Google tests PMax controls as AI ads face buyer pressure
Google is testing new Performance Max controls as advertisers push for more transparency and steerable AI advertising.
Google is testing a new way for advertisers to reclaim some control inside Performance Max, its automated campaign product for running ads across search, display, YouTube and other Google inventory.
The pilot lets selected media buyers exclude inventory from third-party search partners and Google Display Network. The feature appears as two checkbox options inside PMax consoles, enabled by default but removable by advertisers who want tighter control over where spend flows.
That may sound like a modest product change. Strategically, it points to a larger shift in AI advertising: platforms can keep pushing automation, but they increasingly need to prove that automation is steerable.
Table of contents
Jump to each section:
- Why PMax control matters now
- What Google is testing inside Performance Max
- The strategic tension behind automated media
- What marketers should know about PMax controls
Why PMax control matters now
Performance Max has always asked advertisers to accept a tradeoff. In exchange for scale and machine-led optimization, marketers give up some visibility into exactly how the system routes budget across placements.
That bargain becomes harder to defend as AI ad tools mature. Buyers are no longer asking whether automation can improve campaign execution. They are asking whether the automation gives them enough evidence, control and confidence to justify deeper spend.
The important signal is not that Google is abandoning automation. It is doing the opposite. The company is showing that automated media products may need more human steering, not less, if advertisers are expected to keep trusting them.
Patrecia Meliana
What Google is testing inside Performance Max
The test gives selected advertisers the option to remove Search Partners and Google Display Network from PMax campaigns. Those inventory sources have long been bundled into the product, which created frustration among some search buyers who wanted clearer separation between mainline search, display and third-party partner inventory.
Several buyers described the change as meaningful because it adds a lever that PMax historically lacked. It follows earlier updates that gave advertisers more reporting visibility, negative keyword options and first-party audience exclusions.
Google has described the feature as a limited pilot. Some media buyers in the United States and United Kingdom have access, while others have requested access and been denied.
That uneven rollout matters. A control feature is only strategically useful when marketers know whether it is a test, a coming default, or a privilege granted to certain accounts. Until Google clarifies the path to broader availability, advertisers will treat the pilot as both a useful signal and an incomplete answer.
The strategic tension behind automated media
The common assumption is that better AI makes human intervention less necessary. The reality inside paid media is more complicated. As automated systems absorb more campaign decisions, the value of human oversight shifts from manual execution to setting boundaries.
PMax shows the tension clearly. Advertisers may want machine learning to optimize bids, audiences and placements, but they still want the ability to reject inventory they consider misaligned with client goals. Automation can find efficiency. Marketers still need to define acceptable risk.
That distinction matters because AI ad platforms are becoming less like tools and more like operating environments. Once a campaign system decides where budget moves, what inventory counts as valuable and how performance is attributed, platform design becomes media strategy.
Control is becoming a trust feature.
This is also where competitive pressure enters the story. Google faces a changing search advertising market as AI-native ad products develop around conversational interfaces and agent-like experiences. Giving advertisers more practical control over PMax may be less about reversing course and more about protecting confidence in Google Ads while the search category changes around it.
What marketers should know about PMax controls
For marketing teams, the lesson is not to reject automated campaign systems. It is to become more precise about where automation should be allowed to act freely and where it needs constraints.
Inventory quality needs explicit governance. If PMax controls become more widely available, brands should decide in advance which placements are acceptable, which need testing and which should be excluded by default.
Measurement has to separate platform lift from inventory mix. When campaign results improve, marketers need to understand whether gains came from better automation, broader inventory, stronger creative, or more aggressive budget routing.
Agency value moves toward interpretation. Buyers who understand how automated systems behave across inventory types will be better positioned to advise clients than teams that only monitor output-level performance.
Platform trust will depend on configurable automation. The next stage of AI advertising will not be defined only by how much platforms can automate. It will be defined by how responsibly marketers can shape the automation before spend scales.
The broader implication is that AI media buying is entering a more mature phase. The first wave rewarded platforms that could simplify campaign setup and absorb complexity. The next wave will reward systems that expose enough of that complexity for professionals to manage risk.
Google’s PMax pilot is small in interface terms, but it speaks to a bigger question for every AI ad product: can the system be powerful without becoming opaque?
For marketers, that question should sit at the center of every platform conversation. The future of AI advertising will not belong to full automation or full manual control. It will belong to teams that know which decisions to delegate, which to constrain and which to keep close to the brand.
