Google's AI Max Auto-Upgrade: What to Do Before September

Every Dynamic Search Ads campaign you're running gets rewritten whether you touch it or not. Starting in September 2026, Google auto-upgrades Dynamic Search Ads to AI Max, and if you don't act first, the algorithm picks your new settings fo

Every Dynamic Search Ads campaign you're running gets rewritten whether you touch it or not. Starting in September 2026, Google auto-upgrades Dynamic Search Ads to AI Max, and if you don't act first, the algorithm picks your new settings for you. Google is framing it as a feature bundle. For anyone who built a DSA campaign around a specific page feed, a hard-won negative keyword list, or a URL exclusion that took a support ticket to sort out, it reads more like a renovation crew showing up with a two-week notice and their own blueprint.

What Actually Changes When DSA Becomes AI Max

Dynamic Search Ads used to work by crawling your site and generating headlines and landing pages straight from your own content, sidestepping keyword lists entirely. AI Max keeps that trick and stacks three more on top: search term matching that reaches beyond your site's own text, automated text customization pulled from your assets and landing pages, and final URL expansion that can send traffic to a different page than the one you set if Google's model thinks it converts better. All three now live inside a standard Search campaign instead of a separate DSA campaign type, which is the structural part most advertisers miss. Google's own setup documentation spells out the new controls, and it's worth ten minutes even if you're not touching anything yet.

The September Date Is a Default, Not a Deadline

Google's own timeline runs in two phases. Voluntary upgrade tools are live right now, and Google is actively encouraging advertisers to use them, because doing so ports your existing settings into the new structure instead of starting from Google's assumptions. The automatic phase begins in September, at which point new DSA campaigns can no longer be created at all and eligible existing ones start converting on Google's schedule. Search Engine Journal's coverage is blunt about the framing: this isn't an optional new feature sitting next to the old one, it's a retirement with a grace period attached.

Three Things Break First If You Let It Auto-Migrate

The failure mode here isn't dramatic. Nobody's account gets shut off. What actually happens is quieter and more expensive: a site section that was excluded from paid traffic for a good reason starts showing up in reports again, a brand-safety exclusion softens back to a broader default, or the campaign's effective budget pacing shifts because the new matching logic is casting a wider net than the old one did. None of that trips an alarm. It just shows up three weeks later as a CPA that's crept 15 to 20% off target and nobody can point to the exact day it happened.

This is the same failure pattern we've written about in dynamic ad budget liquidation: the damage isn't the change itself, it's the gap between the change happening and someone noticing. A campaign that drifts for three weeks before a human checks the dashboard has already spent three weeks of budget on the drift.

Do You Migrate This Yourself, or Hand It Off

For a business owner who checks their ad account once a week between everything else running the business, doing this migration properly is a real Tuesday: export your current DSA performance as a baseline, list every URL exclusion and negative keyword by hand, upgrade voluntarily instead of waiting, then run the new setup in parallel with the old one and watch search term reports daily instead of weekly for a couple of weeks. That's not a complicated task. It's a task that gets skipped the first time something more urgent comes up, and September has a way of being busy for everyone.

The upgrade isn't the risk. The three weeks after it, when nobody's watching the account closely, is where the budget actually leaks.

— Bunny Honey Club

This is also exactly the shape of problem workflow automation is built for: not a one-time migration, but a standing rule that catches the drift the moment it starts instead of a month later in a CPA report. If you'd rather hand the whole thing off, that's the same work either way. We wire a CPA ceiling to the account, the same pattern we run for clients who treat ad budgets like open trading positions, and it fires the pause before the drift becomes a bad month, migration or not.

The Checklist, If You're Doing It Yourself

Five things, in order, before September:

1. Pull your current DSA performance as a written baseline: impressions, clicks, conversions, CPA. You'll need something to compare against once the new setup is live.
2. List every URL exclusion and negative keyword. These are the settings most likely to reset to a default if you let the automatic upgrade handle it instead of migrating manually.
3. Upgrade voluntarily, using Google's own migration tools, rather than waiting for September. Manual migration ports your existing settings; automatic migration applies Google's defaults.
4. Run the new setup in parallel with the old campaign for two to three weeks before fully retiring it, checking search term reports daily rather than at the end of the week.
5. Set a hard CPA ceiling that pauses the campaign automatically if it drifts past a threshold you choose. Don't rely on remembering to check.

If Google Ads is one piece of a bigger stack, this same logic applies wherever automation quietly rewrites your defaults. We covered the rest of that stack, Meta, TikTok, Microsoft, in our 2026 paid ads breakdown, and the honest math on where automation actually pays for itself versus where it's cost theater in what a small business should automate first. An ad account that isn't quietly leaking budget isn't a cost saved. It's revenue you'd have otherwise lost to a setting nobody was watching, which is the same argument we make in why automation is a revenue driver, not just a cost cutter.