AI is not the problem.
The vague, lazy prompt you gave it probably is.
That might sound blunt, but it is also good news.
Because when AI gives you generic, beige, painfully obvious content, you do not need a new tool. You need to give the tool better direction.
AI can help you research customers, write ads, create emails, map funnels, analyse campaigns, develop lead magnets, and generate content faster.
But it cannot read your mind.
The quality of the result depends heavily on the quality of the instructions.
So before you blame ChatGPT for sounding like a robot, check whether your prompt gave it enough useful information to work with.
Here are six things you need to know if you want AI prompts that produce genuinely useful results.
- Start With a Clear Outcome
Most people start with a task.
“Write me a Facebook ad.”
“Create some social media posts.”
“Give me some email ideas.”
That tells AI what to make, but not what the content needs to achieve.
There is a big difference between:
“Write a Facebook ad for my bookkeeping business.”
And:
“Write a Facebook ad designed to generate qualified leads from established business owners who are frustrated by poor cash flow and want better financial visibility.”
The second prompt gives AI a commercial outcome.
That matters.
Your prompt should explain what success looks like.
Are you trying to:
• Generate qualified leads
• Sell a product
• Book consultations
• Build trust with a cold audience
• Retarget people who already know you
• Educate prospects before making an offer
• Increase the value of existing customers
AI works better when it understands the job the content needs to do.
Do not just ask for words, ask for a result.
- Tell AI Exactly Who It Is Speaking To
Your audience is not “business owners.”
That is too broad.
A plumber running a $300,000 local business has different concerns from a finance company turning over $10 million.
They may both own businesses, but they do not think the same way, use the same language, or make buying decisions for the same reasons.
Your prompt should include details such as:
• The type of person you want to reach
• Their business stage
• Their current problem
• What they have already tried
• What they are frustrated by
• What they want instead
• What may stop them from taking action
Good marketing, AI or not, starts with understanding the customer before you write the ad, email, page, or post.
AI can help speed up this research, but you still need to give it a useful starting point.
When you understand what your audience wants, what they fear, and what is holding them back, your messaging becomes more specific and far more effective.
3. Give It a Role, But Make the Role Specific
“Act as a marketing expert” is not enough.
There are plenty of marketing experts who would give completely different advice.
Be more precise.
For example:
“Act as if you are a direct-response copywriter who specialises in Facebook ads for established service businesses.”
Or:
“Act as a customer research strategist. Analyse the following reviews and identify recurring frustrations, desired outcomes, objections, and phrases customers use.”
Or:
“Act as a funnel strategist. Review this offer and identify where prospects may lose trust or drop out before buying.”
You can even ask AI to act as a specific public figure if you like their writing style:
“Act as if you are David Ogilvy. Write a Facebook ad for my offer using clear, persuasive direct-response principles, a strong headline, specific benefits, credible proof, and no hype.”
You can also tell AI what not to prioritise.
For example:
“Do not focus on engagement, likes, or follower growth. Focus on qualified leads, sales, and profit.”
That one instruction can completely change the output.
4. Show It What Good Looks Like
AI performs much better when you give it examples.
This is especially important when you want content to sound like you.
You can provide:
• Previous emails
• Blog posts
• Ad copy
• Sales pages
• Video transcripts
• Social posts
• Phrases you regularly use
• Phrases you never want used
Then tell AI what to pay attention to.
For example:
“Study these examples and identify the tone, sentence length, level of directness, rhythm, structure, and repeated language patterns. Then write the new piece using those patterns.”
Do not simply say, “Write in my voice.”
Give it evidence of your voice.
You should also explain what you do not want.
That might include:
•No hype
• No corporate language
• No empty motivational lines
• No fake urgency
• No long introductions
• No generic phrases
• No unnecessary adjectives
• No robotic summary at the end
5. Add Constraints
Constraints make AI more useful.
Without them, it tends to overwrite, repeat itself, use tired phrases, and explain things nobody asked it to explain.
Useful constraints could include:
• Keep paragraphs to two or three lines
• Use plain English
• Avoid jargon
• Do not use clichés
• Include three practical examples
• Write at a Year 8 reading level
• Keep the email under 250 words
• Include one clear call to action
• Do not open with a greeting
• Do not use emojis
• Do not make claims without proof
• Do not mention the product until the second half
You can also give structural constraints.
For example:
“Open with a blunt observation. Explain the problem. Give three practical fixes. Address the main objection. Finish with one natural next step.”
This gives the content direction without forcing it into a stiff formula.
The goal is not to make the prompt complicated.
The goal is to remove ambiguity.
6. Treat the First Output as a Draft
This is where people either get good at AI or give up on it.
They enter one prompt, read the answer, and decide AI is brilliant or useless, and neither conclusion is particularly helpful.
The first response is often a starting point.
You can improve it by giving clear feedback.
Instead of saying:
“Make it better.”
Say:
“The opening is too generic. Lead with the financial cost of getting this wrong.”
“The tone is too formal. Make it sound like an experienced business owner explaining this over coffee.”
“The copy is focused too heavily on features. Rewrite it around the customer’s frustrations and desired outcome.”
“The call to action is too vague. Make the next step specific and low-friction.”
“This sounds like every other coach online. Remove the hype and make the argument more commercially intelligent.”
AI improves when your feedback is specific.
You can also ask it to critique its own work.
Try:
“Review this copy as a sceptical buyer. Identify anything vague, repetitive, unbelievable, or too salesy. Then rewrite it.”
Or:
“Score this prompt out of ten for clarity, context, audience detail, commercial objective, constraints, and tone. Tell me what is missing before completing the task.”
That one step can save a lot of editing later.
A Simple AI Prompt Structure You Can Use
You do not need a 47-page prompt manual.
Use this structure:
Role: Who should AI act as?
Task: What exactly do you want it to create or analyse?
Audience: Who is the output for?
Objective: What business result should it support?
Context: What does AI need to know about your offer, customer, or situation?
Proof: What evidence, examples, or results can it use?
Tone: How should it sound?
Constraints: What must it include or avoid?
Output: What format do you want?
The real skill is clear thinking.
AI forces you to clarify:
• Who you are speaking to
• What they care about
• What you want them to do
• Why they should trust you
• What makes your offer worth considering
Those are not AI questions.
They are marketing questions.
Used properly, AI can help you move faster, research deeper, test more ideas, and create better marketing without staring at a blank screen for three hours.
But it should accelerate your strategy, not replace it.
And if you’re sick of wasting money on marketing that doesn’t work, come along to my AI-driven Facebook Ads Masterclass.
https://app.socialmediaplusone.com/master-meta-7196
Kerry x
