For twenty years, search advertising had one reassuringly simple principle at its core: someone typed a query, an advertiser bought a keyword, and Google decided whether the two were close enough to show an ad.
The machinery around that transaction became increasingly sophisticated, of course. Automated bidding replaced a lot of manual work. Broad match became broader. Smart Bidding learned which auctions were worth entering. Responsive Search Ads started assembling creatives dynamically. But even as Google automated almost everything around paid search, advertisers could still think in keywords.
That is becoming harder to do.
Key takeaways
- Since September 2026, eligible Search campaigns with automatically created assets or campaign-level broad match are being moved into AI Max. Dynamic Search Ads follow in February 2027.
- AI Max lets Google choose the query, adapt the ad text and change the landing page.
- Google reports an average of 7% more conversions at a similar CPA or ROAS, based on its own data.
- The competitive edge moves downstream: defining which traffic is actually valuable matters more than finding the keyword first.
What changed in September
This September, Google began moving eligible Search campaigns using Automatically Created Assets (now called text customization) and campaign-level broad match into AI Max for Search. Dynamic Search Ads are heading in the same direction, although Google has pushed their automatic migration back to February 2027.
Taken individually, none of these changes looks particularly dramatic. Put them together and the direction becomes much clearer.
Google is steadily moving Search away from a system where advertisers tell it exactly which queries to buy, which ad to show and which page to send the user to. AI Max allows Google's systems to make decisions across all three.
The keyword isn't disappearing. But it is losing some of its authority.
What AI Max actually does
AI Max can use an advertiser's existing keywords, ads and website content to find relevant searches beyond the original keyword list. Its text customization feature can generate or adapt creative to better fit a particular search. Final URL expansion can decide that the landing page selected by the advertiser isn't actually the best destination and send the user to another relevant page on the same domain instead.
In other words, Google is beginning to connect the query, the creative and the destination itself.
That's a meaningful change for anyone whose business depends on buying traffic.
The 7% figure, and what it costs
Google says advertisers using the full AI Max feature set (search term matching, text customization and Final URL expansion) saw an average 7% increase in conversions or conversion value at a similar CPA or ROAS compared with advertisers using search term matching alone. The figure comes from Google's own internal data for non-retail advertisers, so it shouldn't be read as a universal performance promise. Still, it explains why Google is pushing the product so aggressively.
The interesting part isn't really the 7%.
It's what advertisers are giving up in exchange for it.
A traditional performance marketer might spend hours identifying a profitable cluster of queries, write an ad specifically for those searches and build a landing page designed around exactly that intent. The economics of the campaign could then be measured relatively cleanly: this keyword produced this click, which went through this funnel and generated this conversion.
AI Max makes that chain more fluid.
The query may come from Google's expansion rather than the advertiser's original keyword list. The headline may be customized by Google's AI. The landing page may be different from the one initially entered into the campaign.
Google isn't simply optimizing the bid anymore. It is increasingly participating in constructing the customer journey.
Why performance marketing feels it first
For a large retailer with thousands of products, that's an obvious opportunity. If somebody searches for something buried three levels deep inside a website, Google's AI may be better at finding the right product page than a campaign manager maintaining thousands of keyword and URL combinations manually.
For performance marketing, things get more complicated.
Imagine you're buying insurance traffic. You know that one group of queries converts into qualified policy requests at €42, while another looks almost identical at the top of the funnel but produces customers worth half as much. Or you're buying loan traffic and the real KPI isn't the form submission but whether the applicant eventually gets approved. In iGaming, a registration and a depositing player are two very different things.
That distinction matters enormously when margins are tight.
More conversions don't necessarily mean better traffic.
And that may be the central challenge of AI-driven Search. The better Google becomes at finding additional traffic, the more sophisticated advertisers have to become at telling Google which traffic was actually valuable.
The advantage moves downstream
This is already changing the job of a search marketer.
For years, one of the advantages of a strong PPC team was its ability to discover profitable keywords before competitors did, organize them intelligently, build tightly matched ads and landing pages, and continuously remove traffic that didn't work.
Some of that craft isn't going away. But its relative importance is changing.
The competitive advantage increasingly sits downstream:
- Can you tell the difference between a lead and a good lead?
- Can you send offline conversion data back into the platform?
- Do you know the lifetime value of users acquired through different types of intent?
- Can you distinguish between a cheap conversion and a profitable customer?
If Google is going to make more of the acquisition decisions, advertisers need to become much better at defining the outcome.
Reporting becomes a necessity
There's another important development here that shouldn't be overlooked. Google is giving advertisers more visibility into what AI Max actually does.
The updated reporting can show which search terms came through AI Max expansion and whether that expansion originated from broader matching or keywordless technology. Advertisers can also inspect combinations of the search term, headline and landing page that Google actually served.
That sounds like a reporting feature. In reality, it's almost a necessity.
Once an advertising platform starts making decisions that used to belong to the advertiser, sophisticated buyers need to know what those decisions were.
What it means for affiliates and arbitrage
This is particularly important for affiliates and traffic arbitrage businesses.
Search arbitrage has always depended on understanding small differences in traffic value. Buy attention for X, monetize it for Y, and make sure Y remains comfortably higher than X. A seemingly minor change in query mix or user intent can destroy that equation very quickly.
If Google starts discovering more of those users itself, the old question, "Which keyword should we buy?", becomes less useful.
A better question may be: what kind of user are we actually trying to buy?
That's a much bigger shift than it sounds.
Keywords are convenient because they are visible. Intent isn't. Two people can type slightly different things and want exactly the same product. Two others can type almost identical queries and have completely different commercial value.
AI is supposed to become better at understanding that difference.
If it succeeds, the future of paid search may look less like buying words and more like buying outcomes.
Beyond Google: buyers want outcomes, not volume
And that development goes beyond Google.
Across performance advertising, buyers are gradually becoming less interested in traffic as a generic commodity. They want to know what kind of audience they're getting, where it originated, how it behaves and whether it produces actual business results.
Publishers face the same change from the opposite direction. Saying "we have finance traffic" or "we have 500,000 monthly users in Germany" isn't particularly useful if a buyer can't understand what that audience is worth.
This is one of the ideas behind Traffberry. The traffic marketplace is designed to make media supply easier to discover: advertisers can browse traffic offers across different GEOs, verticals and formats, or compare ad networks side by side, while publishers can post a listing and put their available traffic in front of buyers beyond their existing network.
As platforms like Google automate more of the buying decision, having more ways to discover and compare supply becomes increasingly useful. The technology used to find traffic may change. The commercial question doesn't.
Is this the right audience, at the right price, for the result I'm trying to achieve?
Keywords still matter, but the center of gravity is moving
For now, keywords still matter. Google Ads isn't suddenly becoming a keyword-free product, and experienced search marketers aren't about to throw away their keyword strategies.
But the center of gravity is moving.
The advertiser used to tell Google what traffic to buy.
Increasingly, the advertiser tells Google what result it wants and lets the machine work out how to get there.
That may turn out to be the much bigger change.
Sources
- Google: Dynamic Search Ads are upgrading to AI Max
- Google: Unlock next-level performance with AI Max for Search campaigns
- Google Ads Help: How AI Max for Search campaigns works
- Google Ads Help: Text customization frequently asked questions