Singapore

Google Ads in AI Mode: 10+ Countries and a Reporting Gap

As of November 2026, your existing Search, Shopping and Performance Max campaigns can serve ads inside AI Mode in 10+ countries, in English, with no new campaign or separate budget needed, and since 9 November those placements also reach follow-up questions in a US conversation. What you cannot do is see them as a line of their own: Google reports AI Mode traffic only inside campaign totals. That single fact decides what an sem agency should brief a client on before the question "are we in AI Mode?" arrives.

What changed on 9 November 2026?

Google announced the expansion from its official Google Ads account on November 9, adding more than 10 countries where ads can appear in AI Mode and switching on multiturn ads in the U.S. for follow-up questions inside a conversation. An earlier announcement had already named the country list; the November post made the follow-up placement real.

Two details in that list matter more than the headline number. Pakistan shows text ads only, and the language for placements within AI Mode is English. If your Singapore account runs English search, you are in scope by default, because eligibility does not require an opt-in. Ads above or below an AI Overview are a different placement and reach 200+ countries and territories in all languages where AI Overviews exist, so a lot of what a marketer calls "AI search" today is that wider, long-standing surface rather than AI Mode itself.

Marketer pinning markers across a wall map of the world in a Singapore office while a colleague takes notes.

Which campaigns are in, and which keywords are not?

The documentation is unusually specific about matching, and this is where an account loses or gains coverage.

Campaign or targeting type	Matched to the direct query	Matched contextually to query plus AI response
Search with exact or phrase match	Yes, when the query shows direct commercial intent	No
AI Max for Search, broad match	Yes	Yes
Shopping campaigns	Yes	Yes
Performance Max	Yes	Yes

Read the first row carefully. Exact and phrase match keywords are not eligible for contextual matching on broader exploratory questions. The explore-type question, where a user describes a problem and the AI answer surfaces a product category, is reachable through broad match, AI Max, Shopping or Performance Max. An account built entirely on tight exact-match lists has a narrower footprint in AI Mode than the campaign manager assumes, and no report will tell them so.

Campaign guardrails still apply, and any sem agency running a tight account should check them before the next review: negative keywords and asset pinning carry over, every ad in these placements carries the Sponsored label, and advertisers cannot target AI Mode only or opt out of it separately from Search. Billing runs on the same auction and the same CPC or conversion-based bidding, with no extra charge, and standard invalid traffic protection applies.

Why can't I see AI Mode performance separately?

Because Google includes those queries in aggregate. Campaign reporting carries AI Mode impressions, clicks, conversions and cost inside the existing metrics, and the search terms report folds AI Mode queries into the total rather than listing them apart. There is no AI Mode filter, so an AI Mode CPA or ROAS is not calculable from the Google Ads interface.

For a reporting lead the consequences are concrete. If a campaign's cost per conversion improves in the weeks after the rollout, the improvement cannot be attributed to AI Mode on timing alone; the placement is mixed with everything else in the same row. The reverse also holds: a placement that underperforms hides inside a healthy campaign average. Two datapoints explain why this gap is harder to close than it looks:

Each platform enforces its own attribution window, and the platform applies a longer one than most B2B deal cycles need. A deal that closes after that window is rejected from platform reporting, so a platform report is a partial count even before placements are merged.
Everything between the click and the signed deal, from on-site interaction to lead progression and CRM follow-up, sits outside the ad account and is never in the platform's report.

Put together, the aggregate view is not something a better dashboard fixes. It is the boundary of what first-party platform data can answer.

Analyst at a desk studying campaign charts on a monitor, with the lower part of the screen in shadow.

What an agency should report while Google reports in aggregate

The sensible answer is to stop trying to isolate AI Mode and start measuring the funnel the placement feeds. The metric and its time range go into the brief before the first report goes out, so the client reads one number with one meaning.

A workable reporting set for a Singapore SME with lead generation in mind:

Agree the conversion definition and its window with the client in writing, for example a qualified enquiry booked inside the attribution window above.
Report campaign-level cost per qualified lead next to the platform's own conversions, and state which of the two decides budget.
Feed qualified leads back to Google and Meta as offline conversions, so bidding optimises toward revenue rather than form fills.
Reconcile the platform totals against the client's CRM once a month, and keep the differences visible instead of smoothing them.

Step 2 is the one that survives the AI Mode problem. If the reported number is a qualified lead the client's sales team accepted, whether the click came from a classic results page, an AI Overview or an AI Mode conversation stops being a reporting blocker and becomes a targeting question. Step 4 is where a platform report stops being treated as ground truth.

A small calculation that shows the blind spot

A lead-generation account runs two campaign lines, one for brand and one for generic service terms. Look at two months of the generic line. In the month before the rollout its cost per conversion sits at one level; the month after, at a lower level. The client asks whether AI Mode is working.

The honest response is that both figures are blended over every placement that line served, AI Mode among them, and that the movement between them could come from auction mix, seasonality, a landing page change or AI Mode, in any combination. What the account can still test is the part it controls: move the exploratory terms to broad match or AI Max, give the campaign accurate conversion values, and compare cost per qualified lead from the CRM before and after. The placement remains invisible; the business outcome does not.

An sem agency that keeps the placement question out of the monthly report and the funnel question in it is doing the work this situation actually asks for. That is the habit that matters here: separate what changed from what caused the change, and only claim the second when the measurement supports it. A number that moves is not evidence until the definition behind it is pinned down.

What the country-by-country rollout means for a Singapore account

Singapore is on the documentation's country list for ads within AI Mode, and English is the only language there, which suits most local search accounts that already run English copy. Two practical consequences:

No budget change is needed to become eligible; the campaigns participate through the existing auction when relevant.
No negative-keyword list can exclude AI Mode alone, so brand exclusion and negative keywords remain the only levers over when and how ads serve.

For a team that has to explain a move to clients, a little context from the sem agency singapore side helps: the country list is also the answer to "where did this come from": ads within AI Mode are a named placement in the documentation, but their results still land in the campaign row the client already reads.

Multiturn ads are a US launch for now, and Google has not said when follow-up placements travel. For a Singapore advertiser that means the conversational placements inside a long AI Mode session are not yet a planning assumption, while the expanded country coverage is. Building a report that cannot separate the two is how a team ends up with a number nobody can explain.

Person at a coffee shop typing a follow-up question on a phone as a colleague compares a printed report with a laptop.

Topkee's point of view

The product knowledge behind this article holds no statement about partnership with Google's AI Mode placements, and none should be read into it. What it does describe is a delivery structure built around the parts of the funnel a platform report cannot see.

Topkee's Performance Google Ads work is scoped across the full funnel, from impression to click, landing page, enquiry, proposal and close, with CAPI and offline conversion setup and attribution reconciliation treated as core deliverables rather than add-ons. Each lead carries a ttoclid, and sales-stage events come back to Google Ads Conversions, Bing UET, the Facebook pixel and GA4 through TTO Tag. When generative placements like AI Mode hand more of the auction and the assembly to the platform, the agency's value shifts from operating the console to supplying assets and signals, which is the end of the work those systems cover. The inference is ours, not an official claim: a cleaner conversion signal and a reconciled funnel are what any expanded placement needs in order to be judged on outcomes.

What to do next

Before the next client review, write the conversion definition, the attribution window and the CRM reconciliation date into the reporting brief, then check whether your exact and phrase match lists are doing all the work in accounts that now rely on contextual matching. If you want a starting point, the page on how brands use Google Ads with first-party data sets out the same argument from the data side, and the note on AI-powered Performance Max conversions covers the conversion signal that broad match and AI Max lean on.

For an advertiser choosing a partner for this work, the questions to ask a sem agency singapore team are simple: which conversion counts, over which window, and what happens when the platform total and the CRM disagree. A team that answers those three without hesitation is the one whose AI Mode reporting you can trust, whatever the placement mix turns out to be. The same questions belong in the brief a sem agency hands back before any campaign manager starts optimising toward a number nobody defined.

FAQ

Are my existing Google Ads campaigns already showing in AI Mode?

If they are Search, Shopping or Performance Max campaigns serving in English in one of the listed countries, they are eligible by default. You do not create a new campaign or a separate budget. Whether a given query shows an ad depends on the auction and on whether the query has commercial relevance.

Can I see how much spend went to AI Mode?

Not as a separate figure. AI Mode impressions, clicks, conversions and cost are included in aggregate in campaign reports, and AI Mode queries are folded into the search terms report total. There is no filter that isolates the placement.

Why do my exact match keywords not show in AI Mode conversations?

Exact and phrase match keywords are eligible for direct query matching when the query shows clear commercial intent, but not for contextual matching on broader exploratory questions. Broad match, AI Max, Shopping and Performance Max are the targeting types eligible for that contextual matching.

Is it a problem that AI explainers are written by Google?

AI explainers are generated by Google's Gemini models for some AI surface placements, and advertisers cannot write, edit, approve or opt out of them. The ad copy you control is still the campaign asset; the explainer is context Google generates. Keep website content clear and current, since public sources including your site can inform what the explainer says.