AI Max for Search: When Google’s Black Box Actually Delivered (...Sometimes)

After testing Google’s new AI Max across clients, we’re sharing what worked, what didn’t, and how marketers can make the most of it.

TL;DR

We rolled out AI Max for Search across campaigns for clients in multiple verticals, including eCommerce, online education, and nonprofit. In most cases, it captured 5–10% of Search spend, produced ~30% stronger ROAS than non‑AI Max traffic, and uncovered incremental volume. However, it also proved to “cherry-pick” volume from existing keywords, which masked some inefficient spend, emphasizing the need for strong conversion signals and constant guardrails.

Here’s what to watch for — and how you can stack the odds in your favor.

What Is AI Max? 

AI Max is a new feature layer in Google Search campaigns that lets you combine up to three AI-driven capabilities. Search Term Matching is required at minimum when it’s enabled, while the other features are optional.

1. Search Term Matching (Keywordless & Broad Expansion): Match to queries beyond your keywords using context from assets, landing pages, and existing keywords.

2. Text Customization: Dynamically generates or customizes headlines and descriptions.

3. Final URL Expansion: Routes users to what it deems the most relevant page on your site.

Unlike Performance Max, AI Max came out of the gate with some reporting transparency, surfacing 'AI Max' as its own match type in Search Term & Keyword reports.

Why We Decided to Test It

1. The potential for incremental conversions beyond our existing keyword sets, including an improved ability for ads to appear in AI Overviews

2. To see if AI could optimize landing page routing better than static URL selection

3. The need to future‑proof: in new campaigns, AI Max is now required for brand inclusion/exclusion controls

What We Found

AI Max consistently captured 5–10% of Search spend and in most cases:

• CPCs were lower – likely from tapping into broader, less competitive search terms

• ROAS exceeded non‑AI Max traffic by ~30%

• Total campaign conversion volume increased anywhere from 5-100% depending on the prevalence of broad match keywords in the campaign

• Minimal improvement of landing page routing

Digging Deeper

While these top-line results were strong, the question remained: does AI Max “cherry-pick” conversion volume from existing keywords, and if so, was that masking other inefficient spend brought on by AI Max?

Search-term analysis showed that much of AI Max’s conversion volume came from short- and mid-tail queries that overlapped with existing keywords. The gains were real — lower CPCs and modest incremental volume — but required rethinking campaign structures, ad-group targets, and negatives to harness AI Max effectively. Some examples of where changes were necessary:

• Client A (Strong AI Max Performance): AI Max capitalized on generic non-brand terms closely related to the brand name. We expanded negatives and added in broad branded terms to intentionally capture this overlap in our brand campaign, which typically has a higher target ROAS.

• Client B (Weaker AI Max Performance): AI Max spent on queries tied to keywords that we’d previously avoided having broad match versions of due to poor performance. Our next steps are to refine ad-group targets and/or further adjust conversion values to reflect true downstream quality.

Ultimately, clients with the strongest conversion signals achieved the best AI Max results, underscoring the value of robust, down-funnel conversion data in an AI-powered advertising landscape.

Lessons Learned & Recommendations

1. Strong input signals lead to stronger results. Enable AI Max where your structure is mature and conversion signals are strong.

2. Don’t trust it blindly. Search term audits are necessary. Segment and evaluate the AI Max match type separately.

3. Cherry-picking can be okay. AI Max can bring in existing search traffic at lower CPCs, so cherry picking is okay, as long as it doesn’t mask other inefficient spend.

The Risks of Not Testing AI Max

• You risk missing auction-level efficiencies and new demand.

• Some account features require AI Max.

• Google continues to push toward AI-driven search campaigns.

Conclusion

AI Max is neither a miracle nor a disaster. It’s an evolving tool that responds to the quality of the signals you provide. Done right, it added revenue efficiently across most of our test accounts. Done carelessly, it has the potential to seek out low‑value traffic.

If you’re ready to explore AI Max but want a partner to navigate its blind spots, reach out to us here at Working Planet!