Dynamic Search Ads Move in 2027; Two AI Max Upgrades Are Live

Google Ads Help sets February 2027 as the start of automatic upgrades for campaigns using Dynamic Search Ads (DSA). Advertisers can test AI Max against an existing DSA campaign before that rollout: keep the DSA campaign as the control and upgrade the treatment through a custom experiment.

Google’s updated announcement keeps September 2026 as the start of automatic upgrades for campaigns using automatically created assets (ACA) or the campaign-level broad match setting. That rollout is already underway for eligible campaigns, though an individual campaign may not yet have changed. The page’s older headline, summary and rollout descriptions still mention September for DSA; its June update gives DSA the later deadline.

Match the deadline to the legacy setting

Inventory settings by campaign, because a campaign can contain dynamic ad groups alongside other legacy features. A single account-wide migration date will miss changes that may already have reached a Search campaign.

Dynamic Search Ads: Automatic upgrades begin in February 2027. Record which campaigns contain dynamic ad groups, their dynamic targets, page feeds and URL exclusions. The date marks the start of a rollout, not a promise that every campaign changes that day.
Automatically created assets: Automatic upgrades begin in September 2026. Check the campaign’s current AI Max and text customization settings before assuming its original ACA configuration is still the testing baseline.
Campaign-level broad match: Automatic upgrades also begin in September 2026. Record this setting separately from ordinary broad match keywords and from any dynamic ad groups in the campaign.

For either September feature, use campaign settings and account notifications to establish what has happened in your account. A future DSA deadline does not reverse an ACA or broad match upgrade that has already occurred.

Capture the DSA baseline before changing the campaign

Start with a campaign whose DSA traffic matters to the decision. Record its dynamic ad groups, targets, page feeds, URL exclusions and negative keywords. Inspect the pages receiving traffic, including pages for unavailable products or content outside the campaign’s intended scope. DSA uses website content to select landing pages and generate headlines, so a change to the site can change the campaign’s behavior during a test.

Save the campaign’s bidding strategy, budget, location targeting, conversion actions, attribution settings and brand controls. Establish a reporting baseline for spend, conversions, conversion value and the efficiency measure used to manage the campaign. Keep its search terms, generated headlines and landing pages alongside those totals: an apparent gain in volume is less useful if traffic reaches unsuitable destinations or queries.

Note whether the campaign is entirely dynamic or mixes dynamic and standard ad groups. Also flag any ACA or campaign-level broad match setting that has already moved to AI Max. Those details determine which experiment can answer the DSA question and what the untreated campaign actually represents.

Use the experiment path that tests DSA

For a campaign made entirely of dynamic ad groups, use a custom experiment. At the draft campaign’s “Make Changes” stage, expand Dynamic Search Ads settings and select “Upgrade campaign.” The treatment then moves to AI Max when the experiment launches, while the original DSA campaign supplies the control. Leave automatic application off if the decision requires a review of performance and destinations.

A templated AI Max experiment is unavailable for a campaign made entirely of dynamic ad groups. In a mixed campaign, its treatment upgrades only the standard ad groups; the dynamic ad groups stay dynamic. That template can answer a question about the standard portion, but its result alone cannot establish whether replacing the DSA portion worked. Choose the custom upgrade path for that decision.

Check the AI Max experiment rules before building the trial. The built-in flow excludes campaigns with text customization already enabled, Display Network targeting, portfolio bidding, shared budgets, bidding exploration or an active experiment. The documented custom-flow limits also apply except for the Display Network restriction. Resolve eligibility before interpreting a missing experiment option as a DSA deadline issue.

Preserve the controls you intend to measure

The built-in AI Max experiment enables search term matching, text customization and final URL expansion in its treatment by default. Campaign-level text customization and final URL expansion can be deselected during setup. Ad-group changes, such as disabling search term matching or adding URL inclusions, require an edit after creation. Record the treatment’s final configuration; otherwise the result may reflect a bundle of changes you did not intend to test.

Review URL exclusions and permitted destinations, and keep negative keywords aligned with the control. Brand inclusions or exclusions added during built-in experiment setup apply to both arms while it runs. If brand controls must change, record that shared change rather than crediting its effects solely to AI Max.

Avoid changing conversion tracking, bidding or landing pages during the measurement period where possible. If a substantial change is unavoidable, log when it happened and decide whether the comparison remains usable. The same discipline applies to a mixed campaign whose standard ad groups or September legacy settings may change during the trial.

Set the result and rollback rules in advance

Choose a primary measure that matches the campaign’s objective, such as conversion value at an acceptable return on ad spend or conversions at an acceptable cost per acquisition. Compare treatment and control after allowing for the campaign’s normal conversion lag. Review spend, search terms, generated text and landing pages alongside the primary measure so that an efficiency result does not hide a loss of traffic quality.

Define reasons to stop the treatment or end without applying it: ads reaching prohibited pages, failed brand boundaries, unreliable tracking or search terms materially outside the intended market. If performance misses the agreed efficiency threshold after a useful measurement period, retain the DSA control and investigate the difference. The approaching migration window is not evidence that a weaker treatment should be applied.

Application has a specific consequence in a custom DSA experiment. “Update your original campaign” converts the original to AI Max; “Convert experiment to a new campaign” converts the treatment and pauses the original. Review that choice and the final settings before applying a successful test. A built-in AI Max experiment ended without application returns its AI Max settings to their pre-experiment state, allowing a revised trial without treating the first result as a permanent campaign change.