Google AI Max GA: DSA Sunset Playbook for Agencies

What April 15 Actually Shipped

April 15 was a GA state change, not a net-new product launch. AI Max for Search had been in open beta since May 2025 — roughly 11 months — with "hundreds of thousands of global advertisers" already using it per Google's own announcement. What changed:

• GA status. AI Max is now the default recommended campaign type for Search, not a beta feature.
• Auto-upgrade announcement. September 2026 is the hard deadline for DSA, ACA, and campaign-level broad match to be absorbed into AI Max.
• Voluntary upgrade tools. Rolling out the week of April 15 — in-UI banners for ACA and broad-match users, one-click A/B experiment tools, and DSA-specific upgrade flows.
• New DSA creation blocked in September. Across the Google Ads UI, Google Ads Editor, and the Google Ads API.

Google's communication explicitly framed this as "announcement to completion in 5.5 months," shorter than any prior major campaign-type sunset. That compression is the single most important operational detail in this announcement.

Three Campaign Types Being Sunset

The September auto-upgrade absorbs three separate legacy campaign features into AI Max:

Notable non-sunsets: standard keyword campaigns continue as-is. Exact match and phrase match stay. Google's own spokesperson told Search Engine Land: "Keywords remain an essential component of a successful campaign strategy, providing the 'fuel' for our AI." If a client's account is pure exact and phrase match with no DSA, ACA, or campaign-level broad match, nothing auto-upgrades in September.

The scope matters for triage. DSA was historically the catch-all for service businesses, lead-gen, and publisher sites without product feeds. Those advertisers — unlike retailers — have fewer fallback campaign types. The September migration is most operationally painful for lead-gen and B2B agencies, least for retail agencies that have already moved to Performance Max.

Auditing the 7% Lift Claim

Google's April 15 post states: "AI Max for Search campaigns see an average of 7% more conversions or conversion value at a similar CPA/ROAS when using the full feature suite." Four things about that claim require unpacking before agencies quote it to clients:

Two previous AI Max lift numbers have circulated in Google's communications. Agencies should know exactly how they compare to avoid quoting them incorrectly:

What Independent Tests Show

The independent testing record on AI Max is more varied than Google's 7% headline suggests. Four data points worth knowing:

The pattern is consistent across agencies: AI Max tends to expand reach and impressions, but whether that translates to conversions or revenue varies wildly by vertical, account maturity, and brand-control setup. SMEC's query- expansion audit (1M+ impressions) found AI Max still delivers 80% of impressions on exact-match queries, 20% phrase, and less than 1% broad — but that small broad slice had 49–63% overlap with legacy broad match, and one account saw 69% of impressions on competitor-brand terms.

Agency takeaway: plan for variance, not uniform lift. The campaigns most likely to see the 7% headline are mature accounts with clean conversion tracking, defined brand-safety guardrails, and non-retail verticals with long-tail query space. Everyone else should expect more variance in both directions.

Smart Shopping → PMax Precedent

The only comparable precedent for this kind of forced migration is the Smart Shopping → Performance Max transition in 2022. The timelines diverge meaningfully:

Two lessons from 2022 that apply to 2026:

1. Learning periods bite the agencies that wait. In 2022, campaigns migrated in July and August hit learning- period issues right through September — peak back-to-school spend. Campaigns voluntarily migrated in April and May had the learning period behind them. The 2026 compressed window makes this risk more acute.
2. Reporting continuity is softer than the marketing claims. Many agencies lost access to pre-migration Smart Shopping benchmarks in specific dimensions after the auto-upgrade. Best practice: export everything useful before migrating, regardless of platform promises.

Default Feature Activation After Upgrade

Not every upgraded campaign gets the same default settings. Google ports different configurations depending on which legacy feature the campaign started from:

The practical effect: a DSA campaign that auto-upgrades inherits the most aggressive AI Max configuration — all three features active by default. If the client is not ready for Google's AI to generate creative, match broader queries, and route to expanded landing pages, you want to voluntarily migrate before September so you can opt individual features off per campaign. After auto-upgrade, each feature is still toggleable, but you're already reporting on traffic Google has chosen.

The Cost-of-Delay Model

Three costs accumulate the longer agencies wait to voluntarily migrate:

• Control cost. Every week past July without voluntary migration means less time to dial in text guidelines, brand safety rules, and final URL expansion controls before Google applies its defaults.
• Learning-period cost. Campaigns migrated in September hit their 1–2 week learning period in the final two weeks of Q3 — the highest-spend B2B window of the year. Any learning-period underperformance directly reduces quarter-end results.
• Reporting-continuity cost.Historical DSA benchmarks not exported before auto-upgrade may become partially unavailable. Without clean pre/post comparison data, it's harder to prove AI Max improved or hurt performance — which matters for client retention discussions.

Agency Migration Calendar May–September

The month-by-month calendar agencies should be running against their DSA book of business:

What Google Has Not Clarified

Seven unresolved items agencies should flag to clients so nobody assumes certainty where there isn't any:

• The exact September 2026 auto-upgrade start date — Google says "September," not a specific day
• The 7% lift's sample size — undisclosed
• The Retail benchmark — Retail is excluded from the 7% figure with no replacement number published
• Mixed-campaign behavior — accounts with both DSA and standard keyword ad groups in the same campaign; whether AI Max applies to all ad groups or only the dynamic ones
• Historical DSA reporting retention period after migration — not specified
• A formal opt-out window — does not appear to exist
• Bulk migration via Google Ads Editor — minimally documented; high-volume accounts may face operational friction

The First 30-Day Client Playbook

Week-by-week for the first 30 days of a client engagement specifically focused on DSA → AI Max migration:

Conclusion

DSA → AI Max is not a radical product change — the underlying capability has been available in beta for 11 months. The April 15 news is the timeline. Agencies that treat September as a deadline rather than a casual direction-of-travel will save clients meaningful quarter-end performance. Agencies that assume "Google will figure it out" will watch October reports explain the drop.

Voluntary early migration is the only real control lever. The migration window is 40% shorter than the Smart Shopping → PMax precedent. Google's 7% lift claim is conservative and narrowly framed — do not promise more than that to clients, and expect variance in both directions based on independent agency tests. Export your baseline data before you migrate anything. That's the short version.

Frequently Asked Questions