Use this banner section for site-wide announcements, news updates, big changes and more.
.svg){kind=logo}
If you built your Q4 plan around a September deadline for Dynamic Search Ads, you need to change two things. The DSA deadline moved to February 2027. And a different September deadline, the one almost nobody is talking about, is still live and probably affects more of your account.
That gap between what advertisers think is happening and what is actually happening is where the damage gets done. By the end of this you will know exactly which parts of your account change in September, which change in February, what breaks specifically in B2B accounts when Google flips the switch for you, and how to run a migration test that produces a real answer rather than a noisy one.
On 15 April 2026, Google announced that Dynamic Search Ads were being retired and moved AI Max for Search out of beta into general availability. The original plan was blunt: from September 2026, every eligible campaign running DSA, automatically created assets (ACA) or campaign-level broad match would be force-migrated into AI Max, with no opt-out and no ability to create new DSA campaigns afterwards.
Then advertisers pushed back. On 11 June 2026, Google Ads Liaison Ginny Marvin confirmed the DSA auto-migration was being pushed from September 2026 to February 2027, and that the ability to create new DSA campaigns would be restored from 15 June. Google's stated reason was advertiser feedback and a desire to avoid disrupting Q4 planning.
Here is the part that gets missed. The delay applies to Dynamic Search Ads only. Automatically created assets and the campaign-level broad match setting still migrate to AI Max on the original September 2026 timeline.
For most B2B accounts, the nearest deadline is measured in weeks, not months. If you have automatically created assets switched on anywhere, or campaign-level broad match applied at the campaign setting level, those campaigns are getting AI Max features enabled this autumn whether you have tested them or not.
There is a related signal worth registering. Google rewrote its Ads Terms of Service on 1 July 2026, its first substantive update since 2018, and the new language explicitly authorises Google's automated systems to format, select or generate targets, ads and destinations on the advertiser's behalf. AI Max is also now the default setting when you create a new Search campaign. The direction of travel is not ambiguous.
AI Max is not a new campaign type. That distinction matters more than it sounds, because it changes how you think about testing and rollback. It is a feature layer that switches on inside your existing Search campaigns, built from three components.
Search term matching expands your reach beyond your keyword list using broad match and keywordless technology, matching queries based on intent signals rather than the terms you have explicitly added.
Text customisation uses Gemini to generate headlines and descriptions in real time, pulling from your landing page content and existing assets to write copy tailored to the individual query.
Final URL expansion lets Google choose the landing page. Rather than sending traffic to the URL you specified, it selects whichever page on your site it judges most relevant to the search, which is the mechanic that most closely replicates old DSA behaviour.
Alongside those, AI Max carries controls that Performance Max never gave you: brand inclusions and exclusions at campaign and ad group level, URL exclusions at campaign level and URL inclusions at ad group level, and locations of interest for targeting geographic intent rather than physical location. You also keep full search term reporting, including visibility into which headlines ran and which landing pages were served. Google added further controls in July 2026, including text guidance and URL exclusions for automatic landing page expansion.
That reporting layer is the single biggest practical difference between AI Max and Performance Max, and it is why the migration is survivable. You can see what the automation did. Whether you act on what you see is a different question.
The defaults differ depending on what triggered your migration, and they are not conservative.
Structurally, dynamic ad groups become standard ad groups and historical performance data is preserved in place. That sounds clean. In practice, forced migrations rarely inherit the nuance of a well-built account. Carefully segmented page feeds tend to get superseded by broad final URL expansion. Negative keyword logic built around dynamic targeting does not always map cleanly onto keywordless matching. Anyone who lived through the Smart Shopping to Performance Max transition in 2022 will recognise the pattern, where advertisers who waited absorbed two to four weeks of instability and advertisers who migrated early largely did not.
Google's published figures are respectable. Enabling AI Max features together delivers roughly 7% more conversions or conversion value at similar cost per acquisition compared with using search term matching alone, and case studies like Australian utility connection service MyConnect report 16% more leads at 13% lower cost per action.
Those are averages, and averages are built mostly on ecommerce and consumer lead gen. B2B SaaS and FinTech accounts have four structural characteristics that make the migration behave differently.
A typical ecommerce site is 80% product pages. A typical B2B SaaS site is 80% content: blog posts, glossary entries, comparison pages, help documentation, changelogs, webinar replays, careers listings, customer stories. Every one of those is a candidate landing page once final URL expansion is on.
The failure mode is not that the traffic is irrelevant. It is that a genuinely high-intent buyer searching a bottom-of-funnel query gets routed to a 2,000-word educational blog post with no demo form on it, converts at a fraction of the rate your pricing page converts at, and Smart Bidding quietly learns that the query is unprofitable. You lose the click and the signal.
Support documentation and careers pages carry a second problem. They attract existing customers and job seekers, both of whom convert on nothing and cost you money in a category where clicks are rarely cheap.
Start with final URL expansion off, or restricted to a vetted page feed, and expand from there. This is the single highest-leverage decision in the entire migration for a B2B account.
Gemini writing your headlines from landing page content is efficient. It is also a claims risk in any regulated or contractual category.
In FinTech, that means generated copy touching on regulatory status, returns, protections or eligibility. In SaaS, it means implied claims about pricing, free trials that do not exist, integrations you do not support, or security certifications you have not completed. AI-generated ad copy is your legal responsibility, not Google's.
Before migrating, audit which of your responsive search ad assets are pinned for legal, pricing or promotional reasons, and treat text guidelines as a compliance control rather than a tuning knob. If your brand or legal team has ever reviewed ad copy line by line, they need to be in this conversation before September.
B2B category terms are ambiguous in ways consumer terms usually are not. Queries around "payment platform", "workflow automation" or "compliance software" sit across enterprise buyers, small business owners, students writing dissertations, developers looking for open-source alternatives and jobseekers. Keyword match types were doing quiet work distinguishing between them.
Keywordless expansion removes that filter and replaces it with intent modelling trained on aggregate behaviour, which is thinner in low-volume B2B categories than it is in retail. Brand controls become genuinely important here, both for excluding competitor associations and for keeping your ads away from adjacent brands you do not want to appear next to.
Expect to work your search terms report harder than you did with DSA, not less. The queries you will need to negative out are ones you have never seen before.
This is the one that matters most and gets discussed least. In B2B lead generation, conversion volume frequently increases under AI Max while lead quality declines, and you will not detect it unless offline outcomes are feeding back into the platform.
If Smart Bidding is optimising towards form fills, it will find you more form fills. With a 90 to 180 day sales cycle, the CRM data that tells you whether those form fills became qualified pipeline arrives a full quarter after the bidding decisions were made.
Offline conversion import is not an optimisation nicety for this migration. It is the prerequisite. Feed SQLs, opportunities and closed revenue back into Google Ads before you enable expanded matching, not after. An account that optimises on CRM-verified pipeline can absorb a broader query surface. An account optimising on raw form fills will scale its own noise.
The default advice on this migration is "test for two to four weeks". That advice was written for accounts with hundreds of conversions a week. If your campaign generates 40 conversions a month and your qualification signal lags by a quarter, a four-week test tells you very little.
Here is a sequence that works with B2B PPC economics.
You cannot evaluate a migration without knowing what you are migrating from, and the window to capture clean pre-migration data is closing.
Export campaign, ad group, search term, landing page, device and geography performance for every DSA, ACA and campaign-level broad match campaign in the account. Record CPA, ROAS, conversion rate, and critically the downstream metrics: qualified lead rate, SQL-to-opportunity conversion, average contract value. Document your page feeds, custom labels, negative keyword lists and URL exclusions before anything reshuffles them.
Then audit the site itself. List every URL that should never be a landing page and build that exclusion list now. Careers, support docs, thin blog content, outdated pricing pages, anything with broken tracking. This list is your primary defence against final URL expansion.
Confirm offline conversion import is live and accurate. Check that CRM stages map cleanly to conversion actions and values. Verify tracking templates handle dynamic landing pages without breaking attribution, because final URL expansion will send traffic to pages your tracking template may never have been tested against.
If measurement is broken, migrating first and fixing later means the automation spends three months learning from bad data. That is not a delay you can undo.
Do not enable all three features at once and try to read the result. Pick two or three campaigns covering different use cases, one high-volume, one niche, one heavy page-feed user, and layer features in sequence.
Start with search term matching only, at ad group level where you want to be aggressive and off where you do not. Add text customisation next, with text guidelines set conservatively and pinned compliance assets protected. Add final URL expansion last, restricted to a vetted URL inclusion list rather than the whole site.
Judge on CPA and ROAS, not CPC. Cost per click will rise as you enter broader auctions, and that is a mechanical consequence of expanded matching rather than evidence of worse efficiency. Where conversion volume is too thin for statistical confidence inside a quarter, use conversion value and pipeline contribution as your read, and be honest about the confidence interval rather than declaring a winner from 14 days of data.
If you sell into EMEA, North America and APAC from one account, locations of interest is the AI Max control most worth understanding. It targets geographic intent at ad group level rather than physical presence, which means you can reach a buyer researching your UK entity from a laptop in Singapore.
That is useful. It also creates new ways for market segmentation to leak. Expanded matching combined with intent-based geography can pull queries across market boundaries you deliberately separated for currency, language, compliance or pricing reasons. If your German landing pages carry different regulatory disclosures from your UK ones, final URL expansion crossing that boundary is a compliance event, not a performance one.
Segment testing by market. A migration result from your UK campaigns tells you very little about how the same settings will behave in a market where your content depth, brand recognition and competitive density are different.
When exactly do my Dynamic Search Ads stop working?
They do not stop working in September. New DSA campaign creation ends in January 2027, and automatic migration of remaining DSA campaigns to AI Max begins in February 2027. Campaigns using automatically created assets or campaign-level broad match migrate earlier, in September 2026.
Can I opt out of the AI Max migration?
No. Every affected campaign ends up on AI Max. The only variable is whether you configure the settings or Google's defaults do.
Should I migrate voluntarily or wait for the automatic upgrade?
Migrate voluntarily. Voluntary upgrade preserves historical data in the same ad groups and lets you choose which features are enabled and how aggressively. The automatic upgrade turns everything on with default settings built for the average advertiser, and B2B accounts are not the average advertiser.
Will AI Max hurt my lead quality?
It can, particularly if Smart Bidding is optimising towards form fills rather than CRM-verified outcomes. Accounts that import offline conversions and optimise on SQLs and pipeline tend to hold quality through the transition. Accounts optimising on top-of-funnel conversion actions tend not to.
What happens to my page feeds?
Dynamic ad groups convert to standard ad groups with AI Max features enabled. Page feeds can still be used as a targeting input, and for B2B accounts starting conservatively, page feed targeting combined with URL inclusions is closer to the control you had under DSA than open final URL expansion is.
Does AI Max replace keywords entirely?
No. Keywords remain a foundational intent signal that steers the system toward the audiences you want. Broad match keywords plus Smart Bidding is the recommended base configuration, with AI Max layering query expansion and creative generation on top.
Google gave the industry an extra five months on DSA. The trap is reading that as a problem deferred rather than a problem clarified, shelving the work, and arriving in January with the same amount of preparation done as you had in June.
What the extension actually bought is the ability to migrate on evidence rather than on a deadline. Between now and February you can build a clean baseline, fix the offline conversion plumbing that B2B lead quality depends on, and run staged tests long enough for a real sales cycle to report back. That is not glamorous work. It is the difference between AI Max finding you incremental qualified pipeline and AI Max finding you a larger volume of leads your sales team will not take.
If you are running meaningful monthly spend across Google and want that migration handled on your terms, with CAC and pipeline as the measure rather than conversion count, get a free proposal and we will audit what is exposed in your account before September makes the decision for you.
Lever Digital is proud to be a 2026 UK Paid Media Awards finalist, recognised for outstanding performance-led paid media campaigns across B2B and SaaS.
.png)
