How to Trial Google’s New AI Max Campaigns

Google’s rolling out another AI feature, this time it’s AI Max for Search campaigns that promise to provide much higher visibility and relevant ads for broader search terms. Google has shown promising results for big names who tested their new campaign, like L’Oréal and MyConnect, of up to 14% more leads for 16% less cost in their release blog.

But these are early days, and cherry-picked case studies from huge accounts don’t guarantee success for smaller spends. Like any new Google Ads feature, AI Max needs careful testing to determine if it’ll help or hinder your specific campaigns. We’ll walk you through exactly how to trial AI Max safely.

What is Google AI Max?

AI Max is an optional one-click setting that adds Google’s latest AI-powered targeting and creative features to your Search ads. Think of it as Google’s middle ground between traditional keyword campaigns and the fully automated Performance Max, more automation, but you keep some control.

Unlike regular Search ads that rely on your keyword lists, AI Max uses broader matching from your site content to potentially find searches you’re not currently capturing. It also dynamically adjusts ad headlines and landing page selection based on each user’s query.

When Will AI Max Be Available?

As of mid-2025, it’s available worldwide in beta to all advertisers. There’s no special access needed and no minimum spend requirements.

How Could AI Max Impact Your Google Ads?

Here’s what AI Max promises, though your mileage may vary:

Potentially expanded reach

AI Max’s broader matching might find queries your current keywords miss. The system could show your ads for long-tail and conversational searches you haven’t thought to target or wouldn’t make it through if you’re running exact match keywords. Whether these additional impressions convert is something you’ll need to test.

Dynamic ad customisation

The system adjusts ad headlines and landing pages to match user intent. Google’s data suggests this could drive 14% more conversions at similar CPA, with campaigns using only exact or phrase match keywords potentially seeing up to 27% more conversions with AI Max added. These are Google’s numbers though, your results could differ significantly.

Smarter landing page selection

Here’s where AI Max gets interesting for content-rich websites. The system doesn’t just pick from your designated landing pages, it can scan your entire site and send traffic to the most relevant content. This means:

• Blog posts could become landing pages for informational queries
• FAQ pages might capture “how to” searches
• Product subcategories could match specific feature searches
• Case studies might serve comparison shoppers
• Location pages could target local intent queries

For example, if someone searches “small business website design process Perth”, AI Max might send them to your detailed blog post about web design timelines rather than your generic web design service page. This could improve relevance and quality scores, or it could send valuable traffic to pages that don’t convert well.

The upside? Your existing content library becomes a potential asset for capturing diverse search intents.

The risk? You might need to audit your entire site for conversion optimisation, not just your main service pages.

Retained controls (mostly)

Unlike Performance Max, AI Max keeps some Search campaign controls:
– Negative keywords still work
– Search term reports remain available
– Location targeting and brand exclusions stay active
– You can remove auto-generated headlines

That said, you’re still giving Google more control than with traditional Search campaigns.

Is AI Max Worth Testing for Your Business?

Remember, AI Max means giving Google more control. For specialised services with narrow audiences, the broader approach could waste budget on irrelevant clicks. While you retain some guardrails (negatives, geographic controls), you’re still trusting Google’s algorithm more than before.

How to Trial AI Max in Your Account (Step-by-Step)

Here’s how to test AI Max without gambling the farm:

Step 1: Pick the Right Test Campaign

Choose a mid-performing Search campaign, definitely not your biggest revenue driver. This limits risk while still providing meaningful data. You need a campaign with enough volume to generate statistically significant results within a reasonable timeframe.

Before starting, document your baseline metrics: CPA, ROAS, conversion rate, and impression share. You’ll need these for comparison.

Step 2: Enable AI Max with an Experiment

Google’s Experiments feature makes setup straightforward:

1. Navigate to Experiments in your Google Ads account
2. Click New experiment > Campaign Features
3. Select AI Max from the options
4. Choose your test campaign

The system splits traffic 50/50 between Control (AI Max off) and Trial (AI Max on) within the same campaign. This provides a fair comparison without the hassle of duplicating campaigns.

The experiment activates both Search Term Matching and Asset Optimisation by default. You could disable either to test specific elements, but for a proper AI Max trial, leave both enabled initially.

Step 3: Run, Measure, Decide

Run the experiment for at least 4 weeks or one full conversion cycle, longer if volume is low. Monitor these metrics closely:
– Conversion rate changes
– Cost per acquisition movements (did you get more leads of lower quality)
– Search term quality and relevance
– Overall account performance (watch for cannibalisation)

Check the “AI Max” match type in search terms reports weekly. Look for concerning patterns: irrelevant queries, geographic mismatches, or brand term confusion.

After gathering sufficient data:
– If metrics improve: Consider expanding the test or applying to the campaign.
– If results are mixed: Dig deeper into what’s working and what isn’t.
– If performance drops: End the experiment and analyse what went wrong.

Don’t just look at the experiment in isolation, monitor whether AI Max impacts other campaigns through competition for the same queries.

Tips to Monitor Your AI Max Trial

• Stay vigilant – Check performance every few days initially. If spend accelerates unexpectedly or irrelevant queries flood in, pause and reassess. There’s no shame in ending an experiment early if it’s clearly not working.
• Review auto-generated assets -AI Max creates headlines automatically. Check these regularly and remove any that miss the mark or sound generic. Poor ad copy could tank performance regardless of targeting improvements.
• Watch your budget distribution – The 50/50 split means each arm gets half your daily budget. Ensure both have enough to generate meaningful data – $10/day per arm won’t tell you much.
• Monitor broader impacts – Check if AI Max affects your other campaigns. Are CPCs rising elsewhere? Is impression share dropping on priority campaigns? The experiment might have ripple effects.
• Document everything – Keep notes on what you observe – unusual queries, performance fluctuations, competitive changes. This context helps interpret results beyond just the numbers.

The Bottom Line on AI Max

AI Max is another Google automation feature that could potentially improve performance – or could waste budget on irrelevant traffic. Like automated Google Ads features before it, success depends entirely on your specific situation.

If you have the volume and budget flexibility, it’s worth a controlled test. Approach it like any experiment: hypothesis, test, measure, conclude.

Don’t assume it’ll work just because Google’s case studies look impressive.

The reality is Google will continue pushing automation whether we’re ready or not. Those who test methodically and make data-driven decisions will adapt better than those who either blindly adopt or stubbornly resist every change. For guidance on improving your Google Ads ROI through strategic testing, the fundamentals still apply.

Thinking about testing AI Max but want expert eyes on the experiment? Our Google Ads management team can help design a proper test, monitor for red flags, and interpret results objectively.

Request a quote

Zac Fitzpatrick

PPC Service Lead