AI Max for Search Is No Longer Optional: What the September Migration Means for Every Advertiser
Google is auto-migrating DSA, broad match, and auto-assets to AI Max from September 2026. Here is what you actually control, and how to migrate on your own terms first.
Google has a habit of making decisions for advertisers and calling it automation. AI Max for Search is the latest example, but this one carries a harder deadline than most: from September 2026, Dynamic Search Ads campaigns, campaign-level broad match, and automatically created assets will be migrated to AI Max whether you are ready or not. There is no opt-out path. The question now is not whether you will use AI Max, but whether the migration happens on your terms or Google's.
This piece is not a product overview. It is a field-level briefing for practitioners managing real accounts: what the feature actually does, what the performance numbers mean in context, which controls survive the migration and which do not, and the practical steps to audit and migrate DSA campaigns before the forced switch removes the option to do it cleanly.
The September timeline is close enough that accounts with significant DSA spend need to start this work now, not in August.
What AI Max for Search Actually Is (and Is Not DSA)
The confusion is understandable. AI Max for Search does what DSA did: it finds queries beyond your keyword list, generates ad copy dynamically, and routes users to the most relevant page. But the architecture is different enough that treating AI Max as a renamed DSA will lead you to manage it wrong.
DSA was a campaign type. You created a separate DSA campaign, pointed it at your website or a page feed, and it ran independently from your keyword campaigns. AI Max is a feature toggle applied within existing Search campaigns. You activate it at campaign level, and it layers on top of your keyword structure rather than sitting alongside it in a separate campaign.
This distinction matters operationally. When you migrate a DSA campaign to AI Max, you are not just switching campaign type: you are bringing that campaign's budget, bidding, and audience signals into a structure that also includes whatever keywords you have or choose to add. The interplay between keyword matching and AI Max's search term matching is not the same as the DSA-vs-keyword-campaign relationship you have been managing.
AI Max operates through three sub-features: search term matching (the DSA-equivalent query expansion), text customisation (dynamic headline and description generation from your landing pages and assets), and final URL expansion (routing to the best-matching page rather than the ad's stated URL). You can enable these independently. Enabling all three is what Google's data calls the "full feature set," and that combination produces the highest reported uplift, though it also represents the most autonomy handed to the system.
The feature became generally available in April 2026 after an extended beta. Per Search Engine Land reporting at the time, Google announced it had been adopted by "hundreds of thousands" of advertisers globally before the GA announcement, which is the stated basis for the decision to retire DSA.
The Migration Timeline: Voluntary Now, Automatic in September
The current state, as of June 2026, is that DSA campaigns can still be created and run normally. The migration tools are live: you can upgrade individual DSA campaigns to AI Max through the Google Ads interface, and the workflow preserves campaign history. That window closes in September.
From September 2026, per the Google Ads blog announcement:
- New DSA campaign creation will be disabled in Google Ads, Ads Editor, and the Ads API.
- Existing DSA campaigns will be auto-migrated to AI Max by Google.
- Campaign-level broad match and automatically created assets (which have their own existing toggle) will be brought under the AI Max umbrella in the same migration.
That last point is one that many practitioners have missed. The September migration is not just about DSA. If you have campaigns running automatically created assets or using campaign-level broad match as a feature, those are also in scope. Check all three before you assume only your DSA campaigns need attention.
The Smarter Ecommerce team noted in March 2026 that adoption had accelerated sharply in February, and flagged that the curve "could go vertical" once the DSA retirement was confirmed. That inflection has now happened. If you are still waiting, you are in the trailing segment.
Performance Data: What Google's Own Numbers Show
I will give you the numbers, but I will also give you the appropriate level of scepticism to apply to them.
Google's internal data from May 2025 showed a 14% average uplift in conversions or conversion value at similar CPA or ROAS for advertisers activating AI Max. For campaigns running predominantly on exact and phrase match keywords, the figure rose to 27%. By May 2026, Think with Google data was citing 27% more conversions at similar efficiency as the headline figure for AI Max adopters.
The 7% figure is separate and worth understanding: that is the incremental uplift from enabling the full AI Max feature set (all three sub-features) compared with enabling search term matching alone. In other words, adding text customisation and final URL expansion on top of search term matching delivers a further 7% on average, per Search Engine Land's reporting of the GA announcement data.
Case study numbers from the Google Ads blog: L'Oréal reported a 2x higher conversion rate at 31% lower cost-per-conversion, specifically attributing it to queries like "what is the best cream for facial dark spots?" that AI Max captured and DSA had not. MyConnect, an Australian utility connection service, saw 16% more leads at 13% lower cost-per-action, with 30% of those conversions coming from net-new queries.
Three caveats worth applying before you forecast these numbers into your own account:
First, all of these figures are Google-sourced or Google-facilitated. There is no independent audit. The direction is credible, but the magnitude should be treated as upper-bound until your own data shows otherwise.
Second, the "similar CPA or ROAS" framing conceals the question of whether efficiency held or improved. "Similar" means within a tolerance Google defines, not necessarily flat.
Third, case studies select for success. L'Oréal's query expansion working beautifully for skincare does not predict what happens when a B2B software account activates the same feature.
Keywords remain an essential component. AI Max helps advertisers expand beyond keywords while retaining control.
Google Ads blog, April 2026
That quote, from a Google spokesperson cited by Search Engine Land, is worth holding onto: it is the clearest statement that AI Max is additive to keyword strategy, not a replacement for it. The accounts where AI Max performs well tend to be those where the keyword layer is already well-structured, not those hoping AI Max will do the keyword work for them.
What Controls You Gain (and What You Lose)
This is the section most migration guides skip over in favour of reporting the headline uplift. Let me be specific.
What you gain with AI Max:
Brand controls at both campaign and ad group level. You can specify brands to include (ensuring AI Max's query expansion stays within defined brand territory) and brands to exclude (protecting competitor spend and keeping brand campaigns clean). This is more granular than anything DSA offered.
Location-of-interest targeting at ad group level. DSA did not have this. AI Max allows you to target users whose search intent signals a geographical location, regardless of their physical location. For service businesses with geographical scope this is a meaningful addition.
URL inclusion and exclusion controls at the campaign level, carrying over from the DSA page feed model. You can still direct AI Max at specific sections of the site rather than the whole domain.
What changes or narrows:
Ad copy control. Text customisation pulls from your existing assets and landing page content. You can provide custom parameters and asset signals, but the system assembles the final headline and description. If you have been writing DSA headlines manually via expanded dynamic ad groups, that precise control goes away.
URL routing transparency. Final URL expansion selects destination pages algorithmically. You can review and exclude specific URLs, but you are not approving individual query-to-URL pairings in advance.
Search terms visibility. This has been a persistent concern with AI Max in beta: the search terms report does not show all matched queries, only those meeting Google's privacy thresholds. This is the same limitation that exists across Smart Bidding generally, but practitioners coming from DSA (which showed more term-level detail) will find the opacity a step back.
| Control | DSA | AI Max |
|---|---|---|
| Campaign type | Standalone campaign | Feature on existing Search campaign |
| Brand inclusion/exclusion | Not available | Campaign and ad group level |
| Location-of-interest targeting | Not available | Ad group level |
| URL controls | Page feed or domain | URL inclusion/exclusion lists |
| Ad copy generation | Auto from page content | Auto from assets + page content |
| Final URL routing | DSA auto-target determines URL | Final URL expansion (can be toggled) |
| Search term visibility | Partial | Partial (same privacy thresholds) |
| Keyword coexistence | Separate from keyword campaigns | Layered within same campaign |
The DSA Upgrade Workflow: Preserving History and Settings
The upgrade path Google provides does preserve campaign history, which matters if you are using DSA conversion data to inform Smart Bidding in the migrated campaign. The history carries over; the bidding model does not restart from scratch.
The practical steps in sequence:
Before you start: Export your DSA campaign settings, ad group targets, and URL exclusions as a baseline. You want a record of what you had before the migration in case something unexpected changes in the first week.
Review your page feed or domain targets. AI Max will inherit these, but this is a good moment to audit whether your site's crawlable content is what you actually want the system indexing. If product pages have thin copy or URLs you do not want in the mix, add exclusions before the migration.
Clean your negative keyword lists. Negatives apply to AI Max's query expansion the same way they apply to keyword matching. Any gaps in your negative list will be visible in AI Max's search term report within days of activation.
During migration: Use the upgrade tool in Google Ads rather than creating a new AI Max campaign from scratch. This preserves the bidding history. If you create a new campaign, you start the Smart Bidding learning period again with no historical signal.
Enable AI Max features selectively at first. Search term matching is the core feature and the closest analogue to DSA's query expansion. Enable that first and run for two weeks before switching on text customisation and final URL expansion. This makes it easier to isolate where performance changes come from.
After migration: Set up a comparison view in reporting before you flip the switch, so you have a clean before/after baseline from the same campaign. Monitor search terms daily for the first two weeks. The expansion range will likely be wider than your DSA campaigns, particularly if you had conservative page feed targets previously.
Early-Adopter Results and Where the Scepticism Comes From
The accounts that have performed well with AI Max tend to share a few characteristics: strong conversion volume feeding the bidding model, well-maintained negative keyword lists, and domains where the page content is genuinely descriptive of what users are searching for. That last point sounds obvious, but it eliminates a lot of accounts where page copy is written for the brand rather than for search intent.
The scepticism in the practitioner community is concentrated in a few areas. Query visibility is the most consistent complaint: without full search term transparency, it is difficult to diagnose when AI Max is matching on intent you did not intend to serve. The brand controls help with the obvious cases, but there is a middle layer of query expansion that remains opaque.
The second concern is about the "full feature set" interaction between text customisation and final URL expansion. When both are on, the system is simultaneously generating the ad and choosing the destination. If the landing page that AI Max routes to is not as strong as the one you would have specified manually, the conversion rate impact can offset the query volume gains. This is why monitoring at URL level, not just campaign level, is important in the first few weeks.
A third sceptical view holds that the reported uplift figures reflect the novelty effect of expanded reach rather than improved efficiency. Under this reading, the 14-27% conversion uplifts are a volume story, not an efficiency story, and accounts that were already doing sophisticated query expansion through phrase and exact match would see smaller gains. There is enough heterogeneity in the data to make this plausible, even if it is not the modal outcome.
How to Audit Your DSA Campaigns Before the Forced Switch
Run this audit on every DSA campaign you control, ideally in June or July. Waiting until August reduces your ability to course-correct before Google does the migration for you.
Step 1: Identify scope. Pull a list of all DSA campaigns, including any campaigns with campaign-level broad match or automatically created assets enabled. These are all in scope for September.
Step 2: Check conversion volume. For each DSA campaign, what is the monthly conversion volume over the last 90 days? AI Max requires meaningful conversion signal to constrain its query expansion. Campaigns below 30 conversions per month will need attention: either consolidate them before migration, or accept that the learning period will be longer and the reach broader.
Step 3: Review page feed quality. If your DSA campaigns use page feeds rather than domain targets, review the feed for accuracy. Expired products, out-of-stock pages, thin content pages, and duplicated URLs should be cleaned out before you hand the feed to AI Max.
Step 4: Audit URL exclusions. List every URL exclusion currently applied and verify it is still valid. Then identify any URLs you would want to add that you never got around to excluding from DSA. AI Max will index the same URLs DSA did, so any gaps you tolerated in DSA will surface again.
Step 5: Rebuild your negative keyword lists. Run the search terms report for your DSA campaigns across the last 90 days. Pull every query that spent without converting and has a CPA above your target. These become negatives before migration. This is the most time-intensive step but the most consequential.
Step 6: Map your URL performance. Before migration, identify which landing pages DSA was sending traffic to, and what the conversion rates were by URL. This gives you a baseline to measure final URL expansion against once AI Max is live.
Building Your Migration Checklist
The migration itself is not technically complex. The preparation is where most accounts will succeed or struggle.
Do not activate AI Max on any campaign without: a clean negative keyword list built from a 90-day STR review, URL exclusions reviewed and updated, and a conversion volume check confirming the campaign has adequate signal. These three steps are the difference between a managed migration and one that requires emergency intervention two weeks later.
The checklist, in order:
- Identify all campaigns in scope (DSA, campaign-level broad match, auto-assets).
- Check conversion volume for each; flag any below 30 per month for consolidation.
- Clean page feeds or domain targets; remove expired and thin-content URLs.
- Run a 90-day STR audit; build pre-migration negatives from it.
- Review and update URL exclusions.
- Export baseline reporting views (conversion by URL, query category, cost by ad group).
- Enable AI Max with search term matching only; run for two weeks.
- Review search terms daily for the first fortnight; add negatives as needed.
- Enable text customisation; monitor headline performance by ad group.
- Enable final URL expansion last; watch URL-level conversion rate against baseline.
- Assess brand controls; add brand inclusions/exclusions relevant to your account.
The accounts that will have problems in October are the ones that let Google run the auto-migration in September without doing steps one through six first. The auto-migration will run on whatever state your campaigns are in. If your page feeds are stale and your negatives are thin, that is what AI Max inherits.
This is a migration worth doing carefully, before September, on your terms.
Frequently Asked Questions
- Will AI Max campaigns share a budget with my existing keyword campaigns?
- Yes. AI Max is a feature applied within existing Search campaigns, not a separate campaign type. If you migrate a DSA campaign to AI Max, it runs with that campaign's budget. If you enable AI Max on an existing keyword campaign, the same budget covers both keyword-triggered and AI Max-triggered traffic. This is a meaningful difference from DSA, which had its own dedicated campaign and budget.
- Can I turn off final URL expansion and text customisation while still using AI Max's query expansion?
- Yes. The three AI Max sub-features (search term matching, text customisation, final URL expansion) can be enabled independently. Google recommends the full feature set for maximum performance, and the 7% incremental uplift figure applies to enabling all three together. But you can start with search term matching only and add the others once you have reviewed the initial behaviour.
- What happens to DSA campaigns that Google auto-migrates in September if I have not prepared them?
- Google will apply AI Max to the campaign in its current state. Your existing page feed or domain target, URL exclusions, and bidding settings will carry over. The risk is that any weaknesses in your existing setup (stale page feeds, thin negative lists, low-quality landing page URLs) will be inherited by AI Max and potentially amplified by its broader query expansion. The time to clean these up is before September, not after.
- Does AI Max replace the need for a keyword list?
- No. Google has been explicit that keywords remain central to Search campaigns. AI Max's search term matching expands beyond your keyword list to find additional queries, but it works alongside your keywords, not instead of them. The best-performing AI Max campaigns in Google's own data are those with well-structured keyword layers, not those relying on AI Max to do all the targeting work.
Sources & Further Reading
- Google Ads blog: AI Max for Search campaignshttps://blog.google/products/ads-commerce/google-ai-max-for-search-campaigns/
- Google Ads blog: DSA upgrade to AI Max 2026https://blog.google/products/ads-commerce/dsa-upgrade-to-ai-max-2026/
- Search Engine Land: Google to retire Dynamic Search Adshttps://searchengineland.com/google-retire-dynamic-search-ads-ai-max-474262
- Smarter Ecommerce: The ultimate guide to AI Max for Google Searchhttps://smarter-ecommerce.com/blog/en/google-ads/the-ultimate-guide-to-ai-max-for-google-search/
- Google Ads Help: AI Max for Search campaignshttps://support.google.com/google-ads/answer/15910187?hl=en