Google AI Max for Search: When It Works, When It Wastes Budget, and How to Decide
Google says AI Max delivers 14% more conversions. Independent analysis tells a different story. Here’s what actually determines whether it works for your account.
Google says AI Max delivers 14% more conversions at a similar CPA. Independent analysis of 250+ campaigns tells a different story: +13% revenue but +16% higher cost per acquisition. One large-scale test found that 99% of AI Max impressions generated zero conversions across 30,000 search terms. So who is right? Both – and the difference comes down to whether your account has the foundation to make AI Max work.
We have tested AI Max across multiple client verticals since its beta launch in 2025. This guide breaks down exactly when AI Max for Search campaigns delivers real value, when it burns budget, and how to decide before you flip the switch.
What Is AI Max for Search Campaigns?
AI Max for Search is not a new campaign type. It is a one-click enhancement layer you apply to your existing Search campaigns. Google announced it at Google Marketing Live in May 2025, and it has been rolling out globally since late Q3 2025.
It combines three features that previously existed in isolation:
Search Term Matching
expands your keyword reach beyond what you have explicitly targeted. It uses broad match plus keywordless technology to find queries Google predicts will convert – even if those queries do not match any keyword in your account. Think of it as broad match on steroids, powered by signals from your ad creative, landing pages, and conversion history.
Text Customisation
(formerly Automatically Created Assets) uses AI to generate headlines and descriptions dynamically. It pulls from your existing ad copy and landing page content to assemble ads tailored to each user’s query. The goal is higher ad relevance without you writing dozens of ad variations manually.
Final URL Expansion
replaces your ad’s destination URL with whatever landing page Google’s algorithm determines is most relevant to the user’s search. Instead of sending everyone to your chosen page, it routes traffic across your domain based on query intent.
Together, these three features give Google significantly more control over who sees your ads, what those ads say, and where clicks land on your website.
How AI Max Differs from Performance Max
This is the most common question we hear. Performance Max is a cross-channel campaign type that runs ads across Search, YouTube, Gmail, Display, Maps, and Discovery – all from a single campaign. You hand Google a budget and creative assets, and it decides where to place them.
AI Max for Search stays within the Search network only. You keep your existing campaign structure, keywords, and ad groups. AI Max enhances what is already there rather than replacing it. You also get significantly better reporting – including a search terms report that shows “AI Max” as a match type, plus visibility into which headlines, descriptions, and URLs were served for each query.
For advertisers who want AI-powered optimisation without surrendering visibility into what is actually happening, AI Max offers a middle ground that Performance Max does not.
What Google’s Published Results Show
Google’s official data claims that advertisers activating AI Max typically see 14% more conversions or conversion value at a similar CPA or ROAS. For campaigns still heavily using exact and phrase match keywords, the typical uplift rises to 27%.
Google has published case studies showing enterprise brands achieving dramatic results: conversion increases of 160% to 260%, CPA reductions of 22% to 73%, and conversion rate improvements of 2X or more.
These numbers are real, but they come with a critical asterisk. These are enterprise accounts with massive budgets, clean conversion tracking, deep website architecture, and dedicated teams monitoring performance daily. Google’s 14% benchmark also conspicuously excludes retail – a significant omission for ecommerce advertisers.
What Independent Testing Actually Reveals
The gap between Google’s marketing and real-world results is significant.
- An independent analysis of 250+ campaigns found that while median revenue increased by 13%, median CPA also increased by 16%. The ROAS range swung wildly from +42% to -35%, meaning AI Max is essentially a coin toss for many accounts. You may see a revenue lift, but efficiency often does not follow.
- A separate four-month test comparing AI Max against traditional match types found AI Max cost $100.37 per conversion compared to $43.97 for phrase match and $52.69 for exact match. AI Max lost to every single match type tested, including close variants that are typically criticised for poor performance.
- Large-scale testing across approximately 30,000 search terms found that 99% of AI Max impressions generated zero conversions. Less than half of search terms showed they were matched to an actual keyword, meaning the majority functioned as keywordless ads without clear targeting logic.
- Perhaps most concerning: up to 63% of queries AI Max captured were not new at all – they were cannibalised from existing keyword coverage. And in one documented case, AI Max funnelled 69% of total Search impressions into competitor brand terms, generating over 8,000 impressions on keywords the advertiser had intentionally avoided.
When AI Max for Search Actually Works: 5 Use Cases
Based on our testing and the available data, AI Max delivers genuine value in these specific scenarios:
- 1. Large Accounts with 100+ Monthly Conversions and Clean Tracking AI Max’s machine learning needs data to optimize. The algorithm requires a meaningful volume of conversion signals to identify patterns and make good decisions. Accounts with fewer than 100 monthly conversions simply do not generate enough signal for the system to learn effectively. If your conversion tracking is accurate, deduplicated, and focused on real business outcomes – not inflated micro-conversions – AI Max has the raw material it needs to work.
- 2. Campaigns Still Relying on Exact and Phrase Match That Need Expansion This is where Google’s 27% uplift figure comes from, and it makes sense. If you have been running tightly controlled exact and phrase match campaigns, you are likely missing relevant queries. AI Max’s search term matching can surface high-intent long-tail queries you never thought to bid on. The key is that your existing campaign provides a strong performance baseline for the algorithm to learn from.
- 3. Lead Generation Accounts Wasting Performance Max Budget on Display and YouTube Many lead gen advertisers run Performance Max but find that most of their budget gets consumed by Display and YouTube placements that generate impressions but not qualified leads. AI Max for Search lets you capture the AI-powered query expansion benefits while staying exclusively on the Search network, where bottom-of-funnel intent is strongest.
- 4. Accounts with Deep, Well-Structured Websites Final URL Expansion is AI Max’s most underrated feature – but only if you have the landing page inventory to support it. Advertisers with dozens or hundreds of well-optimised, conversion-ready pages give the algorithm meaningful options. If you have five landing pages total, Final URL Expansion has nowhere useful to send traffic.
- 5. Brands Already Succeeding with Broad Match and Smart Bidding AI Max builds on the same foundational infrastructure as broad match and Smart Bidding. If your account is already performing well with these features – meaning you have simplified your campaign structure, consolidated ad groups, and let the algorithm optimize at auction time – AI Max is a natural extension. It amplifies what is already working.
When AI Max Is Wasted Budget: 5 Scenarios
These are the situations where we consistently see AI Max underperform or actively harm results:
- 1. Small Budgets Under $750 Per Day Google’s official minimum is $50 per day, but real-world testing shows that $750 per day is the threshold where AI Max begins to perform consistently. Below that, the algorithm cannot participate in enough auctions to learn, and budget-restricted campaigns lose massive impression share. We have seen campaigns losing 28% to 52% of impression share due to budget constraints alone – turning on AI Max in that environment just accelerates wasteful spending.
- 2. Niche or Compliance-Heavy Verticals Legal services, healthcare, financial services, and other regulated industries face a specific problem with AI Max. Text Customisation can generate ad copy that violates compliance requirements or makes claims you cannot legally support. Search Term Matching can expand into queries where advertising is restricted or inappropriate. If your industry requires strict control over ad messaging and targeting, the automation that powers AI Max becomes a liability.
- 3. Accounts with Broken or Inflated Conversion Tracking AI Max optimises toward whatever you define as a conversion. If your tracking is inaccurate, counting duplicates, or inflated with low-value micro-conversions like page views or scroll depth, AI Max will optimise enthusiastically toward the wrong outcomes. We see this frequently: the algorithm reports strong conversion numbers while actual lead quality or revenue craters. Fix your tracking before you touch AI Max.
- 4. Limited Landing Page Inventory If your website has a handful of pages, Final URL Expansion has nowhere useful to route traffic. Worse, it may send users to pages that are not optimised for conversion – your About page, a blog post, or a resource that lacks a clear call to action. AI Max’s Final URL feature shines on deep, well-structured sites. On thin sites, it is a liability you should disable.
- 5. Brands That Cannot Afford Competitor Term Leakage AI Max’s search term matching aggressively expands into competitor brand queries. The system sees competitor searches as “available pockets of traffic” because the products or services are similar. If your brand cannot afford to appear alongside competitor names – whether for strategic, legal, or budget reasons – this behaviour can consume the majority of your impressions on terms that rarely convert. Negative keyword lists help, but they require constant maintenance as AI Max continuously discovers new query variations.
The Foundation You Need Before Turning It On
AI Max amplifies whatever is already in your account – good or bad. Before enabling it, make sure these foundations are in place:
Account Structure Simplification
Consolidate traffic into fewer, larger ad groups and campaigns organised by business objective, not by match type, device, or audience segment. Smart Bidding handles those signals at auction time. Granular structures create data fragmentation that starves the algorithm.
Conversion Tracking Hygiene
Ensure your conversions are accurate, deduplicated, and measuring real business outcomes. Enable Enhanced Conversions for better signal quality. If you are tracking phone calls, make sure call duration thresholds are meaningful.
Budget Sufficiency
Your tCPA and tROAS campaigns must not be budget-restricted. Budget-constrained campaigns cannot maximise volume within your target, and adding AI Max to a budget-starved campaign just wastes the additional reach on queries you cannot afford to serve.
Landing Page Quality
Every page AI Max might route traffic to should have a clear value proposition, strong calls to action, and fast load times. Audit your site before enabling Final URL Expansion.
Negative Keyword Lists
Build comprehensive negative keyword lists before launch, not after. Include competitor brand terms, irrelevant service categories, and informational queries that do not convert. Review and update weekly.
ScaledOn Recommendation: How to Test AI Max Without Burning Budget
Testing AI Max requires discipline. Here is the approach we use with clients:
- Start with a single high-volume campaign. Do not enable AI Max account-wide. Pick your strongest-performing campaign with the most conversion data and test there first.
- Enable all three features simultaneously. Independent testing across 23 campaigns found that using all three AI Max features together delivered 40% higher uplift compared to activating search term matching alone. Partial adoption underperforms.
- Run a 50/50 campaign experiment. Use Google’s built-in experiment tool to split traffic between your existing campaign settings and an AI Max-enabled version. This gives you a clean comparison without risking your entire budget.
- Set up competitor brand exclusions from day one. Do not wait to see competitor terms in your search terms report. Proactively add competitor brand names as negative keywords before launch.
- Monitor search terms weekly for the first month. AI Max generates massive search term reports – sometimes tens of thousands of rows. Review them weekly and add negatives aggressively. After the first month, bi-weekly monitoring is usually sufficient.
- Give it time, but set a kill switch. The algorithm needs two to four weeks to calibrate. Expect choppy performance in the first week or two. But if CPA is more than 25% above your target after four weeks, pause and reassess. Do not let it run indefinitely on hope.
- Watch for Search Partner Network waste. One documented case found a campaign generating 500,000 monthly impressions on the Search Partner Network with a 0.07% conversion rate versus 3.04% on standard Google Search. Check your network placement reports.
The New Search Landscape: Why This Matters Now
AI Max is not just about campaign optimization. It is Google’s answer to how search itself is changing.
Longer queries are growing 1.5x faster than shorter queries. Google processes over 5 trillion searches annually. AI Overviews now reach 1 billion monthly users. Voice and image searches hit 20 billion monthly.
Google is placing ads above, below, and within AI Overviews, and has begun experimenting with ads inside AI Mode. AI Max’s search term matching and text customisation are designed specifically to capture these longer, more conversational queries that traditional keyword targeting misses.
The advertisers who build the foundation now – clean tracking, simplified structures, sufficient budgets – will be positioned to capture demand as search behaviour continues shifting. Those who ignore these changes will find their traditional keyword strategies covering a smaller and smaller share of actual search activity.
The Bottom Line
- AI Max for Search is a powerful tool, but it is not a universal one.
- Google’s incentive is to get every advertiser to enable it – more automation means more ad inventory means more revenue. Your incentive is different: profitable growth.
- After 18 years of managing performance marketing campaigns, our position is straightforward: AI Max is worth testing if your account meets the threshold criteria outlined in this guide.
- If it does not, invest in building that foundation first. The feature will still be there when you are ready, and it will work significantly better on a solid base.
Need help determining if AI Max is right for your Google Ads account?
Contact ScaledOn for a free account audit and personalised recommendation.
👉 Contact ScaledOn