by Austin LeClear
on October 6, 2025
Google Ads AI Max for Search is officially here, and if you’re wondering whether you should activate AI Max on your Search campaigns, we’ve got some honest insights to share. After testing this across our client accounts at Grow My Ads, we’re diving into what AI Max offers, how it actually performs in real campaigns, and whether it delivers better performance than traditional Search campaigns.
Prefer video? Check out Google AI Max: Why 99% of Businesses Will Lose Money (Setup + Results) on YouTube here:
Still with us? Good, let’s dive in.
What Is Google’s AI Max for Search?
First off, let’s break down what AI Max is. At its core, Google Ads AI Max combines two main features: search term matching and asset optimization.
AI Max isn’t revolutionary technology. Google has essentially taken existing features like Dynamic Search Ad (DSA) technology and automatically created assets, then packaged them into one simple toggle that you can turn on or off. When you turn it on, you’re giving Google permission to:
- Create AI generated headlines and descriptions using your landing page content (text customization)
- Send traffic to different pages it thinks will convert better (final url expansion)
- Tap into broad match keyword targeting, even if you’re running phrase match keywords or exact match keywords
The comprehensive AI Max suite also includes some interesting additions like geographical intent targeting at the ad group level, brand inclusions and brand exclusions (finally at campaign level), and URL controls. But honestly, most of this isn’t new — it’s just been repackaged with better (AI) branding.
Testing AI Max in a Real Account
Here’s what we’re seeing in a real campaign. This is a sectional furniture company where we’ve been running an existing Search campaign with AI Max turned on. We’re using both broad match and exact match in this account, so we were already tapping into Google’s AI for matching relevant searches.
To use AI Max, all you have to do is switch the toggle to on. It will then automatically set up text customization and final URL expansion.
You can turn these off, which will essentially mean AI Max is limited to tapping you into broad match keyword technology, but you don’t have to. Test to see which settings work best for your campaign.
When you activate AI Max, nothing dramatic happens initially. It doesn’t create a new Search campaign or build new ad groups. Instead, it runs in the background, and you have to actively check the search terms report to see what’s happening.
There’s a specific filter for “search terms and landing pages from AI Max” where you can see the AI Max section results.
Here’s where things get interesting.
AI Max: The Good, The Bad, and The Bizarre
The automatically created assets? Some are decent. Headlines like “Made in USA” and “Easy Clean Washable Fabrics” make sense. The asset optimization pulls from existing ads and landing page copy, which can sometimes work well.
But there are some changes which are questionable at best.
In this campaign, we found search terms like “home” triggering ads for a furniture company. No context, no intent — just the word “home.”
We had to add it as a negative keyword to stop wasting budget on impressions and clicks.
Further, AI Max matched this American-made furniture company with the search term “buy England furniture online.” The headline it served? “Made in USA.” That’s not user intent matching — that’s the complete opposite of what someone searching for England furniture wants to see.
These sorts of AI-driven search terms make us question whether Google’s AI Max works as advertised (at least, so far).
Our Honest Recommendations After Testing
After extensive testing across numerous accounts, here’s what we’ve found on when to test AI Max:
- Larger accounts already using broad match successfully: AI Max seems to perform okay here, though it doesn’t dramatically expand beyond what broad match was already delivering. The incremental performance is minimal, but it’s not actively hurting performance metrics either.
- Smaller accounts or those not using broad match yet: Avoid it. If you’re not successfully using broad match, AI Max for search campaigns probably won’t work well for you yet. Master broad match first, then consider testing.
- The Conservative Testing Approach: Run a 50-50 campaign experiment. Split your existing Search campaign traffic — 50% goes to your current structure without AI Max, 50% goes to the campaign experiment with AI Max settings enabled. This gives you deeper insights into actual performance (without risking your entire budget).
- Try to let it run for a few weeks without touching anything.
What to Monitor If You Test AI Max
If you decide to test, keep a close eye on the search terms report. Look at what landing page content Google is using, what ad copy the text customization is generating, and what search terms are triggering your ads. You’ll likely see at least a few searches that make zero sense for your business.
But here’s the key: don’t make rash decisions. Let it run for a few weeks to give AI Max time to calibrate and learn from customer ad journeys. That said, you can add obviously misaligned terms as negative keywords straight away.
On accounts where it does work, the first week or two are often choppy, then it finds its groove. Use it as a mining tool — take the successful search terms AI Max discovers and build them into your existing keywords and ad group structure. Don’t just let AI Max take over completely.
Final Thoughts
We’re not too impressed with AI Max yet, but that doesn’t mean you shouldn’t test it or that it’ll never work. Google often releases features that aren’t great initially but improve significantly over time. Performance Max was like this. Smart bidding was like this back in the day. Even broad match needed time to mature.
Google’s AI Max will probably be excellent a year from now, maybe sooner. Right now, it’s learning from advertiser data across millions of campaigns, and gathering those crucial data points that make machine learning work. With over a decade of Google internal data behind it, the potential is there.
But for now? It’s all about testing conservatively with sufficient budget to gather meaningful data. Monitor your campaign settings religiously. And have the risk tolerance to let it run for several weeks before pulling the plug.
