prompt easier for AI to interpret
A prompt becomes easier for AI to interpret when it removes avoidable guesswork.
In practice, that usually means being clear about five things.
The audience
Who is this for. A new gardener needs different wording from a regular horticultural customer. A team briefing needs a different tone from a customer email.
The task
What exactly do you want. Write, rewrite, summarise, compare, tidy, simplify, organise, or suggest are all different jobs.
The format
Should the answer come back as bullet points, short paragraphs, a table, a list of subject lines, or three options to choose from.
The constraints
What limits matter. Word count, tone, reading level, style, key facts to include, and things to avoid all help narrow the task.
The context
What does the tool need to know before it starts. The season, the product type, the customer type, the business setting, and any supporting notes all help guide the result.
Common prompt mistakes that weaken the result
Most poor prompts are not poor because the user has done anything foolish. They are poor because people are busy and the first instruction that comes to mind is usually too broad.
Here are some of the most common issues.
Vague verbs
Words like write, help, improve, or fix are too open on their own.
If you ask the tool to “improve this”, it does not know whether you want clearer wording, a stronger sales angle, fewer words, better grammar, or a friendlier tone.
Missing audience
If the AI does not know who the reader is, it often chooses safe middle-of-the-road language.
That is one reason outputs can sound too formal for everyday customers or too simple for an internal planning note.
Unclear outcome
Many prompts do not say what the final answer should look like.
Do you want one finished version, three options, a list of headlines, a summary with actions, or a tidy rewrite that keeps the original meaning. The tool needs that steer.
Too much in one go
AI can help with complex work, but a prompt that asks for a strategy, a sales message, a team briefing, and a social post all at once is likely to come back muddled.
It is usually better to break bigger jobs into steps.
No constraints
Without any limits, AI often wanders into generic wording.
A short instruction like “keep it under 70 words”, “avoid technical jargon”, or “use plain language for beginners” can improve the result straight away.
Weak prompts rewritten into stronger ones
The easiest way to improve prompting is to look at weak prompts and rewrite them with AIM.
Customer-facing example
Weak prompt: Write a sign for our compost offer.
Stronger prompt: You are helping a UK garden centre write a short in store sign for a compost offer. The audience is everyday gardeners looking for spring planting supplies. Keep the tone clear and friendly. Mention that the offer applies to selected lines only. Give 5 options, each under 12 words.
Why it is stronger: the task, audience, tone, and limits are all clear, so the tool does not have to guess what kind of sign you need.
Internal example
Weak prompt: Sort out these notes.
Stronger prompt: You are helping a garden centre manager tidy up meeting notes after a staff planning session. Turn these notes into three sections: decisions made, actions agreed, and questions to follow up. Keep the wording plain and do not add anything that is not in the notes. Here are the notes: [paste notes or attach photo].
Why it is stronger: it tells the AI what kind of material it is working with, how to structure the answer, and what not to do.
Customer service example
Weak prompt: Reply to this customer.
Stronger prompt: You are helping a UK garden centre reply to a customer asking whether our café has gluten free cake options this weekend. Use a warm and helpful tone. Keep the reply short and clear. Include this information: [paste details]. Write 2 versions.
Why it is stronger: the AI knows the exact subject, the tone, the length, and the number of options needed.
Website copy example
Weak prompt: Write about our lavender.
Stronger prompt: You are writing a short website description for a UK garden centre. The product is a range of lavender plants for customers who want colour and scent in sunny spots. Keep the wording simple, practical, and customer friendly. Mention [key details]. Write 3 options between 40 and 60 words.
Why it is stronger: it replaces a broad request with a clear purpose and a defined customer use.
Where it helps in a garden centre
Better prompting helps most when the job is clear but the wording still needs shaping.
Marketing and events
Prompts become more useful when they include the audience, season, and output type.
You might ask for three subject lines for a summer holiday club email, a short website paragraph for a grow your own weekend, or a Facebook post for a pottery restock. Clear prompts help the drafts come back closer to what you actually need.
Customer service messages
AI can draft cleaner replies when the request, tone, and facts are provided.
This works well for routine questions such as refund wording, workshop booking queries, or whether dogs are allowed in certain areas. The clearer the brief, the less reworking the reply tends to need.
Internal organisation
Prompting also helps behind the scenes when notes need sorting or messages need tightening.
That could be turning a messy order update into a clear team summary, rewriting a rough staff notice into plain English, or organising a supplier conversation into actions and follow up points.
Product and display wording
A prompt with the right customer context can improve product descriptions, POS wording, and simple category introductions.
This is especially useful when the first draft sounds too technical, too stiff, or too generic for a garden centre setting.
Try this in 10 minutes
This week, use the tool you picked in week 1, or the one you already use most often. Stick with it. The point of this exercise is to improve how you prompt, not to compare platforms.
Mini task: turn one weak prompt into a stronger one
Choose a real task from your week, such as:
a customer email reply
a short bit of website copy
an internal summary
a sign or social post
Then follow these steps:
Write the quick version you would normally type.
Look at it and identify what is missing.
Who should the AI act as
What information does it need
What exactly should it produce
Rewrite the prompt using AIM.
Actor: who is it helping as
Input: what facts, notes, or context does it need
Mission: what should it create, in what format, and with what tone or limits
Run both versions and compare the results.
Keep the stronger prompt in a note on your phone or computer so you can reuse the structure later.
The real lesson is not the individual task. It is learning to notice what a prompt is missing before you send it.
Saveable tip sheet
Stick with one AI tool while you build confidence.
Use AIM: Actor, Input, Mission.
Always say who the audience is.
Use specific verbs such as rewrite, summarise, compare, or simplify.
Tell the tool what the final answer should look like.
Add useful limits such as length, tone, format, or things to avoid.
Break bigger jobs into smaller steps.
Ask for 2 or 3 options when you want choice.
Paste in source material when accuracy matters.
Save strong prompts so you can reuse the structure.
Template prompt pack
Rewrite customer copy more clearly
You are helping a UK garden centre improve customer wording. Audience: [customer type]. Rewrite this so it sounds [your tone], uses plain language, and stays under [word count]. Keep the meaning the same. Text: [paste text].Draft a short event paragraph
You are writing a website paragraph for a UK garden centre event. Audience: [families / local customers / regular gardeners]. Event: [event name]. Include [date, time, booking details]. Keep it welcoming, practical, and under [word count]. Write 3 options.Turn notes into actions
You are helping a garden centre manager tidy up notes after [meeting type]. Turn these into headings for decisions, actions, and follow up. Keep the wording clear and do not add anything not in the notes. Notes: [paste notes or attach photo].Write short in store signage
You are writing in store signage for a UK garden centre. Audience: [customer type]. Create 5 short sign options for [product / offer / department]. Keep them clear, friendly, and easy to read quickly. Mention [key detail if needed].Draft a customer reply
You are helping a UK garden centre reply to a customer asking about [topic]. Use a warm and helpful tone. Include these details: [paste details]. Keep it concise and write 2 versions.Create website product copy
You are writing a short website description for [product range]. Audience: [target customer]. Keep it practical and easy to understand. Mention [key features]. Write 3 versions between [word count range].Improve a weak prompt
Here is the prompt I was about to use: [paste prompt]. Before rewriting it, ask me up to 5 short questions, one at a time, so you can understand what is missing. Help me clarify the role, context, audience, and task. Then write a stronger final prompt using AIM and briefly explain why it is better than my original version.Simplify something technical
You are helping a UK garden centre explain [topic] to everyday customers. Rewrite this in plain English, keep it friendly, and avoid jargon. Text: [paste text].
What’s next
Week 1 explained why AI can feel disappointing when the instruction is thin. Week 2 is about giving it a clearer steer. Once you start using simple structures like AIM, prompting becomes much less mysterious and much more practical. In the next article, we can build on this again by looking at how to refine, check, and improve AI outputs once the first draft comes back.