Listen to the podcast below or read the full article:
This week, I was auditing a new client’s Google Ads account, and I spotted something that’s becoming all too common.
They’d switched on AI Max a few weeks back. Prompted by Google’s swift marketing of the new product.
The algorithm was running wild, pulling in search terms for regions they didn’t even service. Money was going out the door for clicks that had zero chance of converting.
When I asked them why they turned it on, they said, “Well, everyone’s saying AI does it better, so we thought we’d give it a go.”
That right there? That’s the problem.
Recently, I came across an article on Search Engine Land by Rémi Kerhoas that I think also really hit the nail on the head. It’s called “3 PPC myths you can’t afford to carry into 2026,” and it breaks down exactly why blindly trusting AI in your paid ads is lazy strategy.
Today, we’re going to unpack that, combine it with what I’m seeing in real client accounts, and give you a clear plan for how to actually use AI properly in your paid media.
The Problem With “Just Trust the Algorithm”
Here’s what I’m seeing across the board right now.
Businesses are blindly switching on AI features in Google Ads and Meta without understanding what those features are actually doing. They’re reading articles and “gurus” that say “AI targeting outperforms manual targeting,” so they consolidate everything, let the algorithm loose, and then wonder why their results are average at best.
The Search Engine Land article mentioned above breaks down three big myths that spread like wildfire in 2025. And I’m seeing all three of them play out in client accounts right now.
But before we get into those myths, you need to understand one thing:
AI and algorithms are incredibly powerful. They’re taking a lot of the guesswork out of marketing for us.
But – and this is critical – they only work when you feed them the right data.
Bad signals and weak foundations just get scaled faster.
That’s it.
That’s the game.
Myth 1: Manual Targeting Is Dead and AI Does It Better
Let’s start with the first myth: that manual targeting is outdated, and AI just does it better.
There’s truth in that statement, but only under very specific conditions.
Here’s what the article points out: AI performance depends entirely on your inputs. If you don’t have enough volume, the AI can’t learn. If you don’t have clean conversion signals, the AI optimises for the wrong thing.
For big e-commerce brands that are feeding purchase data back into Google Ads and consistently getting at least 50 conversions per month per bid strategy? Yeah, AI targeting makes sense. They’ve got the volume and the signal quality.
But for most businesses? Especially service-based businesses or B2B companies tracking leads instead of purchases? That logic breaks down fast.
A Real Example From Our Client Work
We’ve got a client who came to us spending between $750 to $1,000 a day across multiple campaigns. They weren’t hitting that budget every day though – they’d hit a ceiling. We couldn’t figure out how to reach more people with the keywords they were targeting.
So we tested AI Max. We set it up properly – made sure conversion tracking was clean, made sure the website was solid, set up negative keywords. And you know what? It worked. We found new customers they didn’t even know were searching for them.
But then this week, I audited a different client who’d turned on AI Max without any of that groundwork. No proper location exclusions. Messy conversion tracking. And despite having location settings configured, AI Max was showing ads to people in regions the client doesn’t even service.
Why? Because the algorithm was looking for anyone who might be interested in that service, regardless of whether the business could actually help them.
That’s what happens when you blindly trust the algorithm without understanding what signals you’re sending it.
How to Fix Your AI Targeting Strategy
Before you hand over targeting decisions to AI, ask yourself three questions:
- Are your campaigns optimised against a business-level KPI? Not just “get leads,” but “get leads that cost less than X” or “get sales above Y% margin.”
- Are you sending enough conversions back to the ad platform? You need volume for AI to learn from.
- Are those conversions reported quickly, with minimal delay?
If you answered no to any of those questions, 2026 needs to be about fixing your fundamentals.
And don’t be afraid to go old school when the situation calls for it.
The Search Engine Land article gives a great example. The author doubled a client’s margin by implementing a match-type mirroring structure and pausing broad match keywords. That runs completely counter to what Google tells you to do, but it worked because they looked at the data.
Their broad match keywords were delivering leads at $33 each. Sounds great, right? Except that those leads had a customer acquisition cost of over $2,000. Meanwhile, exact match keywords cost $35 per lead but had a CAC of $450.
Google’s algorithm was doing exactly what it was told: deliver the cheapest leads possible. But those cheap leads weren’t converting to customers.
When they took back control and focused the budget on exact match – the keywords that were actually converting – margins doubled.
That’s the point. The algorithm is literal. It does what you tell it to do. If you’re telling it the wrong thing, you get the wrong results.
Myth 2: More Creative Variations Mean Better Campaign Performance
The second myth is all about creative volume.
In 2025, Meta rolled out Andromeda, and suddenly everyone started talking about how more creative means more learning, which means better results.
And look, there’s logic there. More creative variations give the platform more options to match messages to different people and contexts.
But in practice, this myth more reliably increases creative production costs than it improves results.
Here’s why: creative volume only helps when the ad platform is receiving enough high-quality conversion signals. Without those signals, more ads just means more assets to rotate through. The AI has nothing meaningful to learn from.
I see this all the time. Businesses pump out 20 different ad variations, all generated with AI, and they’re confused about why performance stays flat.
It’s because they haven’t fixed the conversion tracking. They haven’t improved their customer journey. They’re just adding more noise to a system that doesn’t have the data it needs to make good decisions.
What to Focus On Instead of Creative Volume
When resources are limited – and for most SMEs, resources are always limited – creative volume is not your best move.
Instead, focus on conversion rate optimisation.
Review your tracking. More tracked conversions improve performance. If you’re only tracking form fills but the real value is when someone books a consultation, track that consultation.
We’ve got a client who runs a quiz to self-qualify leads before they book a consultation. Initially, they were tracking quiz completions as the conversion. Great for reporting, but it told us nothing about lead quality.
When we switched to tracking actual consultation bookings, the algorithm completely changed its behaviour. It started finding people who were more likely to complete the quiz AND book the call, not just anyone who’d click through.
That’s the power of clean conversion signals.
After that, improve your customer journey. Make it easier for people to convert. Then map higher-margin products or services so the algorithm knows what actually makes you money.
Only then – once you’ve got clean data flowing through – should you start testing creative variations at scale.
Creative scale follows signal scale. Not the other way around.
Your AI-Ready PPC Checklist
If you’re considering using AI features in your Google Ads or Meta campaigns, here’s what you need to check first:
1. Conversion Tracking
Is your conversion tracking accurate? And I mean actually accurate, not just “we think it’s working.”
If you’re tracking form fills but the real value is qualified leads, you’re sending the wrong signal. If you’re tracking clicks to your website but not what happens after that, you’re blind.
The AI makes decisions based on what you tell it is a conversion. Make sure you’re tracking the actions that actually matter to your business.
2. Website Quality
Is your website fast? Is it mobile-friendly? Is the customer journey clear?
Because if you turn on AI Max or Advantage+ and it starts sending traffic to different pages on your site automatically, those pages better work. If they’re slow, confusing, or broken, you’re just paying for a bad experience.
3. Data Volume
Are you getting enough conversions for the AI to learn from?
If you’re running campaigns that get 5 conversions a month, AI can’t do much with that. You need volume. The general benchmark is at least 50 conversions per month per bid strategy.
If you don’t have that yet, stick with manual campaigns or focus on fixing your conversion rate until you do.
4. Signal Quality
Are you tracking negative signals?
If someone fills out a form and you mark it as a conversion, but then 80% of those leads turn out to be junk, you’re teaching the algorithm to find more junk.
Track the conversions that actually lead to revenue. And if you can, send negative signals back to the platform when a lead turns out to be low quality.
5. Campaign Structure
Is your campaign structure built for AI, or is it a mess from 2018 that you’ve been patching together?
AI works best with clean, simple structures. If you’ve got 47 ad groups in one campaign, all targeting different things, consolidate.
The Bottom Line on AI and Paid Advertising
So I’m hoping you can see my main point from this entire article: the problem isn’t AI. It’s misuse.
Platforms do exactly what they’re asked to do. They optimise against the signals you provide, within the constraints of your budget and structure.
When your fundamentals are broken, AI can’t fix it.
It just scales the problem faster.
This year, continue your focus on getting your foundations right, feeding the algorithm clean signals, and then letting it do what it’s actually good at.
But the main thing I want you to take away from this is simple:
AI is powerful. But it’s not magic.
And if you’re not feeding it the right signals, you’re just burning money faster.