Google AI Mode Ads: 0 of 9 Controls Act Alone, and the 4 Tests Worth Running

Google took ads in AI Mode to 10 more countries on 28 October 2026, and its help page still says you "can't target ads exclusively to AI Overviews or AI Mode, or opt out of showing ads in them separately from Search". We audited 9 Google AI Mode ads controls and reporting views in Google's documentation. None acts on AI Mode alone. One changes how you are eligible, and 4 tests can still tell you something true.

The short answer

You cannot opt out of Google AI Mode ads without leaving Search, and Google Ads gives you no separate AI Mode report. Negative keywords, brand exclusions and match types still shape delivery across Search. Measure total campaign incrementality with Conversion Lift or a match type experiment, and treat any AI Mode specific number as unproven.

What Google changed for AI Mode ads in October

The news is scale, not a new set of controls. Google's 28 October post names 10 countries where it is "scaling ads in AI Mode". It adds that in Q4 2026 it plans to expand to more locales, including Brazil, Mexico and the United Kingdom. The word "including" leaves that list open.

The same post adds a format change for one market. "To allow ads in follow-up queries in AI Mode, we are launching multiturn ads in the U.S." That sentence matters for measurement. An ad can now appear several steps into a conversation that began with something else.

Nothing in the post addresses advertiser controls, opt-out, reporting or eligibility. Those rules live in Google Ads Help, which is where this audit went.

Why Google has a reason to move quickly

Philipp Schindler told investors on the Q3 call that Google is "scaling high-quality ads within AI Overviews and AI Mode to more users and queries". In the Q&A he gave the reason in plain terms. "The strong growth in AI Mode queries had an impact because of the lower monetization we were seeing in AI Mode," he said, per the published transcript.

He named currency as the single largest driver of the slowdown, so AI Mode was one factor among several. Still, our reading is simple. A surface that earns less than the one it partly replaces will get more ads, and your campaigns are the supply. How that shows up in clicks and prices at company level is covered in the Q3 paid clicks and cost-per-click split.

Google's claim you cannot check

Schindler also said "the majority of AI Mode queries are incremental, creating more monetization opportunities over time". Google has not published the method behind that sentence. Incremental for Google's query count is not the same as incremental for your revenue, and only the second one should move your budget.

AI Mode advertiser controls: the 9-row audit

We went through Google's help page on ads in AI-powered experiences line by line on 8 November 2026. The test for each row was strict: does this control or view act on AI Mode without also acting on the rest of Search?

Our count and our test, applied to Google's wording. The final column reflects what the page documents, and an undocumented account-level beta would not appear here.

Read the table as a map of blunt instruments. Your negative keyword list still works, and that is genuinely useful when AI Mode surfaces a query you would never pay for. It also removes that query from every other Search placement, so it cannot express a preference about the AI surface itself.

The last row deserves more attention than it gets. Google says some placements can show an AI explainer, generated by Gemini, next to your ad. You "can't manually write, edit, approve, or opt out of them," so a regulated advertiser has text appearing beside its ad that its compliance team cannot review in advance. If you sell in a regulated category, we would raise that with your Google representative in writing before the rollout reaches your market.

Pinning needs one caveat. Google lists it as a guardrail, but its effect depends on your settings, as our review of what AI Max takes away from keyword control explains for campaigns using final URL expansion.

Google Ads AI reporting: what the data can show

Google's reporting language is short and specific. Performance for ads in AI Overviews and AI Mode is "included in aggregate" in your standard campaign reports, across metrics such as impressions, clicks, conversions and cost.

The search terms report blends the surfaces

The search terms report gets the same treatment. Google writes that "queries that trigger your ads in AI Overviews and AI Mode are included in aggregate in your search terms report". You can see a query. Nothing in the documentation says which surface that query served on.

Our earlier comparison of ChatGPT, Google and Copilot ad rules treated this as one row in a cross-platform table. The practical consequence goes further. Spend on AI Mode cannot be reconciled from your own reports, so nobody outside Google can tell you what share of your budget went there.

The Top Ads sentence has gone missing

Search Engine Roundtable copied a May 2025 version of Google's help page. It said ads in AI Overviews "are reported as Top Ads" and that Google "currently doesn't offer segmented reporting". The page we read on 8 November 2026 no longer repeats the Top Ads line. It links instead to a page on top and absolute top metrics, which does not mention AI Overviews or AI Mode.

That may be an editorial trim rather than a change in behaviour, and we would not read much into it. The practical point is narrower. Do not build a report that assumes the top-of-page segment is your AI Mode segment. Google's current page no longer says what is in it, and even the old wording described a blend with classic top ads. The organic equivalent of this gap is rated in our look at Search Console AI Mode data.

Match type is the only documented eligibility lever

Google describes 2 ways an ad can reach these surfaces. Direct query matching applies when "the search already expresses clear commercial intent". Contextual matching uses both the query and the AI response, for "a longer, exploratory, or conversational question".

The split between campaign types is where the only real choice sits. Google says exact and phrase keywords on Search "aren't eligible for contextual matching on broader exploratory questions". AI Max, broad match, Shopping and Performance Max are eligible for both routes.

What this lever will not do

Moving a Search campaign to exact and phrase is not an opt-out. A query with direct commercial intent can still serve your ad in AI Mode. Google Ads Liaison Ginny Marvin said on 4 September 2026 that "exact and phrase match keywords are eligible to serve text ads in AI Mode," limited to cases of "explicit and direct user intent".

You also lose whatever broad matching was earning you on classic Search. That cost will usually be larger than any AI Mode exposure you remove. We would not restructure an account purely to dodge AI Mode. The evidence for doing it does not exist yet.

Shopping and Performance Max have no equivalent

The two campaign types most exposed to product queries have no documented match type setting to change. For those campaigns your options are the campaign-wide ones in the audit table, and the expansion to more markets reaches them automatically. Google says existing Search, Shopping and Performance Max campaigns "are automatically eligible".

Four incrementality tests that remain valid

An incrementality test compares what happened with an ad against what would have happened without it. None of the pages below describes a split by AI surface. Each test is valid for the narrower question it can actually answer.

Test 1: Conversion Lift based on users

This measures whether your campaign as a whole causes extra conversions. Google's set-up page lists Search and Performance Max among supported campaign types, and requires at least 1,000 observed conversions and a minimum campaign budget of $5,000. Google allows studies as short as 7 days and typically recommends more than 14.

Access is the catch. Google's overview says Conversion Lift "isn't available for all Google Ads accounts" and that you should contact your account representative. Run it twice around a rollout if you can. A change in campaign lift is the nearest thing to an AI Mode reading this tool can give.

Test 2: Conversion Lift based on geography

Geography suits B2B better. Google says this version uses aggregated geographical data and supports offline data, which lets you feed in qualified pipeline. The geo page requires a campaign targeting a single country and uses Google Marketing Areas as units. In this beta you can run only holdback and go-dark studies.

A go-dark study switches ads off in chosen regions, so it measures the whole campaign, with AI Mode blended in. We would pick this one first for a long sales cycle. A user holdout reads online conversions faster than your buyers actually decide.

Test 3: A match type custom experiment

This is the only test here that changes contextual eligibility on purpose. Google's custom experiment page says the base campaign must use Smart Bidding and include exact or phrase keywords. Build a control with exact and phrase, and a trial that adds broad match.

Use the cookie-based split, which Google offers alongside a search-based split for Search. Our reasoning is about multiturn ads: a search-based split can put one person in both arms during a single conversation. The result combines AI surface contextual matching with classic broad match reach, so report it as the value of that bundle. Google's page does not describe match type changes specifically, so confirm the trial saves before you plan around it.

Test 4: A market launch comparison

Google has named markets it plans to reach in Q4, which creates a rare before and after. If you already run comparable campaigns in the UK, Brazil or Mexico and in a market not on Google's list, record weekly cost per click, top impression rate and qualified conversions now. Compare how the launch market changes against the comparison market.

This is the weakest of the 4 and the cheapest. Google has not published launch dates by market. Currency and seasonal demand will move both sides in Q4, and the rollout may be gradual. Treat a result as a reason to run Test 1 or Test 2, and never as the finding itself.

AI Mode ad tests that look valid and are not

The tests most likely to be proposed to you in a pitch deck are the ones the documentation does not support.

The first row is the expensive mistake, and it is our inference rather than Google's statement. If both arms already qualify for contextual matching, the experiment cannot tell you anything about it. AI Max may still deserve a test on its own merits. Just do not write "AI Mode result" at the top of that slide.

If an agency or tool sells you an AI Mode ROAS number today, ask which Google field it comes from. We could not find one in Google's documentation, and a figure with no field behind it is a model output that should be labelled as one.

Where this argument is weakest

Documentation is not behaviour, and our audit reads only the first. Google's help page carries no revision date, so we cannot tell you when any sentence changed. A control could ship next week, or exist today as a limited beta in accounts we cannot see.

The exact and phrase rule is also less settled than the help page makes it sound. Marvin called it "a small experiment that recently kicked off" in September, according to Search Engine Roundtable as cited by PPC Land. The current help page states the eligibility without that qualifier. Your account may not match the documented rule.

The strongest case against this post is that a separate AI Mode report may not matter. Google itself says monetization there has been lower, and Google has not published the share of ads or spend that AI Mode represents. If AI Mode is a small part of your impressions, campaign-level incrementality is the right question anyway. A dedicated segment would then be a tidy dashboard rather than a business input. We think the share will rise as the Q4 markets arrive, but we cannot show that from anything published.

Our 0 of 9 figure also depends on our list. A different reader could add rows or split them differently. We expect the zero to survive that exercise, because Google's FAQ answers the central question directly. The denominator, though, is ours.

Frequently asked questions

Where to start before the rollout reaches you

Pull one number today: the share of your Search spend in campaigns using broad match, AI Max, Shopping or Performance Max. That is your exposure to contextual matching in AI surfaces, and it is the only exposure figure your account can give you.

Then ask your Google representative one question in writing: is Conversion Lift by geography available on this account, for this campaign, this quarter? If yes, book the study before the next rollout lands. If no, set up the match type experiment on your largest Search campaign, and write the words "campaign incrementality, not AI Mode" at the top of the results.

Related reading