Estimated reading time: 9 minutes
The reality, as always in Google Ads, is in between and depends on how you manage it.
This post is not written for the account manager who is going to set it up. It is written for the CEO or CMO who has to decide whether to give the OK, what questions to ask their team or agency, and what budget risk they are taking. No unnecessary technicalities. With the facts on the table.
The underlying problem: the search you knew has changed
For more than a decade, Google Ads in Search worked with a clear logic: you define keywords, you bid, you control which searches trigger your ad. It was predictable. It was auditable. And it was the kind of channel a CFO could understand in five minutes by looking at a spreadsheet.
That gradually came to an end. First came Smart Bidding, which delegated bidding to AI. Then came responsive search ads, which delegated text to AI. Now comes AI Max Google Ads, which delegates to AI the three main levers of a Search campaign at the same time: what searches your ad responds to, what that ad says, and what page on your website it takes the user to.
It is not a cosmetic update. It’s a model change. And it directly affects how your budget is spent.
What is AI Max for Search, without technicalities
AI Max for Search is not a new type of campaign. It’s a switch, or rather a set of switches, that you can turn on within your existing search campaigns. When you turn it on, you give Google’s AI control over three aspects that were previously handled by you or your agency:
- First, search term matching. Your keyword list becomes a reference, not an instruction. The AI analyzes the context, the user’s history and the intent behind each query to show your ad in searches that you didn’t explicitly include. If you sell project management software, your ad might show up in searches like “how to organize my remote team” even though that phrase is not in your list.
- Second, text personalization. AI generates headlines and descriptions by combining your web content, your keywords and your existing ads, adjusting the message in real time to make it as relevant as possible for each specific search.
- Third, destination URL expansion. Instead of always taking the user to the landing page you specify, the AI can redirect them to any other page on your site if it thinks it’s a better fit for what they’re looking for.
In short: AI Max autonomously optimizes which searches trigger your ad, what that ad says, and where the user ends up. The question is not whether this is technically impressive. The question is whether it makes sense for your particular business.
The real data: what works and where the risk lies
Official Google figures show an average increase of 14% in conversions while maintaining the same CPA. In campaigns that worked with exact or phrase matching, the improvement can reach 27%. L’Oréal doubled its conversion rate and reduced CPA by 31%. Australian company MyConnect generated 16% more leads with a 13% lower cost per acquisition.
These are the cases that Google publishes on its Think with Google website. And they are real. But there’s another side to the story that doesn’t appear in the press releases.
An independent analysis published in Search Engine Land of more than 250 retail campaigns found that AI Max delivered a 35% lower ROAS than conventional campaigns. Another four-month test showed a 90% higher CPA over phrase-matching campaigns. The technical explanation is that AI finds more traffic volume, but some of that traffic has lower purchase intent or lower average order value.
Put bluntly: AI Max tends to bring more conversions in volume, but not always better conversions in quality or profitability. And that difference matters a lot depending on whether your business lives on qualified leads or transactional volume.
Three decisions to make before activating it
If your agency or internal team has presented AI Max to you as “the new Google thing we need to activate now,” there are three questions you should ask yourself before giving the green light.
1. How do you measure the quality of what comes in, not just the volume?
AI Max optimizes for the conversions you have registered. If you only have “form submission” set as a conversion, the AI will look for more form submissions, regardless of whether those leads convert into customers or not. For AI Max to work well in B2B or high-ticket industries, you need to have fed it with qualified conversion data: calls that closed sales, leads that advanced in the pipeline, customers that bought.
Without that quality information, you are giving the AI the steering wheel but hiding the destination from it.
2. Do you have brand controls and messages configured?
Google has included editorial control tools in AI Max that many advertisers don’t set by default. Text guidelines allow you to set which words should never appear in your ads, and branding instructions allow you to give the AI natural language guidelines about your company’s tone and priority messages. Without setting these, the AI can generate text that is perfectly functional but inconsistent with your positioning.
A manager of a financial services company who activates AI Max without these restrictions may encounter ads where the tone is too aggressive or where promises are made that the sales team then has to qualify.
3. Are you prepared to test, not to implement blind?
Google offers the option to activate AI Max as a 50% split traffic experiment, so you can measure the real impact on your account before migrating the entire campaign. It’s the smart way to go. Any agency that proposes you to activate AI Max directly in production, without a prior experiment, either doesn’t have enough experience with the tool or is prioritizing novelty over results.
Always ask for the experiment. Four weeks of real data is worth more than any Google success story.
When it makes sense to activate AI Max and when not to activate it
AI Max works best when you have mature campaigns with solid conversion history, when you are already using Smart Bidding and your campaigns have passed the learning curve, when your industry search volume is high and there is room for audience expansion, and when your product or service has relevance across multiple search variants (different ways of expressing the same need).
AI Max has more risks when the business is very specific and irrelevant traffic has a high cost (legal, financial, industrial services), when the conversions recorded do not reflect real commercial quality, or when the budget is tight and there is no margin to absorb the cost of learning.
In sectors such as general retail, online training or SaaS software with good conversion tracking structure, the data shows more consistently positive results. In B2B lead generation with long sales cycles, implementation requires more monitoring and well-fed qualified conversion data.
The connection to AI Overviews: why this is more urgent than it seems
There is a context that makes AI Max more strategically relevant than simple campaign optimization. Google is progressively integrating Search ads into its generative AI experiences: the AI Overviews that already appear in many searches and the AI Mode it is rolling out in English-speaking markets.
Ads appearing in these generative experiences require campaigns to be set up to operate on AI logic, not keyword logic. Put bluntly, traditional search campaigns have less and less coverage in the results where Google is investing its development. AI Max is, in part, the compatibility mechanism between your current advertising and the search of the near future.
This does not mean that you should activate it tomorrow without preparation. It means that ignoring it indefinitely has an opportunity cost that will grow.
The question you should ask your agency this week
If they haven’t already done it to you, do it yourself: What experiment structure do you propose to test AI Max on our account, what success metrics will you measure beyond raw conversions, and when will you present the results to me?
A good response includes: defined test period (minimum 4-6 weeks), quality metrics in addition to volume (CPA, ROAS, average conversion value), pre-launch brand guidelines configuration, and a proposal of which campaigns are candidates and which are not.
If the answer is “let’s activate it and see how it goes”, you have a problem that goes beyond AI Max.
AI Max Google Ads does not change the objective of your campaigns: to keep getting profitable customers. What changes is who makes the tactical decisions to get there. How much control you give up and what data AI backs up that decision with is what determines whether it works for your business.
Frequently asked questions about AI Max for Search
No. They are different tools with different objectives. Performance Max operates on all Google channels (Display, YouTube, Maps, Gmail, Search). AI Max for Search is exclusively for search campaigns and offers more granularity of control. They can coexist in the same account and in fact Google recommends a combined strategy it calls Power Pack: Demand Gen + Performance Max + AI Max for Search.
Yes, you can enable search term matching without enabling destination URL expansion, for example. This allows for a gradual and more controlled implementation, starting with lower risk functionalities.
AI Max’s best results occur in accounts with sufficient data history. With very small budgets or recent campaigns with no conversion history, AI has little to work with and results are less predictable. In those cases, consolidating conversion history first is the priority.
There is no official confirmation, but Google’s historical trend is to push towards more automation with every update. The same happened with expanded search ads (which disappeared in favor of responsive) and Smart Bidding (which became the de facto standard). Preparing to operate with AI Max is preparing for the Google Ads of the next cycle.
Especialista en SEO y Paid Media | Google Ads, Meta Ads, LinkedIn Ads & Search Console en vena Optimizo visibilidad orgánica + escalo adquisición pagada con ROAS obsesivo y estrategias data-driven.
Especialista en Performance Digital, Core Web Vitals, E-E-A-T, algoritmos de subasta, attribution multi-touch y maximizar LTV/CAC.
