The headlines said Google blinked. In mid-2026, after advertiser pushback, Google pushed the forced migration of Dynamic Search Ads to AI Max from September 2026 out to February 2027. Search Engine Land reported Google's own words: "We've heard your feedback loud and clear: you need more time to transition from Dynamic Search Ads (DSA) to AI Max." For a lot of B2B SaaS teams, that read as permission to close the tab and deal with it next year.
That is the trap. The delay applies to DSA campaigns only. Two other surfaces — automatically created assets (ACA) and the campaign-level broad match setting — still auto-upgrade to AI Max on the original September 2026 schedule. And those two surfaces, not DSA, are what most modern B2B SaaS lead-gen accounts actually run. So the accounts least likely to run legacy DSA campaigns are the accounts most likely to be quietly migrated in September, precisely because their owners believed the deadline moved. This post lays out what actually flips in September, why it lands hardest on B2B SaaS, and the baseline-and-guardrail work to do in the next few weeks. If you want the DSA-specific walkthrough, our DSA to AI Max migration guide covers that path; this post is about the surfaces the delay did not cover.
The split deadline nobody read past
Google's migration is not one event on one date — it is three surfaces on two timelines. The DSA timeline is the one that moved: new DSA campaign creation now runs through January 2027, and automatic migration of any remaining DSA campaigns begins February 2027. But the September 2026 timeline for automatically created assets and campaign-level broad match held firm. As Search Engine Journal noted in its coverage of the extension, the reprieve was scoped to DSA while the ACA and campaign-level broad match transition to AI Max continues on its original September schedule. One announcement, two clocks — and the faster clock is the one most people stopped watching.
This is why "the deadline moved" is a dangerous summary. It is true for the surface fewest B2B SaaS accounts depend on and false for the two surfaces most of them run every day. The result is an attention mismatch: the change that got the press coverage (DSA sunset) is now the least urgent, and the change that got almost none (the September ACA and broad match flip) is the one with a four-week fuse. If you manage a SaaS account and your mental model is "AI Max is a 2027 problem," the calendar disagrees.
What actually flips in September
Two mechanics change behavior. The first is automatically created assets — the feature that lets Google generate additional headlines and descriptions by reading your landing pages and existing ads. Under AI Max, ACA becomes part of a broader automated-creative system that can also expand the final URL a click lands on. The second is the campaign-level broad match setting, the toggle that lets a Smart Bidding campaign match queries beyond its explicit keyword list. Under AI Max this becomes intent-based search-term matching — the same query-expansion engine that made DSA useful, now layered onto your standard Search campaigns.
Put those together and a campaign that was a reasonably bounded, keyword-anchored Search campaign becomes something closer to a semi-automated one: broader query reach plus machine-assembled creative, both steering toward whatever conversion you have defined. That is not inherently bad — it is the same trade every automation layer offers, reach and efficiency in exchange for control. But it is a real behavioral change to live campaigns, triggered automatically, on a date most teams are not tracking. The difference between capturing the upside and importing waste is entirely in the guardrails you set before the flip, which is the whole argument for doing this work in August rather than discovering it in October.
Why the September flip hits B2B SaaS hardest
B2B SaaS keyword spaces are dense with adjacent-but-worthless queries: "free alternative to X," "is Y worth it," "how to do Z in a spreadsheet," "Z jobs," "Z tutorial." Tight keyword campaigns keep those out by construction. The moment campaign-level broad match becomes AI Max query expansion, the campaign starts reaching for intent it thinks is related — and in SaaS, "related" frequently means job seekers, students, free-tool hunters, and competitor researchers. This is the same dynamic we flag in our negative keywords guide: the broader the matching, the more expensive a weak negative list becomes, because a mediocre exclusion set that was tolerable under exact and phrase match turns into a budget leak under expansion.
The deeper problem is measurement, and it is where SaaS is structurally exposed. AI Max optimizes toward whatever you call a conversion, and B2B SaaS funnels are long with wildly variable lead quality — a newsletter signup, a trial start, a sales-qualified demo, and a closed deal are all "conversions" that mean completely different things. If your campaigns optimize toward cheap top-funnel form fills, AI Max's expansion will find you more cheap top-funnel form fills, and your reported CPA will look stable while pipeline quality erodes underneath it. That is the exact trap behind our argument to optimize for SQLs, not leads — and automation makes the cost of measuring the wrong thing much higher, because the algorithm scales whatever you reward.
Lock a baseline before the flip
The single most valuable thing you can do before September is capture a clean before-picture, because after the flip you will not be able to reconstruct it. For every campaign that uses ACA or campaign-level broad match, record the metrics that actually matter for SaaS: cost per SQL and cost per pipeline dollar (not just cost per lead), the search-term report with your current mix of on- versus off-target queries, and your existing conversion-rate and lead-quality distribution. Export the search-term report specifically — it is your evidence of what query quality looked like before AI Max started expanding, and it is the first thing to check for drift afterward.
Baselines only work if they measure downstream revenue, not surface activity. A campaign can post more conversions and a lower CPA after the flip while sending you worse pipeline, and the only way to catch that is to have written down what good looked like on revenue-quality terms beforehand. This is the same evaluation rigor we apply to auditing Performance Max campaigns: automated campaign types must be judged on downstream quality because their surface metrics flatter them. Set a calendar reminder to re-pull the same metrics two and four weeks after September's transition, and compare against the frozen baseline rather than against your gut sense of how things are going.
Tighten the guardrails AI Max will inherit
AI Max is not a black box the way early Performance Max was — it inherits the controls you configure, and the September flip is your prompt to configure them properly. Refresh negative keyword lists first, since broader matching is where a stale list gets expensive. Then build brand inclusion and exclusion lists so query expansion respects your competitive posture rather than bleeding spend onto competitor-confused or irrelevant brand queries — the same discipline behind our competitor analysis for SaaS guide. If your account also runs newer AI Max controls, our breakdown of the new AI Max controls for B2B SaaS covers the URL-expansion and location-of-interest settings worth setting before, not after, the switch.
The non-negotiable guardrail is conversion tracking. Confirm your campaigns optimize toward events that correlate with revenue — qualified demos, opportunities, offline-imported deals — rather than raw form fills, because that definition is what AI Max will faithfully scale. If your tracking still counts every lead equally, fixing that is more urgent than any bid or budget tweak, and our conversion tracking for SaaS guide walks through the setup. The order of operations matters: guardrails and measurement first, then let the flip happen against a system that rewards the right behavior.
Your four-week pre-September checklist
Work backward from September. This week, inventory every campaign that uses automatically created assets or the campaign-level broad match setting — that list is your exposure, regardless of whether you run a single DSA campaign. Next, freeze a baseline for each: export search-term reports and record cost per SQL and cost per pipeline, not just CPL. Then audit guardrails — negatives, brand lists, URL expansion controls — and verify conversion tracking rewards revenue-quality events. None of this is exotic; it is the standard hygiene most accounts skip until an automated change forces the issue, and this is that forcing function with a date on it.
Do not confuse the DSA extension with breathing room on the surfaces that actually flip in September. The teams that come through this cleanly are the ones that treated August as the deadline it effectively is for ACA and broad match, set their guardrails, and locked a baseline they can measure drift against. If you would rather not run this blind, our Google Ads audit can pressure-test which of your campaigns flip in September, whether your negative and brand lists are ready for query expansion, and whether your conversion tracking rewards pipeline over form fills — before the switch happens on Google's schedule instead of yours. For the data behind the upgrade, our look at AI Max search performance data for B2B SaaS sets realistic expectations for what the flip actually changes.