Key takeaways
- AI Max testing now supports multi-campaign A/B tests across budgets and ROI targets, which makes test design more important than platform enthusiasm.
- Brand and location controls staying active during experiments gives UK advertisers a safer way to test automation without dropping essential guardrails.
- Performance Planner will make scaling routes more visible, but forecasts need commercial checks before one-click changes go live.
- Advertisers should segment tests by margin, lead quality and intent, not by campaign convenience.
- The winning rollout tactic is to fix conversion data first, set one business success metric, then scale only when incremental value holds.
AI Max testing has moved from a neat experiment feature into a serious budget decision tool. For UK advertisers, that changes the conversation from “should we try it?” to “what proof do we need before we let automation take more spend?”
The accounts that will benefit are not the ones that switch everything on and hope. They are the ones with clean conversion data, sensible campaign grouping and tight commercial targets. If you have not already audited how automation changes query matching, creative coverage and bidding behaviour, our take on AI Max migration risks is the right starting point before this new layer of testing arrives.
Google is giving advertisers better experiment and planning controls. Good. But better controls do not remove the need for judgement. They move the hard work upstream, into test design.
What’s actually changed in AI Max testing
Google has announced new AI Max testing and planning tools for Search campaigns. The headline change is multi-campaign experimentation. Instead of running isolated tests one campaign at a time, advertisers will be able to test different budgets and ROI targets across multiple Search campaigns within one A/B test.
Google is also adding support for experiments where brand and location controls remain enabled. That matters because many advertisers previously had to choose between testing AI Max and protecting guardrails they rely on for compliance, franchise coverage, local relevance or brand defence.
The final change sits in Performance Planner. Advertisers will be able to forecast how bidding and budget target changes affect existing campaign performance, then apply suggested changes directly to campaigns in one click. That is faster, but it also raises the risk of turning a forecast into live spend without enough commercial scrutiny.
Why this matters for advertisers
AI Max testing changes where budget mistakes happen. The danger is no longer only that automation enters more auctions. The danger is that a poorly designed experiment gives you a convincing answer to the wrong question.
Here is the mechanism. If you test AI Max across several Search campaigns with different budget caps, ROAS targets or CPA targets, Google will redistribute opportunity based on the signals you give it. A campaign with stronger conversion volume and looser efficiency targets will draw more eligible traffic. A campaign with thin data, slow conversion lag or weak landing page relevance will look less scalable. The test result then reflects your data quality and target structure as much as the value of AI Max itself.
That is why commercial segmentation has to come before platform setup. A lead generation advertiser should not group high-margin, sales-qualified campaigns with campaigns that count every form fill as equal. An ecommerce advertiser should not mix low-margin accessories with high-margin core products and then judge the result on total revenue. Automation will optimise towards the conversion definition you provide. If that definition is lazy, spend follows the wrong signal.
Budget forecasting also deserves a harder look. In our accounts, capped budgets quietly cost more clicks than most advertisers assume, and the AI Max planning tools will surface those ceilings far more often. That is an opportunity, not a spend-more nag. The point is knowing which missing clicks are worth chasing, not raising budgets blindly. If your click-through rate is soft on the terms you actually want to scale, fix that first: our guide on how to improve click through rates covers the levers that matter.
The money moves through three routes. First, expanded matching increases eligible impressions. Second, Smart Bidding changes bids as it sees new conversion probability patterns. Third, budget changes decide whether that extra eligibility becomes actual spend. If you loosen all three at once, you will not know which lever caused the result.
That is why UK PPC teams need stricter test boundaries. A clean testing plan isolates one meaningful commercial question: can we increase profitable conversion value, or qualified lead volume, without lowering quality? If the test is built around clicks, impressions or platform-reported conversions alone, it will reward volume before profit.
PPC Geeks’ View
The specific problem advertisers will face is false confidence from multi-campaign test results. A combined experiment looks efficient in the interface, but it hides weak segments when one or two campaigns carry the average. You can finish the test thinking AI Max has worked, then scale spend and discover that most of the incremental budget went to lower-quality queries or weaker regions.
We see this most often in lead-gen accounts running broad match, Smart Bidding and mixed-quality conversion actions. The platform has enough data to spend, but not always enough truth to judge lead value. If sales-qualified leads, booked appointments or closed revenue are missing from the bidding signal, the test will optimise around the easiest conversion event rather than the best customer.
“Do not test automation against a messy account structure and call the result strategy. Fix the signal, segment the risk, then let the experiment tell you something useful.”
— Amy M, Account Executive, PPC Geeks
Our clear takeaway is simple: before you run the new tests, split campaigns by commercial intent and conversion value. Keep brand, competitor, generic and location-dependent campaigns separate where the economics differ. This is exactly the type of issue we look for in a free Google Ads audit, especially where automation, tracking or campaign structure is already affecting performance.
There is also a reporting point. These tools will create pressure to make faster decisions because the planning stage now sits closer to implementation. That is useful for experienced teams and risky for overstretched ones. If your internal report cannot distinguish between revenue, margin, qualified leads and raw enquiries, the new tools will speed up the wrong decisions. Our guide to Google AI tools for UK PPC teams makes the same point across Google’s wider automation stack: cleaner inputs matter more as the platform gets more powerful.
How to roll out AI Max testing without surrendering control
Start with campaign eligibility. Exclude campaigns with broken conversion tracking, undercounted offline sales, duplicated conversion actions or major landing page changes planned during the test window. A test cannot separate AI Max performance from tracking noise if both change at once.
Next, build the experiment around business value. For ecommerce, group campaigns by margin profile and average order value, not just product category. For lead generation, separate enquiry types by sales value and lead acceptance rate. If two campaigns produce conversions with different commercial outcomes, they do not belong in the same readout.
Set one primary success metric before launch. Use profit-weighted ROAS, qualified lead CPA, booked-call CPA or revenue from imported offline conversions. Do not judge the test on conversion volume if the sales team rejects a large share of those conversions.
Lock brand and location guardrails where they matter. If a campaign protects brand terms, regulated claims, dealer territories or service-area economics, keep those controls active during the test. Testing automation does not require you to hand over brand defence or regional budget discipline.
Run a pre-test query and asset audit. Pull the last 30 days of search terms, split them into brand, competitor, generic high intent, research and poor-fit terms, then record the baseline mix. After the test, compare the mix again. If AI Max improves CPA by shifting traffic towards easier branded demand, that is not real expansion.
Use Performance Planner as a scenario tool, not an instruction button. Build separate forecasts for current targets, modest target relaxation and budget-only changes. Then document which assumption you accept before applying anything. Google’s own announcement of the AI Max testing tools confirms the one-click application route, which is exactly why governance matters.
For experiment setup, follow the platform mechanics rather than copying an old campaign draft process. Google’s guidance on Google Ads experiments is the reference point for understanding how drafts and experiments work inside the account. For forecasting, check the Performance Planner support docs for eligibility and planning assumptions before you trust the numbers.
Finally, write a decision rule before the test starts. For example: scale only if qualified lead CPA stays within target and non-brand conversion volume grows. Or for ecommerce: scale only if conversion value rises and margin-adjusted ROAS holds. Without that rule, teams cherry-pick the metric that makes the test look good.
What this means for your campaigns
AI Max testing is a useful step from Google because it gives advertisers a better way to test Search automation across real budget and ROI scenarios. But it also makes weak account structure more expensive. A sloppy test will not just waste a fortnight. It will justify scaling a setup that was never built to measure profit.
UK advertisers should treat these tools as decision support, not permission to accelerate every campaign. Clean the conversion signal, segment campaigns by value, protect brand and location rules, and use forecasts to ask sharper questions about scale. The teams that do this properly will learn where AI Max deserves more budget. The teams that do not will mistake extra spend for progress.
If you want a second pair of eyes on how this affects your account, our team offers a free Google Ads audit, with no strings, and if you would rather hand the whole thing over, our Google Ads management team runs experiments like these every week. For the detail behind this, see the AI Max for Search reporting docs.
Frequently asked questions
What is changing with AI Max testing in Google Ads?
Google is adding multi-campaign A/B testing for budgets and ROI targets, plus support for experiments with brand and location controls enabled. Performance Planner is also being tied more closely to campaign changes.
Should UK advertisers switch AI Max on across all Search campaigns?
No. Start with campaigns that have clean conversion tracking, enough data and clear commercial value. Exclude campaigns with weak signals, mixed lead quality or major landing page changes.
How should lead generation advertisers test AI Max?
Use qualified lead CPA, booked appointment CPA or offline revenue as the main success metric. Do not judge the test on raw form fills if sales quality varies.
How should ecommerce advertisers use the new planning tools?
Group campaigns by margin and average order value before testing. A revenue increase is not enough if the extra spend shifts towards low-margin products.
What is the biggest risk with AI Max testing?
The biggest risk is scaling from a blended result that hides poor segments. Always compare brand, non-brand, location and query mix before increasing budgets.
