How AI Search Summaries Changed Paid Search, Date by Date

Before moving budget because “AI summaries” are supposedly killing paid search, separate the changes by date. The paid-search strategy problem with AI-generated search summaries is not one clean before-and-after event. Google changed the visible SERP, then inserted ads into AI-generated answers, then pushed ads toward AI Mode conversations, and now has a forward-looking campaign-structure migration tied to AI Max.

Those changes do not show up in the account as one neat column. That is the annoying part. A Search campaign can keep serving, CTR can move, CPC can drift, budget can stop clearing the way it did last quarter, and the account still will not give you a clean “AI Overview ad” segment to isolate the placement.

The dated paid-search tracker

AI Overviews changed the page before they changed the ad placement

The first paid-search issue was not that Google immediately put a new paid unit inside every AI-generated answer. It was that the page changed. AI Overviews created a higher, larger answer block that could satisfy part of the user’s task before the user clicked anything. For paid search, that matters because the old account reflex — impressions up, CTR down, check ad copy and rank — becomes incomplete. A lower CTR can come from a changed page, not just a worse ad.

That does not mean every AI Overview query is bad for paid search. Commercial intent still matters. A user comparing vendors, prices, software categories, or local options may still click ads. A user asking for a simple factual explanation may not. The useful account work is to split likely AI Overview-triggering queries by intent and device, then compare movement against the dates above.

The AI Overview ads entry is the one that deserves the most scrutiny

Ads in AI Overviews are not just another cosmetic SERP change. Google Ads Help says ads may appear with AI Overviews in eligible countries, lists 12 countries where the feature is available, says sensitive categories are excluded, and confirms two operational facts that matter more than most launch language: advertisers cannot opt out of ads in AI Overviews, and Google Ads does not provide segmented reporting for ads shown in AI Overviews. [1]

That combination is the account-management problem. Eligibility is mandatory once your campaign and market qualify. The placement is blended into existing reporting. The result is that a media buyer can be funding AI Overview inventory without being able to pull a clean placement report that says how those impressions, clicks, costs, and conversions performed against the rest of Search.

The sensitive-category exclusion is worth noting, but it does not solve the reporting problem for everyone else. If a finance lead asks why nonbrand CPC rose after a page-layout change, “Google says the ads are helpful” is not a report. A usable answer needs splits: AI Overview vs. standard Search, mobile vs. desktop, country, query class, campaign type, conversion action, and new vs. returning user. The platform does not provide the most important split.

That is also why the October 2024 and December 2025 rollout dates matter. A US mobile account could have seen effects earlier than a multinational account whose relevant country entered the availability set later. Treating “AI Overview ads” as one global 2024 event will muddy the read.

The click-impact numbers are real, but they are not interchangeable

The strongest public evidence around AI summaries and click behavior comes from studies that measured different things. That is not a reason to ignore them. It is a reason to stop putting all of them into one headline.

Pew is especially useful because it measures user behavior instead of vendor optimism. But it still measures broad Google search behavior, not the paid auction inside a specific advertiser’s account. Ahrefs gives a sharper CTR-loss number, but it is about position-1 organic CTR under its methodology. Those are different surfaces.

For paid search, the practical question is narrower: did an AI-summary-triggering query group lose clicks, face higher CPCs, or convert differently after the relevant rollout date? If fewer users click anything on a page, and ad inventory remains competitive, CPC pressure is plausible. But the responsible way to handle that is to compare the account’s own CPC, impression share, top-of-page rate, budget-limited status, conversion rate, and query mix before and after the relevant change. If you need the auction-mechanics version of that discussion, see how AI is driving digital ad costs.

AI Mode moves the targeting question from queries to conversations

AI Mode is a different reporting headache from AI Overviews. AI Overviews sit on a recognizable search results page. AI Mode is closer to a chat-style search session. In August 2025 agency briefings, Search Engine Land reported that AI Mode ads were being described around the context of the conversation rather than only the typed keyword; one quoted line captured the operating shift as “from queries to conversations.” [2]

That sounds reasonable as a matching concept. A multi-turn search session can contain more intent than a two-word keyword. The problem is not the idea. The problem is what the advertiser can inspect afterward. If the ad was matched to conversation context, the old search-term report is no longer a full explanation of the impression.

Keep this measured in account reviews. AI Mode ads matter because they point to where Google wants Search monetization to go, not because every Search account in Q3 2026 has already been rebuilt around AI Mode. For current management, keep AI Mode as a separate line in the change log: announced, briefed, tested, or live for the relevant inventory. Do not treat it as the dominant surface unless your own account data and Google’s visible controls support that.

Google Marketing Live 2026 widened the AI-search ad surface

Google Marketing Live on May 20, 2026 added more pieces to the same pattern. Google described a new generation of ads for the AI era of Search, including Ads in AI Mode and new AI-assisted ad experiences across Search and commerce. Its business announcements log also dates “Ads in AI Mode” to May 20, 2026. [3][4]

The formats should not be flattened. Conversational Discovery ads, Highlighted Answers, AI-powered Shopping ads, Business Agent for Leads, and expanded Direct Offers do different jobs. A shopping ad inside an AI-assisted commerce experience is not the same operating problem as a lead agent handling a prospect. The shared issue is that more of the ad experience moves into Google-controlled assistance layers, where the advertiser needs to know which campaign became eligible, what user action counted, and which report proves it.

For a small agency or in-house growth lead, the immediate action is not to rewrite the whole media plan around a keynote. It is to add each format to the account’s change file with three labels: announced, testing, or live in this account. If the label is not live, it should not explain last month’s CPC. If the label is live but not segmented, it belongs in the caveats before anyone presents a clean cause-and-effect story.

The September 2026 DSA-to-AI-Max item is related, but not the same as AI Overview ads

The September 2026 Dynamic Search Ads to AI Max upgrade window belongs in the same tracker because it extends the same automation pattern into campaign structure. It does not prove that AI Overviews caused a paid-search performance change. It means another layer of keywordless or semi-keywordless matching may enter accounts close to the same period when AI-search surfaces are also changing. Google’s announcements log lists the September 2026 upgrade as a forward-looking item, so it should be re-verified before anyone makes structural changes from it. [4]

If you still have legacy DSA campaigns, inventory them now, but do not treat the migration as already completed unless your account says so. Compare current DSA query coverage, landing-page matching, negative controls, budget allocation, and conversion quality against the AI Max setup you expect to inherit. For a deeper structure pass, use the Google AI Max for Search and Performance Max guide before the upgrade window becomes an emergency.

What to check before reallocating budget

• Mark likely AI Overview-triggering query groups. Start with informational and comparison queries, then separate brand, nonbrand, shopping, local, and lead-gen intent.
• Compare performance around the relevant date, not around “AI search” as a vague period. Use May 2024 for the page change, October 2024 for US mobile AI Overview ads, December 2025 for broader country availability, May 2026 for AI Mode ad announcements, and September 2026 only after re-verifying the DSA-to-AI-Max migration.
• Segment what Google still lets you segment: device, country, campaign type, network settings, brand vs. nonbrand, match type where applicable, search terms, conversion action, and new vs. returning users.
• Write down what cannot be segmented. Ads in AI Overviews have no opt-out and no separate Google Ads reporting split, so any claim about their account-level performance is inferred unless Google adds visibility.
• Watch Search, Shopping, Performance Max, and AI Max eligibility separately. A movement in PMax or AI Max should not automatically be blamed on the same surface as a Search text ad inside an AI Overview.
• Use external CTR studies as investigation prompts, not budget rules. Pew, Ahrefs, and your paid account data measure different things.
• Re-check the September 2026 DSA migration in Google’s own announcements and in the account before changing structure. Treat it as forward-looking until the account confirms it.

The safer budget call is slower and more boring than the headline. Do not reallocate because “AI summaries” happened. Re-check each dated platform change against what is live, measurable, and visible in the advertiser’s own account.

References

1. About ads and AI Overviews, Google Ads Help.
2. Google briefs brands on AI Mode ads ahead of Q4 rollout, Search Engine Land.
3. A new generation of ads for the AI era of Search, Google, May 20, 2026.
4. Announcements, Google Business.
5. Google users are less likely to click on links when an AI summary appears in the results, Pew Research Center, July 22, 2025.
6. AI Overviews Reduce Clicks by 34.5%, Ahrefs, April 2025.
7. AI Overviews Reduce Clicks Update, Ahrefs.