AI.
28 October 2026
WRITTEN BY
KATRINA PACE
6 MINUTE READ
SUMMARY
Give AI the background, permission to challenge you, clear writing guidance, room to explore and a push to get specific. You’ll have much more to work with than a vague request for “engaging content”.
RELATED SERVICES
When AI started writing content, I knew pretty quickly that ignoring it wasn’t going to be a sensible business strategy.
I’m a copywriter. Something that could produce a blog in seconds was very much stepping onto my patch.
So I decided early on that I had to use it. And if I was going to keep bringing value to my clients, I needed to get better at using it than someone typing “write me a blog” and hoping for the best.
That’s meant a heap of time playing, questioning, challenging and occasionally wondering how we’d wandered so far away from what I’d asked for.
Along the way, I’ve learnt a lot about giving AI what it needs to produce something useful.
These are five prompts I keep coming back to, plus a bonus that might surprise you.
First, why I think of AI as “Augmented Intelligence”
I find augmented intelligence a much more useful way to think about AI than “artificial intelligence”.
It can extend my thinking. Help me explore an idea. Suggest an angle I haven’t considered. Give me something to pull apart and improve.
It’s my second in command. I’m still responsible for what goes out.
With copywriting, I have the experience to assess what it gives me. I can spot when the message is muddy, the benefit is missing or a lovely-sounding paragraph isn’t actually saying anything.
Taxes? Different story.
I don’t know enough to reliably spot what it’s got wrong. And a confident answer wouldn’t suddenly give me that knowledge.
That distinction matters. Better prompting helps you work with AI, but you still need the judgement to decide whether its answer is any good.
Prompt 1. “This is what we’re doing today…”
AI needs a brief. And “write a blog about my business” leaves an awful lot of room for interpretation.
When I start a piece of content, I explain what we’re creating, who it’s for and what I want it to achieve. I’ll also give it the relevant background: what the reader might be struggling with, what they already know and where this content fits.
For example:
This is what we’re doing today: writing a blog for service-based business owners who use AI for their marketing but spend ages editing it. They know their subject, but struggle to make the writing sound like them. I want to give them practical help and introduce my AI editing service as an option if they’d like support.
Now there’s a direction.
We’re helping a particular person with a particular problem. That gives us something to assess the draft against.
Think about briefing someone who’s helping in your business. You wouldn’t say “make a thing” and expect them to know everything you had in mind.
Give AI the background you’d give a capable person who’s new to the job.
Prompt 2. “You’re allowed to disagree with me if you can justify why your opinion is better”
A string of enthusiastic agreements can feel quite lovely.
But I’m using AI to help me think, and sometimes that means questioning my thinking.
If I suggest an angle that’s weak, I want it to explain why. If I’ve overlooked something my reader needs, I want it to point that out.
So I give it permission to challenge me:
You’re allowed to disagree with me if you can justify why your suggestion would work better for this audience and goal. If I’ve missed something, point it out and explain your reasoning. Don’t disagree just for the sake of it.
That last bit matters. We don’t need to turn every heading into a debate.
It’s easy to work inside a bubble of what we already know and believe. If everything starts with my assumptions and ends with AI agreeing with them, I haven’t necessarily moved my thinking forward.
I want alternatives I can weigh up. I might still stick with my original idea, but I want a good reason for doing so.
Ask for a challenge, then use your own judgement to assess it.
Prompt 3. “This is how I want to sound…”
“It doesn’t sound like me” is one of the things I hear most often from clients who’ve tried writing with AI.
And I get it. You read the draft and think, “Well, the information’s there. But I would never say it like that.”
This is where it helps to get specific about voice and writing style.
“Friendly and professional” is a starting point, but there are lots of ways to interpret those words. Does friendly mean warm and reassuring? Chatty and bubbly? Informal with a bit of cheek?
Try something like:
This is how I want to sound: warm, chatty and confident, like I’m explaining something to a client over coffee. Use everyday language, contractions and practical examples. Vary the sentence length. Let some personality through, but don’t add forced jokes or over-the-top enthusiasm. Here are two examples of my writing to use as a guide.
The examples are important. They show what those instructions look like in practice.
You can also point out what matters in them: how you explain things, the phrases you use, how direct you are or where you add humour.
That gives you something more useful to work with than repeatedly asking, “Can you make it sound more like me?”
If you’re still spending ages getting the voice right, that’s something I can help with through my AI editing service. You bring the draft, and I help shape the voice, message and flow.
Show AI what your voice looks like on the page.
Prompt 4. “Give me two regular and two unexpected suggestions”
Sometimes I want the obvious options. They might be obvious because they make perfect sense.
But I also want to see what else we could do.
This is one of my favourite ways to open up the thinking:
Give me two regular and two unexpected suggestions for approaching this topic. Keep all four relevant to my audience and explain why each could work. The unexpected ideas should offer a useful different perspective, rather than just being quirky.
Say you’re a kitchen designer writing about renovations.
The regular suggestions might be planning your layout and choosing materials. Both useful.
An unexpected suggestion might be to ask readers what annoys them during the breakfast rush. Suddenly, you’re talking about cupboard doors colliding, nowhere to put the toaster and three people trying to reach the same drawer.
That’s a different way into the subject, grounded in something your reader recognises.
The unexpected option won’t always be the best one. And asking for it doesn’t guarantee an original idea. But it gives you more to consider before settling on a direction.
Give yourself a choice beyond the first sensible answer.
Prompt 5. “Where is this generic, and what would make it specific to my business?”
A draft can sound like you and still say very little about your business.
“We provide personalised solutions and exceptional service.”
Lovely. But what does that actually mean?
This prompt helps you look more closely:
Identify anything in this draft that could appear on a competitor’s website unchanged. Ask me for the examples, experiences or details you need to make it specific to my business. Don’t invent them.
Perhaps “personalised service” means you meet the client before recommending anything. Perhaps you explain every option without jargon. Perhaps you’ve built your process around a problem you kept seeing customers struggle with.
Those are useful details.
This is something I look for when I’m editing: where can we replace a broad claim with something the reader can picture, understand or believe?
It’s also where your experience becomes valuable. AI can help you express your approach, but you need to supply the real stories, observations and ways of working.
The details that distinguish your business need to come from your business.
Bonus prompt: “I’m really excited about getting this right…”
This one caught my attention because it sounds suspiciously like giving your computer a pep talk.
Researchers have explored whether adding emotional language to prompts changes AI’s responses. Some studies have found improvements, although the results don’t support a simple rule that positive prompting always wins. Negative emotional prompts have also improved performance in some tests.
Now, I’m not suggesting AI has feelings or needs you to cheer it on.
But I do think there’s a useful idea to try here: explain your enthusiasm, why the work matters and what doing it well would look like.
For example:
I’m really excited about this piece because it could help business owners feel more confident about creating content. I want it to be genuinely useful. Let’s make the advice practical, the examples specific and the explanations easy to follow. Keep challenging anything that doesn’t serve that goal.
Notice that the excitement comes with direction.
“Let’s make this amazing!” doesn’t tell AI much about what amazing means. Explaining what you care about gives it more context.
And keep prompt two handy. Being enthusiastic about an idea doesn’t mean you want unquestioning agreement.
Share why you care, along with what a good result needs to achieve.
Better prompts are the start of better content
You don’t need to memorise a complicated formula to get more from AI.
Start by explaining the job properly. Show it how you write. Ask it to challenge your thinking, explore different approaches and flag where the content could belong to anyone.
Then read what comes back with your customer in mind.
Does it answer their question? Does it give them a reason to care? Does it help them understand your business and take the next step?
That’s the part I bring my copywriting experience to, whether I’m working from an AI draft or creating the content for you.
If you’ve got a draft that needs help, my AI editing service can help you get the voice, clarity and message right.
And if you’d rather hand over the blogs, emails and social posts altogether, my content writing services give you that option too.
Get in touch and tell me what you’re working on. We’ll work out where I can help.
Frequently asked questions about AI prompts
-
No. Use them where they’re helpful. Context and voice belong near the start. Asking for different approaches is useful when you’re exploring ideas, while checking for generic content works well once you have a draft.
-
Clear instructions and examples of your writing can help bring it closer. You’ll still need to review the result for wording you wouldn’t use, details it’s missed and whether the message says what you intended.
-
Yes. Prompting gives the draft direction. Editing is where you check the facts, sharpen the message, remove repetition and make sure the content works for your reader and your business.