Google’s AI Max for Search moves campaign steering away from keywords and toward goals and signals. Google moved it out of beta in April 2026 and is retiring Dynamic Search Ads in its favour. For businesses evaluating agentic advertising, it is the clearest large-scale example of platform automation taking over targeting and copy decisions, and a useful test of what to hand over.
- Google reports that AI Max campaigns already using search term matching see an average 7% more conversions or conversion value at similar CPA or ROAS when text customization and final URL expansion are enabled (Google internal data, global, 2026, non-retail).
- Google gives no sample size or control-group detail for that figure, so treat it as a vendor claim to test, not a forecast.
- Automatic upgrades of campaigns using Automatically Created Assets and campaign-level broad match are reported for September 2026, with Dynamic Search Ads upgrades pushed to February 2027.
- AI Max is platform automation, not an autonomous agent: you still set goals, exclusions, and budget limits.
What AI Max actually changes
AI Max bundles three behaviours that used to be separate manual levers. According to Google’s AI Max for Search announcement page, it uses “broad match and keywordless technology” to find queries an advertiser’s keyword list would not capture, optimises ad text based on a customer’s interests and context, and sends people to what Google calls the most relevant page on the site.
In practice that means three named features: search term matching, text customization, and final URL expansion. Together they shift the unit of control from the keyword to the goal and the landing-page inventory you allow.
The implication for a buyer is that your account structure matters less than your conversion tracking and your site content, because those are now the inputs the system learns from.
| Feature | What it does | What it replaces | What to check |
|---|---|---|---|
| Search term matching | Finds relevant queries beyond your keyword list | Manual keyword expansion | Search terms report for irrelevant queries |
| Text customization | Adapts headlines and descriptions to the query and context | Hand-built ad variants per intent | Brand voice and claims accuracy |
| Final URL expansion | Chooses a landing page from your site | Fixed final URLs per ad group | Which pages receive traffic |
What Google’s own numbers do and don’t show
Google states that AI Max for Search campaigns already using search term matching see an average of 7% more conversions or conversion value at a similar CPA or ROAS by enabling text customization and final URL expansion. The footnote reads “Google internal data, Global, 2026; for non-Retail advertisers.” The page gives no sample size, time period, or control group.
Two points follow. First, the claim is conditional: it concerns advertisers who already used search term matching and then added two further features, so it does not describe the lift from switching AI Max on from scratch. Second, an average hides spread. A 7% mean can sit alongside accounts that improved sharply and accounts that got worse. Larger figures circulate in vendor commentary, but we could not trace them to Google’s own page, so we have left them out.
The migration timeline
Timing matters because some upgrades happen whether or not you opt in. Agency reporting from Adtrak says Google confirmed on 15 April 2026 that AI Max will replace Dynamic Search Ads, covering DSA campaigns, campaigns using Automatically Created Assets, and campaigns using campaign-level broad match. Search Engine Land later reported that Google extended the automatic DSA migration to February 2027 after advertiser feedback, in an announcement dated 11 June 2026, while the Automatically Created Assets and broad-match upgrades stay on the September 2026 schedule.
Dates in this area have already shifted once, so check Google’s current documentation before planning around them. The practical advice is to migrate voluntarily on your own terms, with your own controls set, rather than discover the change in a performance report.
Which decisions you keep
AI Max is an automation layer inside a human-set frame, which is what separates it from the autonomous systems covered in our breakdown of what is actually autonomous in agentic advertising. Adtrak’s guidance notes that advertisers who migrate themselves can choose which brand terms to include or exclude, set URL inclusions and exclusions, and decide which AI Max features to enable.
Google’s page also says text disclaimers, introduced in May 2026, ensure legally required text keeps appearing even with final URL expansion. That matters in regulated categories. The remaining decisions stay with you: the conversion event you optimise toward, the budget ceiling, and the pages and claims you are willing to have shown. Poorly defined conversions are the biggest risk, and the lessons from performance metrics that matter more than CTR apply directly.
A staged test plan
A cautious rollout limits downside while producing evidence from your own account, not from Google’s averages.
- Audit conversion tracking first, and confirm that the primary conversion reflects revenue or qualified leads, not clicks or page views.
- Pick one stable campaign and create an AI Max experiment against the existing setup, with an even traffic split.
- Enable features one at a time, starting with search term matching, then text customization, then final URL expansion.
- Set brand exclusions and URL exclusions before launch, and review the search terms report weekly.
- Compare cost per qualified lead or ROAS against the control for at least one full conversion cycle before expanding.
If you use AI-generated copy elsewhere, read the evidence in what the performance data shows on AI-generated ad creative before letting text customization run unchecked.
What the evidence doesn’t support
The 7% figure is Google’s own, unaudited, and conditional on prior feature use. We found no independent study with disclosed methodology that we could verify, so we do not claim AI Max reliably beats a well-maintained keyword structure. Some commentary cites a 14% conversion lift or a share of advertisers seeing neutral or worse results; the 14% figure does not appear on Google’s announcement page, and the other came from a vendor blog without a traceable method, so we excluded both. Migration dates are as reported by trade press and have already changed once.
Frequently Asked Questions
Is Google AI Max the same as an autonomous AI agent?
No. AI Max automates query matching, ad text, and landing-page selection inside limits you set. It does not choose your goals, budget, or offers, which an autonomous agent would also have to decide.
Do Dynamic Search Ads campaigns have to move to AI Max?
Eventually, yes. Trade reporting says automatic DSA upgrades are now scheduled for February 2027, with other upgrades earlier. Confirm the current dates in Google Ads documentation before planning.
Should a small business switch AI Max on now?
Only after conversion tracking is reliable. Test it against your current setup in one campaign, with brand and URL exclusions in place, and judge it on cost per qualified lead.
Deciding how much of your paid search to hand to platform automation is a measurement problem as much as a technology one. At Grow with Hafsa we help teams set up the tracking, experiments, and guardrails that let AI-driven performance marketing be judged on results from their own account.
Automated campaign types depend on the data you feed them. Our piece on agentic commerce protocols and what they mean for your product feed covers the catalog fields that OpenAI and Google now require for AI shopping.
Meta has a parallel shift underway: see what you still control in Meta Advantage+ when the platform automates audience, placement and budget.
Google has since begun testing ad formats inside AI Mode, covered in ads inside AI answers and what they mean for advertisers.