Google Expands AI Max With New Testing and Planning Tools. Here’s What It Actually Means for Your CPCs.

Google just made it easier to test AI Max before you commit budget to it — which sounds like exactly the kind of control advertisers have been asking for. But easier testing isn’t the same as a safer auction, and it’s worth being clear-eyed about the difference before you flip anything on at scale.

Google introduced new testing and planning tools for AI Max for Search campaigns, according to Search Engine Land and Google’s own Google Ads & Commerce Blog. After 20+ years managing Google Ads accounts, here’s how I’m reading this rollout — and what I’m telling clients before they use it to justify scaling up.

1. What’s Actually Changing

Three new capabilities are rolling out inside AI Max for Search campaigns:

1. Multi-campaign A/B testing (launching September 2026). Advertisers can now test different budgets and ROI targets across multiple Search campaigns in a single A/B test, instead of testing one campaign at a time — giving a clearer read on what scaling does to overall account performance.
2. Brand and location controls inside experiments. AI Max experiments can now run with brand and location guardrails intact, so you can measure AI Max’s impact without having to strip out the controls your business actually needs just to run the test.
3. An expanded Performance Planner. The Planner now forecasts how bidding or budget changes could affect your existing campaign performance, and lets you apply Google’s suggested changes directly to your account in one click.

2. Why This Still Tends to Raise Your CPCs

Google is framing this as removing friction: advertisers who were hesitant to test AI Max because it meant sacrificing brand controls or testing campaigns one at a time now have less reason to hold back. That’s a real improvement for measurement. But measurement isn’t the same as auction pressure, and the auction is where your CPCs actually get set. Once a multi-campaign test shows a positive account-level result, the natural next step is to scale — and scaling into AI Max still means competing for a broader, less controllable pool of search inventory than a tightly managed exact-match campaign. In my experience, any tool that makes it faster to say “yes” to more automation tends to show up a few weeks later as a higher average CPC, because you’re now bidding for volume you weren’t bidding for before.

The one-click Performance Planner is the part I’d watch most closely. Shortening the gap between “here’s a forecast” and “here’s that change, live in your account” is convenient, but it also removes a checkpoint — the pause where an advertiser used to sanity-check a suggested bid or budget change against their own account history before applying it.

3. The Real Issue: Control and Data Visibility

Brand and location controls inside experiments are a genuine win — they let more advertisers test AI Max without abandoning restrictions their business requires. But notice what they don’t restore: granular reporting. You still can’t see which asset, which network, or which specific query drove a given conversion inside AI Max. The new tools tell you the net effect on conversions and ROI at the account or campaign-group level; they don’t open up the black box underneath. And because multi-campaign testing reports success in aggregate, a high-CPC campaign that’s quietly underperforming can be masked by cheaper campaigns in the same test group — you won’t see it unless you break the data out yourself.

How This Connects to Your Bid Strategy

This is the same tradeoff I’ve written about before with Performance Max for lead gen. PMax and AI Max work well for large-budget advertisers who can absorb some inefficiency in exchange for scale — but Search-network traffic quality inside both has always been inconsistent, and better testing tools don’t change that underlying inconsistency. Your bid strategy is what determines how much it costs you. A Maximize Conversions strategy tightens the reins on conversion quality and will limit volume if signals are weak — it protects you from waste, but it can also make a multi-campaign test look artificially flat, since it’s deliberately holding back on volume. Other strategies chase volume and will deliver it in a test, but that volume is volatile and far more exposed to the CPC pressure that scaling introduces. A “positive” test result partly reflects which bid strategy you tested with — not just whether AI Max itself is safe to scale.

The Playbook Before You Test Multi-Campaign Scaling

1. Don’t apply Performance Planner’s one-click suggestions blind. Review the forecast against your actual bid strategy and recent CPC trend before you let it push changes live.
2. Set your bid strategy deliberately before you test scale. Decide whether you’re protecting conversion quality (Maximize Conversions) or chasing volume — testing on top of an undecided bid strategy just tests the strategy, not AI Max.
3. Use the new brand and location controls, but treat them as a floor, not a ceiling. Confirm exactly what they exclude, and audit your search term and placement reports after rollout for what still slips through.
4. Cap the budget for your multi-campaign A/B test. Give Google a defined ceiling to test within, so a positive account-level read doesn’t turn into an open-ended commitment once the test ends.
5. Track CPC and CPA by individual campaign, not just account-level ROI. Multi-campaign testing reports success in aggregate — break it out yourself, or a single high-CPC campaign can hide behind the group’s average.

Thinking about testing AI Max at scale? Let’s build the guardrails before you commit budget.

About the Author

Melissa is the Founder and Lead Strategist at Visionary Search, based in Dallas, Texas. With 20+ years of experience managing high-stakes Google Ads accounts, she specializes in protecting and scaling ad spend for high-CPC industries like personal injury law firms, CPG brands, and hospitality businesses. As Google pushes further into automation, she helps clients keep control of their targeting, their data, and their budgets — instead of handing it all to the algorithm.

Connect with her today to review your account before the next automatic update rolls out.

Book Your Free Consultation