The industry has been bracing for a layout problem: one AI panel merging sponsored ads, organic citations and product cards. The real problem turned out to be different. An AI Overview confident enough to directly contradict the paid ad sitting immediately above it, on the same page, in real time. That is a trust and attribution problem, not a layout one.
Take a classic transactional query: "What is the best plumber for a broken pipe?" At the top sat a textbook high-converting sponsored ad from Eco Plumbers — an $89 leak detection promo, 24/7 availability, 4.9 stars across 20,000+ reviews, a video carousel and location extensions. Every conversion signal an advertiser can pay to deploy.
Directly below, the AI Overview said, with zero hedging: "The best plumber for a broken pipe is a local 24/7 emergency plumbing company like Roto-Rooter Plumbing & Water Cleanup or Amanda Plumbing." The advertiser paying premium CPCs for the top position was not mentioned at all.
For two decades organic search was a candidate pool. Searchers scanned ten blue links, weighed meta descriptions and brand familiarity, and calculated their own trust. Ads sat above or beside them with clear "Sponsored" labels, and users treated both as raw material for a decision they still had to make.
AI Overviews eliminate that evaluation phase through phrasing. They do not offer candidates; they issue conclusions — not "here are options to consider" but "the best plumber for a broken pipe is…" delivered with identical authority whether the underlying data is exhaustive or thin. The Overview is not just competing with the ad for a click; it replaces the user's internal decision process with an algorithmic verdict the ad team had no say in.
The same behavior shows up in ecommerce. For "sweatshirts for anxiety," the standard SERP showed a Shopping carousel of six products from brands like Cloud Nine and Comfrt ($39–$89.95), followed by a Cloud Nine text ad. Cloud Nine dominated the paid real estate across two formats.
Prompted directly on which sweatshirts are best for anxiety, the AI Overview recommended weighted and sensory hoodies — Comfrt ($75), Thera ($158) and Cozy Ghost ($118). Comfrt earned both a paid slot and an organic citation; Thera and Cozy Ghost appeared with no paid presence at all. Cloud Nine, spending aggressively for top-of-page placement, was omitted entirely.
Shift to the informational variant "what is an anxiety sweatshirt," and the model's logic flips: Cloud Nine appears alongside Etsy sellers, reclassified under "graphic and affirmation apparel" rather than "weighted hoodies."
The operational reality is that the auction engine and the retrieval-augmented generation (RAG) engine run on entirely separate logic. Neither the PPC team nor the SEO team controls how they interact, and the logic can flip on subtle prompt variations.
The brands cited had classic fundamentals in place.
SEO has not vanished. The goal moved from earning a position on the page to earning a citation in the answer.
Here is the downstream sequence when an Overview interrupts the funnel:
The Quality Score algorithm does not adjust for an AI Overview undermining your ad's credibility on the same SERP. It registers lower expected CTR, penalizes the keyword and raises your cost — so you pay a premium for diminished traffic on queries where an AI module assigned brand trust to a competitor first.
For a parallel case where AI text lands inside the ad unit itself, see Naver Tests AI Summaries in Power Link Ads; for the blurring line between answer and advertising, see Ads Arrive Inside Conversational AI.
Users convert with the AI-cited competitor without clicking your ad, so you accumulate impressions without clicks. Falling expected CTR lowers Quality Score and raises effective CPC.
The ad auction and the retrieval-augmented generation engine run on separate logic. A small shift from transactional to informational phrasing reclassified brands entirely in testing.
Entity clarity with structured data and consistent NAP details, granular product specs like weight and materials in schema and on-page copy, and an indexed multi-platform footprint.
To apply what you just read to your own site, start with a free audit of where things are now.
A strategist replies within 24 hours on business days.
A new Enhanced matching option is appearing in Customer Match settings, extending first-party audience reach by matching consented users with consented publisher users.
Google is adding multi-campaign A/B testing, experiments that keep brand and location controls on, and expanded Performance Planner forecasting to AI Max for Search campaigns.
Demand Gen starts where there is no keyword and no existing interest. As that protection disappears, PPC specialists need creative strategy, audience research and full-journey measurement.