Ads already serve inside Google's AI Overviews, and ads inside AI Mode are in live testing in the US. Text and Shopping ads from existing Search, Shopping and Performance Max campaigns are eligible to appear within an AI Overview, advertisers cannot opt out of the placement or target it directly, and Google Ads does not yet break it out in reporting (Google Ads Help, "About ads and AI Overviews", checked 20 July 2026).
That is the whole confirmed picture on the paid side. Everything beyond it, including how the new AI Mode formats will be priced and reported, is either in test or unannounced. This post separates the two, because on a topic moving this fast the useful thing is not a prediction, it is a clean line between fact and guess.
What is confirmed about ads in AI Overviews
Google's own advertiser documentation is the primary source here. Checked 20 July 2026 at support.google.com/google-ads/answer/16297775.
| Question | Confirmed answer | Source |
|---|---|---|
| Which ads can serve inside an AI Overview | Text and Shopping ads from existing Search, Shopping and Performance Max campaigns | Google Ads Help, checked 20 Jul 2026 |
| Where else can ads serve | Above or below the AI Overview in all 200+ markets where AI Overviews are available | Google Ads Help, checked 20 Jul 2026 |
| Which countries for ads within AI Overviews | English, mobile and desktop, in Australia, Canada, India, Indonesia, Kenya, Malaysia, New Zealand, Nigeria, Pakistan, Philippines, Singapore and the US | Google Ads Help, checked 20 Jul 2026 |
| Can you opt out | No. "You can't opt out of serving ads in AI Overviews" | Google Ads Help, checked 20 Jul 2026 |
| Can you target the placement | No. "You can't directly target ad placements in the AI Overviews" | Google Ads Help, checked 20 Jul 2026 |
| How is it reported | As Top Ads. Google states it "doesn't offer segmented reporting when ads show within Search AI Overviews" | Google Ads Help, checked 20 Jul 2026 |
| How is relevance decided | Existing auction ranking and signals, matched against query intent and the content of the AI Overview | Google Ads Help, checked 20 Jul 2026 |
Two lines in that table matter more than the rest. You cannot opt out, so the placement is already happening in accounts that never chose it. And you cannot segment the reporting, so nobody can currently tell you what their AI Overview ad performance is in isolation. Any agency quoting you a clean AI Overviews cost per acquisition today is inferring it, not reading it.
What was announced for AI Mode
At Google Marketing Live on 20 May 2026, Google announced a set of Gemini-generated ad formats built to sit inside AI Mode responses rather than beside them (blog.google, "A new generation of ads for the AI era of Search", 20 May 2026; Search Engine Land, 20 May 2026, both checked 20 July 2026).
| Format | Status as announced | What it does |
|---|---|---|
| Conversational Discovery ads | In test, US, mobile and desktop | Creative generated with Gemini that answers the specific question asked |
| Highlighted Answers | In test, US, mobile and desktop | Sponsored suggestions placed within AI Mode recommendation lists |
| AI-powered Shopping ads | Rolling out later in 2026 | Gemini-written explainers on why a product fits the query |
| Business Agent for Leads | Open beta for US advertisers | A brand chat agent inside the ad unit |
| Direct Offers | Expanded pilot | Promotion bundling and native checkout for participating merchants |
If the distinction between the two surfaces is not settled in your head, AI Mode versus AI Overviews covers it, and what Google AI Mode actually is covers the surface itself.
Confirmed against unclear
| Claim | Status | Basis |
|---|---|---|
| Ads serve within AI Overviews from Search, Shopping and PMax | Confirmed | Google Ads Help, checked 20 Jul 2026 |
| No opt out, no direct targeting of the placement | Confirmed | Google Ads Help, checked 20 Jul 2026 |
| Reported as Top Ads, no segmented reporting | Confirmed | Google Ads Help, checked 20 Jul 2026 |
| Four new Gemini formats announced for AI Mode | Confirmed | blog.google, 20 May 2026 |
| AI Mode ad testing is US and English only | Confirmed as of announcement | Search Engine Land, 20 May 2026 |
| Whether AI Mode ads will get their own reporting segment | Unclear | No Google announcement found as of 20 Jul 2026 |
| Whether AI Mode ads price differently from standard Search auctions | Unclear | Not documented |
| Non-US rollout dates for AI Mode ads | Unclear | No published timeline |
| What share of AI Mode responses carry an ad | Unclear | Third-party estimates circulate, none from Google |
| Whether ad presence affects which sources the answer cites | Unclear | Google describes ads as serving through the existing auction, separate from organic |
That last row is the one worth arguing about, and it leads directly to the strategic point.
The ad does not buy the citation
Want to see this in action?
See how every major AI model talks about your brand. Free to start.
An AI Overview has two kinds of brand presence. There is the sponsored unit, which you buy in an auction. And there is the set of sources the answer draws on and links to, which you cannot buy at any price. They sit centimetres apart on the same screen and they are won in completely different ways.
This matters commercially because the cited source often carries the recommendation while the ad carries only the offer. When a buyer asks which product to choose, the answer text is doing the persuading. The ad is arriving after the opinion has formed.
It gets harder still. Semrush found that 62% of AI citations never name the brand being cited, so a page can be feeding the answer without the reader ever registering who wrote it. Ahrefs found AI visibility correlates most strongly with third-party mentions and video rather than on-page work, which means the lever for getting cited sits largely outside your own website and entirely outside your ad account.
So the honest sequencing is this. Find out whether AI answers in your category already name you. If they do, the ad is topping up a position you hold, and it is probably worth paying for. If they do not, and a competitor is being named in the answer text above your ad, you are buying attention inside a page that is recommending someone else. That is a much worse purchase, and no bid adjustment fixes it. An AI visibility audit answers that question before the budget question.
What to measure if you are running these ads
Segmented reporting does not exist yet, so measurement has to be built from proxies and from the organic side.
| What to track | How | Why it matters |
|---|---|---|
| Top Ads impression share shift | Google Ads, by query theme, month over month | AI Overview placements land in Top Ads, so movement here is the closest available signal |
| CTR change on AI Overview-triggering queries | Compare query sets that do and do not trigger an overview | Isolates the surface effect without segmented reporting |
| Branded search volume after campaign start | Search Console and Ads | Assisted demand is often where these placements pay back |
| Citation share for the same buyer questions | AI visibility tracking across engines | Tells you whether you are the recommendation or just the advert |
| Answer volatility on your key prompts | Repeat the same prompt on a schedule | Stops you reading noise as a trend |
That last row is not optional. SparkToro found the same AI query changes its answer roughly 70% of the time. Honeyb's own measurement on 13 July 2026, running 20 buyer prompts three times each on four engines by API for 240 answers, found the top-ranked brand changed between identical runs 44% of the time on Gemini, 43% on Perplexity, 35% on ChatGPT and 28% on Claude. A single before-and-after check on one engine will tell you almost nothing about whether your campaign moved anything.
Top-pick change rate
How often the top recommendation changes between identical runs
For the organic half of the job, how to appear in AI Overviews covers the work that actually earns citations, and an AI Overviews tracker covers how to watch the surface over time. If you are budgeting across assistants rather than just Google, ads in ChatGPT is the parallel question.
The practical position for the next two quarters
Ads in AI Overviews are not a decision you get to make, they are a condition of running Search, Shopping or Performance Max in the eligible markets. AI Mode ads are a US test with real formats and no published economics. Neither of those facts changes what your brand is worth inside an AI answer.
Treat the paid surface as distribution and the citation surface as reputation. Buy the first, earn the second, and never let a rising impression share convince you that AI assistants are recommending you. Check where you actually stand first with the free AI visibility check, then decide what the ad is worth.
