How to Write Badass Ad Copy in the Age of AI

Amy McGrath • May 19, 2026

Share this article

High CTR Hooks, Offer Clarity, Pain Points, Creative Testing and Prompts That Actually Work


Here's the uncomfortable reality of running paid ads right now. Everyone has access to AI. Everyone is using it to write copy. And because everyone is using the same tools, trained on the same data, fed the same generic prompts — the ads flooding every feed are starting to sound identical.


Same hooks. Same structure. Same 'Are you tired of [problem]? Introducing [solution].' Same energy. Same wallpaper.


Which, if you think about it, is actually good news for anyone willing to put a bit more thought in.

AI hasn't made great ad copy obsolete. It's made lazy ad copy invisible. The bar for stopping the scroll has gone up, not down — and the brands that understand that are cleaning up while everyone else wonders why their CPMs keep climbing and their CTRs keep sinking.


This is about writing ad copy that cuts through. The hooks, the offer framing, the pain points, the testing frameworks, and how to use AI as a proper creative tool rather than a glorified copy-paste machine.

The Hook Is Everything. No, Really.


You've got about 1.7 seconds. That's the average time someone takes to decide whether to keep scrolling or stop and look. Your hook — the first line, the first frame, the headline — is the only thing standing between your ad and the void.


Most hooks fail for one of three reasons: they're too clever, too vague, or they say nothing the reader hasn't heard before. The fix isn't complicated, but it does require actually thinking about who you're talking to and what's going to make them feel something in under two seconds.

"On the average, five times as many people read the headline as read the body copy. When you have written your headline, you have spent eighty cents out of your dollar."  — David Ogilvy

The hooks with the highest CTR tend to share a few characteristics. They're specific rather than broad. They create tension — either curiosity, relatability, or a mild sense of challenge. And they don't try to say everything at once.


Hook formats that consistently perform:


  • The specific problem — "Your Meta ads are getting clicks. Nobody's buying. Here's why." Calls out the exact frustration. The right person stops immediately.


  • The counterintuitive statement — "Stop optimising your ads. Fix your offer first." Challenges a belief the audience holds, which forces a reaction.


  • The before/after without the middle — "She went from £800 a month to £14k. Same product. Different copy." Creates a gap the reader wants to close.


  • The direct callout — "If you're a personal trainer still charging by the hour, read this." Talks to one specific person.


  • The uncomfortable truth — "Most freelancers don't have a pricing problem. They have a confidence problem." Slightly stings. Hard to scroll past.


  • The number — "3 reasons your Facebook ads stopped working in 2024." Simple, clear, promises a fast payoff.

💡 Fun Fact: According to research from Copyblogger, 80% of people will read a headline but only 20% will read the rest of the copy. The hook isn't the entrance to the ad — it is the ad. Everything else is just evidence.

Offer Clarity: If They Have to Work to Understand It, You've Lost


A great hook gets the click. A clear offer gets the conversion. And most ad copy — even copy from brands spending serious money — buries the offer in language so vague, so hedged, so wrapped in aspirational fluff that the person reading genuinely isn't sure what they'd be getting.


'Unlock your potential.' What does that mean? 'Transform your business today.' How? 'The solution you've been waiting for.' To what?


Your offer needs to be immediate, specific, and self-explanatory. Someone who's never heard of your brand should be able to read your ad and know exactly what you do, what they get, what it costs, and why now rather than later.

"Make it simple. Make it memorable. Make it inviting to look at. Make it fun to read."  — Leo Burnett

The four things your offer needs to communicate:


  • What it is — exactly, without jargon. Not 'a transformative coaching experience.' A 12-week programme. Six one-to-one calls. A private community. Real things.


  • Who it's for — the more specific the better. 'For ecommerce brands doing £20k–£100k a month' excludes most people and speaks directly to the right ones.


  • What changes — the outcome, not the process. Not 'we'll audit your ads' but 'you'll know exactly where your budget is being wasted within five days.'


  • Why right now — a deadline, a cohort opening, a limited offer, or simply a clear first step. Remove the option to file it under 'I'll think about it.'

💡 Fun Fact: A/B tests run by ConversionXL found that adding specific numbers to offer statements — days, percentages, prices, quantities — increased conversion rates by an average of 36% compared to vague equivalents. Specificity isn't just clearer. It's more believable.

Pain Points: Say the Thing They Haven't Said Out Loud


The reason pain-point copy works so reliably isn't manipulation — it's recognition. When someone reads a line that accurately describes how they feel about a problem they've been quietly carrying around, something clicks. It feels like the brand actually gets it. And a brand that gets it is one you're more likely to trust with your money.


The mistake most people make with pain-point copy is going too surface level. 'Struggling to grow your business?' is technically a pain point but it's so broad it doesn't register. Everyone is struggling to grow their business. That line doesn't stop anyone.


Specificity is what makes pain points land:


  • Too broad: "Struggling to get leads?"  |  Better: "Spending £500 a week on Meta ads and getting enquiries from people who can't afford you?"


  • Too broad: "Tired of feeling stuck?"  |  Better: "You've got the skills. You've got the clients. But every month still ends with you checking your bank balance and doing the maths."


The second versions make the reader feel seen. Not just categorised — actually understood.

Where to find the real pain points:


  • Customer reviews — especially one and two stars. People are brutally specific when they're disappointed.


  • Reddit threads and Facebook groups in your niche — unfiltered, unedited real talk from your actual audience


  • Your own DMs and support inbox — what do people complain about or ask before buying?


  • Post-purchase surveys asking 'what nearly stopped you buying?'


  • Sales calls — what objections come up every single time without fail?

💡 Fun Fact: The Emotional Motivation study by Google and CEB found that B2B customers are 50% more likely to buy when they feel a personal connection to a brand — and that connection is built faster through pain-point recognition than through feature lists or credentials. The principle holds equally in DTC and ecommerce.

"Don't tell me how good you make it; tell me how good it makes me when I use it."  — Leo Burnett

Creative Testing: The Part Most Brands Skip Entirely


Here's something that doesn't get said enough. No copywriter, no matter how good, knows which ad will win before it runs. Anyone who tells you otherwise is either lying or hasn't looked at enough data. The best creatives in the industry don't write one great ad — they write ten decent ones and let the algorithm tell them which three are great.


Most brands test too little, too slowly, and too inconsistently to learn anything useful. They run one version of an ad, it underperforms, they conclude the audience is wrong or the product isn't right for paid media. Meanwhile the issue was a single word in the hook.


A simple creative testing framework:


  • Test one variable at a time — if you change the hook, the image, and the CTA in the same test, you'll never know what moved the needle. Isolate each element.


  • Test hooks first — before you optimise anything else, make sure your hook is working. Run three to five hook variants against the same body copy and let them gather real data.


  • Then test the offer framing — once a hook is pulling clicks, test how you're presenting the offer. Direct vs. curiosity-led. Price-first vs. outcome-first. Short vs. long.


  • Then test format — single image vs. video vs. carousel vs. UGC-style. The format that works for one audience can completely bomb for another.


  • Keep a creative log — date, format, hook, result, notes. If you're not recording what you tested and what happened, you're not learning. You're just spending.

💡 Fun Fact: Meta's own research suggests that creative is responsible for up to 70% of ad performance variability. Not targeting. Not bidding strategy. The creative. Which means the most high-leverage thing you can do for your paid media results is write and test more of it.

Ad Variants: One Idea, Many Angles


Every offer has multiple angles. The problem most copywriters have is picking one — usually the most obvious one — and writing all their ads from there. But your potential customers don't all feel the same thing about the same problem. Some are motivated by fear of loss. Some by aspiration. Some need social proof before they'll consider it. Some need the logic before the emotion lands.


Writing ad variants isn't about creating more work. It's about covering more of your audience's decision-making triggers with the same offer.


For any given product or service, write a variant for each angle:


  • Pain-first — leads with the problem. The reader recognises the frustration and wants the solution.


  • Outcome-first — leads with the result. Aspirational, forward-looking. Works well for transformation-based offers.


  • Social proof-first — leads with a customer result or review. Borrowed credibility before you've built your own.


  • Curiosity-driven — leads with a question or a counterintuitive statement. Works well for audiences who don't yet know they have the problem.


  • Direct and plain — no cleverness, no story. Just 'here's what it is, here's what you get, here's the price.' Underrated. Often performs surprisingly well.


  • Story-led — a short narrative. Someone in the situation your audience is in, who found the solution. Works particularly well in video but can land in static copy too.


Run all of them. You'll almost always be surprised by what wins.



Using AI to Write Better Copy (Not Just Faster Copy)


This is where most people are getting it wrong. They open ChatGPT or Claude, type 'write me a Facebook ad for my yoga studio,' get something generic back, tweak a couple of words, and post it. Then wonder why it sounds like every other yoga studio ad on the internet.


AI is a creative collaborator, not a vending machine. The quality of what comes out is almost entirely determined by the quality of what goes in. A lazy prompt produces lazy copy. A specific, structured prompt with real context produces something you can actually use — or at least a strong draft you can shape into something great.

"The tools have changed. The fundamentals haven't. You still need to know your customer, know your offer, and give people a reason to act."  — Gary Halbert, adapted

Prompt frameworks that produce usable ad copy:

  • For hooks:

    Write 10 Facebook ad hooks for [product/service]. Target audience: [describe specifically — job, situation, frustration]. Main pain point: [one specific thing]. Tone: [conversational / direct / punchy]. Avoid: clichés, vague language, corporate speak. Each hook should be under 15 words.

  • For full ad copy:

    Write a Facebook ad for [product/service]. Format: hook, two-line body copy, CTA. Audience: [specific description]. Their main frustration is [X]. The outcome we deliver is [Y] in [timeframe]. Offer: [what they get + price or first step]. Tone: honest, direct, like a knowledgeable friend. No fluff.

  • For pain-point variants:

    My customer is a [description]. They struggle with [specific problem]. Write 5 ad copy variants, each leading with a different pain point angle. Keep each under 60 words. Use plain language.

  • For testing angles:

    Take this offer: [describe offer]. Write one ad from each of these angles: pain-first, outcome-first, social proof-first, curiosity-driven, and direct/plain. Same offer, different emotional entry points.

💡 Fun Fact: Research from the Content Marketing Institute found that marketers who use AI with structured prompts and human editing produce content rated significantly higher for quality and originality than those using AI output without editing. The tool is a starting point, not a finish line.

A few prompt principles worth keeping in mind:


  • Always give the AI your real audience — not 'small business owners' but 'female founders running product-based businesses, £5k–£15k monthly revenue, stuck on scaling beyond their personal network'


  • Give it the actual pain point, not a category — not 'they want to save time' but 'they spend Sunday evenings doing the admin they avoided all week and it's killing their motivation'


  • Tell it what to avoid — 'no buzzwords,' 'no rhetorical questions,' 'don't start with I' — constraints produce better output


  • Ask for multiple variants, not one version — then combine the best elements across them


  • Always edit the output — not just for grammar, but to add your brand voice, your specific proof points, and the things only you know about your audience



Putting It All Together


Great ad copy in the age of AI isn't about out-prompting everyone else. It's about understanding your audience well enough that your specific insight — the real pain point, the genuine outcome, the thing that only comes from actually knowing your customer — gives your copy an edge that generic AI output simply can't replicate.


That means:


  • Hooks that create tension — specific, fresh, and written for one person rather than everyone


  • Offer clarity that removes confusion — what it is, who it's for, what changes, and why now


  • Pain points that feel personal — found from real customer language, not guessed from a brief


  • A testing framework — running enough variants, measuring properly, letting the data tell you what works


  • Multiple angles for every offer — because your audience doesn't all buy from the same emotional trigger


  • AI as a collaborator — fed with real context, producing real drafts, shaped by a human who knows the brand


The brands winning on paid social right now aren't the ones with the biggest budgets or the fanciest creative studios. They're the ones who understand their customers deeply, test relentlessly, and use every tool available — including AI — in service of that understanding.


Write more. Test more. Learn faster. Edit everything.


That's the job.

Recent Posts

By Anthony McGrath May 19, 2026
Let's be blunt about something. There's probably a version of your product page sitting right now at a decent ranking, getting reasonable traffic, and quietly bleeding sales because the copy is doing absolutely nothing. Not bad copy, necessarily. Just... copy. Words that exist. Manufacturer descriptions rephrased slightly. Dimensions listed. Stuff in a bullet list that nobody asked for. The hard truth is that most ecommerce businesses put enormous effort into getting people to the product page — ads, SEO, influencers , the lot — and then hand them a description that reads like it was written for a stock database. And then wonder why the conversion rate is stuck at 1.8%. Product description optimisation isn't really about writing. It's about removing every possible reason someone might not buy. Handle the objections. Build the trust. Make the decision feel easy. Do all of that while keeping Google happy, and you've got a page that works around the clock without you having to touch it. Here's how that actually looks in practice.
By Amy McGrath May 19, 2026
What Actually Brings Clicks, Leads, and Trust Most people treat SEO like a numbers game. Chase the keyword, hit the word count, grab the ranking, done. But here's the uncomfortable truth: a page can sit at position one and still fail completely. No clicks. No leads. No trust built. Just a metric that looks good in a report nobody reads twice. Real SEO content does three things at once — it earns the ranking, compels the click, and converts the reader into someone who actually does something. Getting all three right requires a different way of thinking about content from the very start. This is that guide.
By Amy McGrath May 14, 2026
You spend ages on your CV. You tidy the font, rewrite the bullet points, move sections around and try to make everything sound sharper. Then you apply for a role that looks right for you and hear absolutely nothing back.  That silence can feel personal, but it is not always a sign that you are underqualified. Sometimes the issue is much simpler: your CV was not read properly by the system before a recruiter ever had the chance to look at it. Applicant Tracking Systems , usually shortened to ATS, are now part of normal online recruitment. They collect applications, extract candidate details, search for keywords and help employers shortlist faster. The important point is this: your CV now has two audiences. It has to be easy for software to scan, and it still has to sound clear, credible and human when a real person opens it. This guide walks through the practical bits that matter: layout, formatting, keywords, section headings and the simple checks you can run before sending your next application.
By Amy McGrath May 13, 2026
You've spent years building skills, climbing ladders, and collecting achievements. Then you condense it all into two pages — and hear nothing. That silence is the most common experience in modern job searching. According to Forbes, less than 3% of submitted CVs result in an interview. The problem, more often than not, isn't who you are. It's how your CV presents you. Here are 15 evidence-based fixes that make the difference.
By Amy McGrath May 8, 2026
Getting people to visit your website is only the first step. The real growth comes from what happens next: do they remember you, trust you, come back, sign up, enquire, buy, or recommend you?  Website traffic is valuable, but repeat attention is more valuable . A visitor who returns three, four or five times is far more likely to become a customer than someone who lands once and disappears. That is why smart brands build systems to keep in touch with visitors after the first click.
By Anthony McGrath May 8, 2026
Writing blog posts that increase site traffic is not about filling a page with keywords and hoping Google notices. Strong blog content works because it answers a real question, gives useful detail, and makes the reader want to stay, click, save, share, or enquire. Google describes good SEO as helping search engines understand your content while helping users decide whether they should visit your site. That means your blog post needs to serve both sides: the person reading it and the search engine trying to understand it.