Why AI-Generated Meta Ads Are Killing Your Cost Per Click

Every service based business's Meta ad looks the same right now and it's not a coincidence. Here's why ChatGPT-generated creatives are quietly inflating your cost per click, with real examples.

Over the last week, I've been screenshotting every ad Instagram's been feeding me.

I've collected almost 100 creatives across completely different industries that have naturally popped up in my Instagram feed.

From plumbers, concreters, locksmiths, spas, beauticians I have really seen it all across such a diverse range of industries.

….But somehow all of these meta ads look almost identical.

Well… That "somehow" is ChatGPT.

Business owners are being sold this delusional shortcut: 

“Skip the creative process, skip the market research, let AI spit out a final product.”

What they're not being told is that the shortcut has a cost, and it shows up directly in their ad spend.

Don’t get me wrong it is costing them fuck all upfront, I understand the appeal totally.

However, it is in the long run that they pay the true price of taking these short cuts.

Understanding what your competitors are doing with their advertising isn't just useful for inspiration.

It's how you spot what's working, and just as importantly, what to avoid so you can actually stand out.

Right now, this is what's flooding the feed:

Logically, you can already see the problem..

IF every ad in the category looks the same, it's harder to stand out. But it's also so much more deeper than that.

Understanding how social media algorithms work and what that means for your business is crucial here.

Put simply: algorithms exist to improve the user experience, because platforms make money by keeping people scrolling. When a platform gets flooded with creatives that all look the same, it doesn't just make your ad blend in visually. 

The algorithm flags it as "more of the same."

That's the part most business owners never connect.

Every platform scores your ad against everything else it's already shown to that user. 

When your creative looks identical to the fifty others they've scrolled past this week, the algorithm reads it as low-relevance content, even if your offer is genuinely different.

That score isn't cosmetic. It directly decides your cost per click.

Low relevance means the platform charges you more to reach the same person, because it has to work harder to justify showing your ad over something fresher.

So the real cost of a ChatGPT-generic ad isn't just "it looks the same as everyone else's." 

It's that you're paying a premium to be ignored! 

Funding the algorithm's assumption that your content isn't worth prioritising, while a business running a genuinely distinct creative gets cheaper reach simply because the platform hasn't seen it a hundred times already this week.

Your ad spend ends up higher than if you'd created something more authentic to begin with. Not because authentic content is inherently better creatives, but because the algorithm is built to reward what it hasn't seen and punish what it has.

Authentic doesn't mean expensive or polished. It means:

• Collecting content on site of your business, with you, your team doing what you do best professionally on camera.
• Built from real market research on what your specific competitors are already saturating the feed with
• Different enough in structure (not just colour) that the algorithm can't pattern-match it to the template above

The market research step everyone's skipping isn't optional anymore. 

It's the difference between an ad that costs you $2 a click and one that costs $8 for saying the same thing, just wearing someone else's outfit.

Want to take your social media content seriously? [Check out how we approach social media strategy here.]

None of this means AI is off the table for your ads. 

It just means the way most people are using it… as an autopilot for the whole creative is the exact thing tanking their results.

Here's the difference between using AI as a shortcut and using it as a tool.

The prompt everyone's running:

"Make me an ad for my plumbing business, here's what we offer and here's our contact details."

That prompt gives the model no direction beyond your industry and your offer list — so it defaults to the most common pattern in its training data for "trade business ad." 

Which is exactly the template you've seen eight times over in this post. Every business typing a version of that sentence gets a version of that ad.

If you want AI in the mix, use it differently:

1. Use it for inspiration, not output.

Ask it to describe five different layout concepts for your offer. Not to design the finished creative. Treat what it gives you as a starting brief, then either build the actual creative yourself, get a designer to execute it, or heavily rework whatever the AI produces so it stops looking like AI's first instinct.

2. Say one thing per ad.

Look closely at every creative in this post and you'll spot the same fault line: they're crammed. Ten services in one grid. Five benefits stacked on a pricing table on a trust bar. None of the information is wrong, there's just too much of it for a thumb scrolling past in half a second. A person can't absorb a full service menu in an ad; they can absorb one idea. Pick the single offer, benefit, or result that matters most this week and build the entire creative around saying that one thing clearly.

3. Replace the AI render with something real.

Stock-style AI photography is its own tell, the lighting is too perfect, the hands are slightly wrong, the job site looks like no job site that's ever existed. Swap it for something actually shot on-site: your team, your van, your actual work. This is also the single easiest way to break pattern-match with every other ad in the category, because no two businesses' real job sites look the same.

4. Let AI help with the words, not the whole ad.

If you want a genuinely useful AI task, ask it to draft five different one-line hooks for your offer, or to help you sharpen a headline down to fewer words. That's a much smaller, more controllable job than "design my whole ad," and it's far less likely to default to a template.

5. Test whether it's actually distinct — not whether you like it.

Before you publish, hold your ad next to two or three competitors' current ads. If yours could have any logo swapped in and still make sense, it's still too generic. The bar isn't "does this look good" plenty of the ads in this post look polished. The bar is "does this look like nothing else in the feed right now."

Do this and the maths flips. Instead of paying a premium for the algorithm to treat you as more of the same, you're the fresh signal it hasn't scored yet.

Which is exactly where the cheap clicks live.