English
Deutsch
PolskiAI Max Migration in September 2026: What Changes Now and Why DSA Waits Until February
Management Summary
AI Max brings automated search matching and advertising content into Google Search campaigns. The migration follows different schedules: some existing automation settings were due to move in September 2026, while Dynamic Search Ads, which use website content to create ads, are scheduled for February 2027. For advertisers, the starting configuration determines what changes. Knowing whether wording, matching or destinations can change helps explain campaign results; a higher recorded purchase count alone does not prove additional sales.
Contents
Google’s AI Max migration has two separate timetables. Search campaigns using campaign-level broad match or standalone automatically created assets were scheduled for gradual migration during September 2026. Dynamic Search Ads, usually shortened to DSA, are excluded from that September migration and follow in February 2027. Treating all three as one September deadline can send a campaign review in the wrong direction. [1, 2]
The migration name is not enough
Campaign-level broad match is a campaign setting; it should not be confused with the mere presence of an individual broad-match keyword. Automatically created assets, or ACA, concern automatically produced advertising content. DSA uses website content to help match searches and generate ads. These starting points call for different checks even when the destination is described as AI Max.
A campaign inventory needs the previous configuration, current settings and migration date. The September rollout window has passed, but an announcement alone does not prove the status of a particular campaign. As checked on 26 October 2026, the published plan still separates that rollout from DSA in February 2027.
Which settings are expected to change
Google’s August announcement describes the migrated defaults. For campaign-level broad match, search term matching is on, while text customization and final URL expansion are off. For standalone ACA, search term matching and text customization are on, while final URL expansion remains off. Existing brand inclusions and exclusions are described as preserved for the broad-match transition. [1]
Search term matching expands which searches can trigger ads. Text customization allows Google to adapt advertising text. Final URL expansion allows a different relevant destination to be selected. The three switches therefore affect matching, wording and destination separately.
| Starting configuration | Text customization | Final URL expansion | Search term matching |
|---|---|---|---|
| Campaign-level broad match | Off | Off | On |
| Standalone ACA | On | Off | On |
This table describes the announced migration defaults, not the current state of an inspected customer account. A campaign that was edited manually needs its actual settings checked. A name in a report cannot substitute for that comparison.
A practical review for a multilingual shop
A fictional retailer advertises separate English, German and Polish product pages. Before interpreting the week’s results, the review records whether advertising text can change automatically and whether destination expansion is enabled. It then compares the visible ad promise with stock, prices and the page reached after a click.
An unchanged destination setting does not by itself prove that every ad is suitable. An automatically produced headline may still need a content review. Conversely, a changed conversion total does not establish that a different destination was selected. Configuration evidence and outcome evidence belong in separate columns.
Campaign review record
Previous feature: standalone ACA
Observed migration date: recorded from the account
Current text customization: verified
Current final URL expansion: verified
Brand restrictions: compared with the saved configuration
Landing pages and conversion goals: reviewed
Owner and follow-up date: recordedDSA has a separate preparation window
The more recent August timeline places DSA automigration in February 2027. It also describes the removal of new DSA ad-group creation during that phase. Earlier announcements contain older dates, so a copied checklist should retain both the source date and the feature it concerns. The broad label “legacy Search migration” loses that distinction. [1, 2]
For an existing DSA setup, the additional time supports a controlled comparison: a defined baseline, a limited test and an explicit acceptance decision. A planned migration record should include the pages and exclusions currently used, rather than only the campaign ID. That record helps explain a later change in traffic mix.
Measuring a change without inventing a cause
A before-and-after chart is useful operational evidence, but seasonality, promotions, budget changes and tracking changes can move at the same time. A test should define its main outcome before results arrive. Purchase value, qualified leads and a platform’s conversion count can answer different questions.
A small migration log beside the report is often more useful than another aggregate score. It makes visible when configuration changed and which assumptions were checked. The existing GAQL Explorer can help prepare readable queries; it does not connect to the account or certify the campaign configuration. The linked article on Conversion Lift explains why attributed results and incremental effects should also remain distinct.
Related tool
GoogleAds – GAQL Query Explorer
Google Ads API v25.1: Reading Conversion Lift Data Without Creating a Study
Questions and answers
Are brand inclusions and exclusions also preserved in the transition from standalone ACA?
The statement that existing brand inclusions and exclusions are preserved, as the article reports it, concerns the broad-match transition. For standalone ACA it establishes nothing either way. That is why the review record contains the line “Brand restrictions: compared with the saved configuration”, and that comparison presupposes that the configuration was saved before the migration.
Why does a rise in conversions after the migration not yet show that AI Max works?
Because a platform’s conversion count records attributed results, not incremental ones. If search term matching reaches new searches, the account can credit the campaign with purchases that would have arrived another way without the ad, such as through another campaign for the same brand or through an organic result. The figure then rises while total revenue stays the same.
A well-designed control group helps estimate the additional effect of advertising. Shop purchase value is a useful business outcome, but a higher shop total alone does not prove that AI Max caused the increase: promotions or seasonality can also change it. Conversion Lift compares exposed and unexposed groups and is not available for every account. A migration comparison needs a test design that actually contrasts the relevant settings; an advertising-versus-no-advertising study answers a different question. [4]
How can a campaign’s migration date be established when the report shows only the current state?
The current setting alone does not reveal its start date. For these two automatic migrations, Google provides campaign.aca_migration_date_time and campaign.broad_match_migration_date_time. The fields record the respective migration time in the customer account’s time zone. Each field needs to be read separately; requiring both fields to be populated would omit campaigns migrated from only one starting feature. [5] Additional records help explain the change:
- Saved configurations show what changed between two observations. For example, a difference between Monday and Thursday places the observed change within that interval. This does not establish the exact instant or reveal temporary changes between observations.
- The API resource
change_eventprovides details for the past 30 days, but does not include every account change. Queries require a date filter and a limit of at most 10,000 results. Its absence is therefore not proof that no migration happened. [3]
The GAQL Explorer prepares query text without connecting to the account. Running a query needs a separate authorised API connection. The two migration timestamp fields are separate from the rolling 30-day change-event history. Regular exports preserve additional configuration history before that window expires.
Sources
- Google Ads: August migration plan
- Google: DSA timetable update
- Change Event reference
- Conversion Lift
- Campaign migration timestamp fields
Sources checked: 26 October 2026. February 2027 remains the announced DSA schedule.
Related articles
