DSA to AI Max Migration: What to Do Before Feb 2027

Luke Costley-White

Adclear and DOJO AI partnership graphic: 'Close the loop on agentic marketing. Compliance at the speed of creation.'
二刀流
Two-Sword Style

Google is retiring Dynamic Search Ads. No new DSA campaigns can be created after the reopened creation window closes in January 2027, and all eligible DSA, ACA, and broad-match campaigns migrate automatically to AI Max by February 2027 (ACA and campaign-level broad match migrate sooner, by September 2026). Google claims a 7% conversion lift from the switch; independent tests show results ranging from strong gains to a 35% ROAS drop. Test before you're forced to.

If you're tracking the dynamic search ads to AI Max migration against a single date on your calendar, you're tracking the wrong one. Google has set two separate deadlines for two separate sets of campaigns, and most advertisers only know about the one that doesn't apply to them.

This isn't a one-time compliance task you clear off your list before a deadline and forget. It's a standing decision that arrives twice, with performance data that disagrees with itself depending on who ran the test. This guide covers what's actually changing, what the real numbers say, and what to do about it before either clock runs out.

What's Actually Changing, and When (The Two-Clock Problem)

Google announced the retirement of Dynamic Search Ads on April 15, 2026, with an update published June 11, 2026 (blog.google, "We're upgrading Dynamic Search Ads to AI Max"). The announcement bundles three separate campaign types into one migration, and they don't move on the same schedule.


Deadline

Applies to

Status

September 1, 2026

Automatically Created Assets (ACA) and campaign-level broad match

Unchanged since the original announcement

February 2027

Dynamic Search Ads (DSA) campaigns

Delayed once; creation window reopened June 15, 2026 through January 2027

The Sept 2026 Clock: ACA and Campaign-Level Broad Match

If you run standard Search campaigns with Automatically Created Assets turned on, or with broad match set at the campaign level rather than the keyword level, that migration date has not moved. It lands September 1, 2026 (Search Engine Land, Apr 15, 2026; PPC Land, Apr 15 and May 20, 2026).

The Feb 2027 Clock: Dynamic Search Ads

DSA campaigns run on a separate, later timeline, and it's already shifted once. Google originally set a September 2026 deadline for DSA too, then delayed it to February 2027 and reopened DSA campaign creation from June 15, 2026 through January 2027 (blog.google; PPC Land, Jun 13 and Jun 26, 2026). Google frames the reopened window as a deliberate side-by-side testing period before the forced migration.

Most advertisers only track one deadline and get blindsided by the other. An account running DSA in one campaign and campaign-level broad match in another is exposed to both clocks, on different dates, with different urgency.

What AI Max Actually Replaces, and What It Doesn't

AI Max bundles three features into existing Search campaigns: keywordless search term matching, AI-generated text customization, and final URL expansion. DSA only ever did one of those three jobs.

The Feature That Replaces DSA

Final URL expansion is the direct successor to DSA. It lets Google route traffic to whichever page on your site it judges most relevant, the same core mechanic DSA used to crawl your site and match queries to landing pages. Search term matching and text customization are new capability layered on top, not a replacement for anything DSA did.

Why Side-by-Side Testing Doesn't Work the Way You'd Expect

Google's own reopened creation window implies you can run DSA and AI Max side by side to compare them. In practice, that comparison breaks down inside a single account.

SavvyRevenue tested this directly and found AI Max generated just 72 clicks out of more than 3,000 total, while a DSA campaign was still active in the same account. Once DSA was paused, AI Max activity increased sharply, effectively "taking over" the volume DSA had been driving (savvyrevenue.com, "DSA to AI Max: A Pragmatic Guide to Migrating Without Breaking Your Account," Jun 5, 2026).

The two systems appear to compete for the same auctions, and Google's own account logic suppresses AI Max while DSA is live. A meaningful test means pausing DSA first, not running both at once.

Google's Performance Claim vs. What Independent Testing Shows

The 7% Figure, and What It Actually Measures

Google's headline number has already moved once. At AI Max's May 6, 2025 launch, Google claimed a 14% lift in conversions or conversion value. In the April 15, 2026 retirement announcement, that figure was revised down to 7%.

The 7% claim carries three qualifiers worth knowing before you cite it internally:

  • It excludes retail advertisers.

  • It compares the full AI Max feature suite against search-term-matching alone, not against DSA as a whole.

  • It's footnoted as "Google internal data, 2026," with no published methodology.

The Independent Data

Four separate independent tests, run by different practitioners in different accounts, tell a less tidy story than Google's single figure:

  • Ezra Sackett, Director of Paid Search at Monks, found 99% of roughly 30,000 AI Max-expanded search terms generated zero conversions (ppc.land, Aug 17, 2025).

  • An independent study across 250+ retail campaigns found AI Max delivered roughly 35% lower ROAS than traditional match types (ppc.land, Nov 8, 2025).

  • Xavier Mantica's four-month test in a single account found AI Max cost $100.37 per conversion against $43.97 for phrase match, in the same account, over the same period (ppc.land, cited Apr 15, 2026).

  • A separate study from Smarter Ecommerce (Mike Ryan) found a median +13% revenue lift paired with +16% higher CPA, and ROAS swinging from +42% to -35% depending on the account (digitalapplied.com, "Google Delays the DSA-to-AI-Max Migration to Feb 2027," Jun 14, 2026).

None of this makes Google's 7% figure wrong. It makes it one data point among several, with the largest independent samples showing far more variance than a single average implies. Treat both Google's number and the independent numbers as directional, not decision-grade, until you've run your own test in your own account.

What to Actually Do Before Your Deadline Hits

Before September 2026, If You Run ACA or Broad Match

  1. Run a brand exclusion audit. Keywordless matching under AI Max can surface queries you'd never manually target, including competitor and brand-adjacent terms.

  2. Refresh your negative keyword list. This is your primary lever once matching broadens beyond your keyword set.

  3. Decide on final URL expansion explicitly. It's on by default. If your site has thin pages or inconsistent messaging, set URL exclusions before the switch, not after.

  4. Set text customization guidelines now. Expanded text controls give you more structured input on brand voice and what the AI avoids generating, but only if you configure them before the migration runs.

Before February 2027, If You Run DSA

  1. Pull a 90-day performance baseline from your current DSA campaigns: cost per conversion, conversion rate, and which landing pages actually convert.

  2. Audit your conversion tracking. Any test you run against that baseline is only as good as the tracking underneath it.

  3. Run an isolated test, not a side-by-side one. Based on the SavvyRevenue finding above, pause DSA in a controlled test account or ad group before turning on AI Max. A concurrent test will understate AI Max's actual behavior.

  4. Use the reopened creation window deliberately. It closes in January 2027. Treat it as your last window to compare a real DSA campaign against a real AI Max campaign, not as extra runway to delay.

If You're Exposed to Both Deadlines

Sequence September first. The brand exclusion and negative keyword work you do for ACA and broad match carries directly into your DSA test, since both run on the same keywordless matching engine underneath.

The Open Questions Google Hasn't Answered

Two practical questions remain genuinely unresolved, and Google has not addressed either directly.

Mixed campaigns. Many accounts run standard keyword ad groups alongside DSA ad groups in the same campaign. Whether the forced migration touches the entire campaign or only the dynamic ad groups within it is an open question raised by practitioners, including Joao de Braganca in a LinkedIn discussion covered by ppc.land. Until Google clarifies this, don't assume either outcome.

Historical DSA reporting. What happens to your existing DSA performance history once a campaign migrates, whether it stays queryable in the same reports or gets siloed, is also unclear. If year-over-year DSA reporting matters to your team, flag this now rather than discovering a gap after the migration runs.

Why This Needs Monitoring, Not Just a One-Time Checklist

Treating this migration as a pre-deadline scramble misses the actual shape of the problem. Google has already moved the DSA deadline once. The feature set changes underneath advertisers with little warning, and the independent data above shows performance swinging from +42% to -35% ROAS across otherwise comparable accounts.

That's not a one-time audit problem. It's a continuous account-hygiene problem: search-term drift as keywordless matching expands, CPA against your baseline as Google's models keep training, and negative-keyword coverage as new query patterns surface. A quarterly check misses the account that drifted 20% off baseline in week three.

This is exactly the kind of standing problem an AI marketing intelligence system like DOJO is built to sit on top of: continuous cross-account tracking of search-term drift and CPA against baseline, catching the erosion before it shows up in a monthly report instead of after.

FAQ

Do I need to do anything before September 2026 if I only run Dynamic Search Ads? No. DSA's deadline is February 2027, not September. But check whether you also run ACA or broad-match settings elsewhere in the account, since those follow the earlier deadline and are easy to miss if DSA is your main focus.

Can I still create new DSA campaigns right now? Yes. Google reopened DSA campaign creation from June 15, 2026 through January 2027, specifically to give advertisers a testing window before the forced February 2027 migration.

Will my negative keywords still work once AI Max is active? Yes. Google confirms negative keywords are respected under AI Max, but they carry more relative weight than before, because AI Max's broader keywordless matching means negatives are one of the few direct controls left on what your ads match to.

Does AI Max actually perform better than DSA? Google claims a 7% lift. Independent data is mixed, ranging from real gains to double-digit ROAS declines depending on the account and vertical. Treat it as untested in your account until you run your own isolated comparison.

If your account is exposed to either deadline, the dynamic search ads to AI Max migration is worth testing on your own terms now, while the reopened creation window is still open, rather than finding out what changed after Google migrates it for you.

Sources Cited

DSA to AI Max Migration: What to Do Before Feb 2027

Luke Costley-White

Adclear and DOJO AI partnership graphic: 'Close the loop on agentic marketing. Compliance at the speed of creation.'
二刀流
Two-Sword Style

Google is retiring Dynamic Search Ads. No new DSA campaigns can be created after the reopened creation window closes in January 2027, and all eligible DSA, ACA, and broad-match campaigns migrate automatically to AI Max by February 2027 (ACA and campaign-level broad match migrate sooner, by September 2026). Google claims a 7% conversion lift from the switch; independent tests show results ranging from strong gains to a 35% ROAS drop. Test before you're forced to.

If you're tracking the dynamic search ads to AI Max migration against a single date on your calendar, you're tracking the wrong one. Google has set two separate deadlines for two separate sets of campaigns, and most advertisers only know about the one that doesn't apply to them.

This isn't a one-time compliance task you clear off your list before a deadline and forget. It's a standing decision that arrives twice, with performance data that disagrees with itself depending on who ran the test. This guide covers what's actually changing, what the real numbers say, and what to do about it before either clock runs out.

What's Actually Changing, and When (The Two-Clock Problem)

Google announced the retirement of Dynamic Search Ads on April 15, 2026, with an update published June 11, 2026 (blog.google, "We're upgrading Dynamic Search Ads to AI Max"). The announcement bundles three separate campaign types into one migration, and they don't move on the same schedule.


Deadline

Applies to

Status

September 1, 2026

Automatically Created Assets (ACA) and campaign-level broad match

Unchanged since the original announcement

February 2027

Dynamic Search Ads (DSA) campaigns

Delayed once; creation window reopened June 15, 2026 through January 2027

The Sept 2026 Clock: ACA and Campaign-Level Broad Match

If you run standard Search campaigns with Automatically Created Assets turned on, or with broad match set at the campaign level rather than the keyword level, that migration date has not moved. It lands September 1, 2026 (Search Engine Land, Apr 15, 2026; PPC Land, Apr 15 and May 20, 2026).

The Feb 2027 Clock: Dynamic Search Ads

DSA campaigns run on a separate, later timeline, and it's already shifted once. Google originally set a September 2026 deadline for DSA too, then delayed it to February 2027 and reopened DSA campaign creation from June 15, 2026 through January 2027 (blog.google; PPC Land, Jun 13 and Jun 26, 2026). Google frames the reopened window as a deliberate side-by-side testing period before the forced migration.

Most advertisers only track one deadline and get blindsided by the other. An account running DSA in one campaign and campaign-level broad match in another is exposed to both clocks, on different dates, with different urgency.

What AI Max Actually Replaces, and What It Doesn't

AI Max bundles three features into existing Search campaigns: keywordless search term matching, AI-generated text customization, and final URL expansion. DSA only ever did one of those three jobs.

The Feature That Replaces DSA

Final URL expansion is the direct successor to DSA. It lets Google route traffic to whichever page on your site it judges most relevant, the same core mechanic DSA used to crawl your site and match queries to landing pages. Search term matching and text customization are new capability layered on top, not a replacement for anything DSA did.

Why Side-by-Side Testing Doesn't Work the Way You'd Expect

Google's own reopened creation window implies you can run DSA and AI Max side by side to compare them. In practice, that comparison breaks down inside a single account.

SavvyRevenue tested this directly and found AI Max generated just 72 clicks out of more than 3,000 total, while a DSA campaign was still active in the same account. Once DSA was paused, AI Max activity increased sharply, effectively "taking over" the volume DSA had been driving (savvyrevenue.com, "DSA to AI Max: A Pragmatic Guide to Migrating Without Breaking Your Account," Jun 5, 2026).

The two systems appear to compete for the same auctions, and Google's own account logic suppresses AI Max while DSA is live. A meaningful test means pausing DSA first, not running both at once.

Google's Performance Claim vs. What Independent Testing Shows

The 7% Figure, and What It Actually Measures

Google's headline number has already moved once. At AI Max's May 6, 2025 launch, Google claimed a 14% lift in conversions or conversion value. In the April 15, 2026 retirement announcement, that figure was revised down to 7%.

The 7% claim carries three qualifiers worth knowing before you cite it internally:

  • It excludes retail advertisers.

  • It compares the full AI Max feature suite against search-term-matching alone, not against DSA as a whole.

  • It's footnoted as "Google internal data, 2026," with no published methodology.

The Independent Data

Four separate independent tests, run by different practitioners in different accounts, tell a less tidy story than Google's single figure:

  • Ezra Sackett, Director of Paid Search at Monks, found 99% of roughly 30,000 AI Max-expanded search terms generated zero conversions (ppc.land, Aug 17, 2025).

  • An independent study across 250+ retail campaigns found AI Max delivered roughly 35% lower ROAS than traditional match types (ppc.land, Nov 8, 2025).

  • Xavier Mantica's four-month test in a single account found AI Max cost $100.37 per conversion against $43.97 for phrase match, in the same account, over the same period (ppc.land, cited Apr 15, 2026).

  • A separate study from Smarter Ecommerce (Mike Ryan) found a median +13% revenue lift paired with +16% higher CPA, and ROAS swinging from +42% to -35% depending on the account (digitalapplied.com, "Google Delays the DSA-to-AI-Max Migration to Feb 2027," Jun 14, 2026).

None of this makes Google's 7% figure wrong. It makes it one data point among several, with the largest independent samples showing far more variance than a single average implies. Treat both Google's number and the independent numbers as directional, not decision-grade, until you've run your own test in your own account.

What to Actually Do Before Your Deadline Hits

Before September 2026, If You Run ACA or Broad Match

  1. Run a brand exclusion audit. Keywordless matching under AI Max can surface queries you'd never manually target, including competitor and brand-adjacent terms.

  2. Refresh your negative keyword list. This is your primary lever once matching broadens beyond your keyword set.

  3. Decide on final URL expansion explicitly. It's on by default. If your site has thin pages or inconsistent messaging, set URL exclusions before the switch, not after.

  4. Set text customization guidelines now. Expanded text controls give you more structured input on brand voice and what the AI avoids generating, but only if you configure them before the migration runs.

Before February 2027, If You Run DSA

  1. Pull a 90-day performance baseline from your current DSA campaigns: cost per conversion, conversion rate, and which landing pages actually convert.

  2. Audit your conversion tracking. Any test you run against that baseline is only as good as the tracking underneath it.

  3. Run an isolated test, not a side-by-side one. Based on the SavvyRevenue finding above, pause DSA in a controlled test account or ad group before turning on AI Max. A concurrent test will understate AI Max's actual behavior.

  4. Use the reopened creation window deliberately. It closes in January 2027. Treat it as your last window to compare a real DSA campaign against a real AI Max campaign, not as extra runway to delay.

If You're Exposed to Both Deadlines

Sequence September first. The brand exclusion and negative keyword work you do for ACA and broad match carries directly into your DSA test, since both run on the same keywordless matching engine underneath.

The Open Questions Google Hasn't Answered

Two practical questions remain genuinely unresolved, and Google has not addressed either directly.

Mixed campaigns. Many accounts run standard keyword ad groups alongside DSA ad groups in the same campaign. Whether the forced migration touches the entire campaign or only the dynamic ad groups within it is an open question raised by practitioners, including Joao de Braganca in a LinkedIn discussion covered by ppc.land. Until Google clarifies this, don't assume either outcome.

Historical DSA reporting. What happens to your existing DSA performance history once a campaign migrates, whether it stays queryable in the same reports or gets siloed, is also unclear. If year-over-year DSA reporting matters to your team, flag this now rather than discovering a gap after the migration runs.

Why This Needs Monitoring, Not Just a One-Time Checklist

Treating this migration as a pre-deadline scramble misses the actual shape of the problem. Google has already moved the DSA deadline once. The feature set changes underneath advertisers with little warning, and the independent data above shows performance swinging from +42% to -35% ROAS across otherwise comparable accounts.

That's not a one-time audit problem. It's a continuous account-hygiene problem: search-term drift as keywordless matching expands, CPA against your baseline as Google's models keep training, and negative-keyword coverage as new query patterns surface. A quarterly check misses the account that drifted 20% off baseline in week three.

This is exactly the kind of standing problem an AI marketing intelligence system like DOJO is built to sit on top of: continuous cross-account tracking of search-term drift and CPA against baseline, catching the erosion before it shows up in a monthly report instead of after.

FAQ

Do I need to do anything before September 2026 if I only run Dynamic Search Ads? No. DSA's deadline is February 2027, not September. But check whether you also run ACA or broad-match settings elsewhere in the account, since those follow the earlier deadline and are easy to miss if DSA is your main focus.

Can I still create new DSA campaigns right now? Yes. Google reopened DSA campaign creation from June 15, 2026 through January 2027, specifically to give advertisers a testing window before the forced February 2027 migration.

Will my negative keywords still work once AI Max is active? Yes. Google confirms negative keywords are respected under AI Max, but they carry more relative weight than before, because AI Max's broader keywordless matching means negatives are one of the few direct controls left on what your ads match to.

Does AI Max actually perform better than DSA? Google claims a 7% lift. Independent data is mixed, ranging from real gains to double-digit ROAS declines depending on the account and vertical. Treat it as untested in your account until you run your own isolated comparison.

If your account is exposed to either deadline, the dynamic search ads to AI Max migration is worth testing on your own terms now, while the reopened creation window is still open, rather than finding out what changed after Google migrates it for you.

Sources Cited

Wired magazine logo

WIRED Magazine Top 100 Startups

Wired magazine logo

WIRED Magazine Top 100 Startups

g2 logo

G2.com 4.9 out of 5 stars

g2 logo

G2.com 4.9 out of 5 stars

Join over 100+ brands
already growing with us.

Join over 100+ brands
already growing with us.

Start a demo

FAQ

Frequently asked questions

Frequently asked questions

What is DOJO AI?

DOJO is an intelligent marketing system that watches every channel continuously, builds a living knowledge graph of your brand's marketing reality, and deploys specialised agents that execute work autonomously before you've had to ask. Not a tool. Not a platform. A system. Every signal your brand produces flows in, every action feeds back, and the system compounds its understanding over time. Most marketing software gives you data. DOJO gives you a system of record, context, and execution: one place where everything is captured, connected, and acted on. Instead of switching between Google Ads, Meta, LinkedIn, GA4, and social dashboards, you get one intelligent system that shows you what's working across all channels - and tells you exactly what to do about it. Specialized AI agents analyze your campaigns 24/7, identify opportunities competitors miss, and help you move faster than companies 10x your size.

Who is DOJO built for?

DOJO is built for marketing teams that want to spend their time on decisions that require human judgment, not on tasks that don't. If your team is stretched across too many channels, too many tools, and too many reports, DOJO replaces the operational burden with a system that runs continuously and arrives with work already done. It's used by in-house marketing teams, agencies managing multiple client accounts, and founders who want the output of a full marketing department without the overhead of one.

Is DOJO suitable for marketing agencies?

Yes. Agencies are one of DOJO's core use cases. The system connects across multiple client accounts, automates reporting and content production, and runs campaign monitoring continuously — so account managers spend time on client relationships and strategy, not on manual tasks that don't require their judgment. DOJO builds a separate knowledge graph for each client, so every recommendation and every piece of content is grounded in that client's actual brand history, not generic best practice.

How does DOJO work with existing tools?

DOJO connects to your existing channels through proprietary connectors and a live web crawler. Google Ads, Meta, LinkedIn, your website, brand mentions, competitor movements — everything flows in automatically, with no manual pulls required. You don't have to replace your stack to use DOJO. The system reads your existing data, connects it, and builds context on top of it. Over time, that context becomes the foundation for every recommendation and every action DOJO takes on your behalf.DOJO builds a separate knowledge graph for each client, so every recommendation and every piece of content is grounded in that client's actual brand history, not generic best practice.

What ROI can I expect?

DOJO customers typically see measurable cost reductions and efficiency gains within the first 90 days, with outcomes compounding as the system builds context over time. Here's what customers have reported: 79% drop in cost per acquisition(Morningstar) 3x conversion volumein the same 23-day window (Morningstar) 40% drop in acquisition costs(Broadvoice) 15x faster marketing reporting(Ozone API) 3x more efficient Google Adsquarter over quarter (Ecologi) 290% increase in content output(Broadvoice) 20 hours saved per month, returned to strategy (Morningstar) The compounding effect matters here. The longer DOJO runs, the more context it builds, and the more precisely it acts. Early results are strong; they get better.

How does DOJO compare to HubSpot, Jasper, or other AI marketing tools?

Most AI marketing tools fall into one of two categories: workflow automation (HubSpot, Marketo, ActiveCampaign) that executes campaigns you set up, or content generation (Jasper, Copy.ai) that produces copy on demand. Both share the same limitation: they start from scratch every session. No memory of your brand history, your previous campaigns, or what your competitors have been doing. DOJO maintains a continuously updated knowledge graph of your entire marketing reality and runs specialised agents that read it daily, surface what needs attention, and execute work before you've asked. The longer DOJO runs, the more precisely it acts — because it compounds what it learns about your specific brand, market, and competitors. If you're evaluating options: Email and workflow automation: HubSpot, Klaviyo, Marketo AI content writing: Jasper, Copy.ai A system that watches every channel, builds brand context, and executes proactively: DOJO

Does AI marketing software actually improve over time, or does it reset every session?

Most AI marketing software resets every session. It has no memory of your brand, your campaigns, or what worked before. Every interaction starts from a blank slate. DOJO works differently. Every signal it captures, every workflow it runs, every recommendation it makes is fed back into the DOJO Graph. The system learns what works for your specific brand, in your specific market, against your specific competitors. It builds institutional knowledge that no other system carries. A team that's been using DOJO for six months has a system that understands their brand history, their campaign patterns, and their market in detail. That depth of context changes what the agents can do. The advantage grows every day the system runs, and it never stops running.

How does DOJO handle data security and privacy?

DOJO is built on enterprise-grade infrastructure with security and data privacy at its core. Your brand data, campaign history, and knowledge graph are kept entirely separate from other customers' data. For detailed information on data handling, storage, and compliance, see our Privacy Policy and Data Processing Agreement, or speak to our team directly when you book a demo.

Join over 100+ brands already growing with us.

Book a demo

Wired magazine logo

WIRED Magazine Top 100 Startups

Wired magazine logo

WIRED Magazine Top 100 Startups

g2 logo

G2.com 4.9 out of 5 stars

g2 logo

G2.com 4.9 out of 5 stars