There is a new failure mode in B2B SaaS paid search, and it is not the one most teams are watching. Everyone knows AI Overviews push ads down the page and dilute click-through rate. The subtler and more expensive problem is contradiction: your ad wins the top paid slot, the searcher sees it, and directly beneath it an AI Overview answers their question by naming two of your competitors as the authoritative solution. You paid for the placement, and Google's own AI summary is undercutting it in the same screen. As Search Engine Land frames the scenario, your ad may win the auction while "an AI Overview appears below your ad, naming two competitors as the authoritative solution," per its analysis of what happens when AI Overviews contradict paid search ads.
This is distinct from the CTR story we cover in our post on the impact of AI Overviews on SaaS paid CTR, and distinct again from AI summaries rendered inside the ad unit. Contradiction is a message-control problem, not a volume problem. This post is a tactical guide: how to find the queries where the AI Overview is working against your ad, and how to move spend and messaging toward the queries where paid search still owns the answer.
What the contradiction effect actually is
The contradiction effect happens when two things co-occur on one search results page: your paid ad in the top slot, and an AI Overview below it that resolves the searcher's question without sending them to you — often by recommending named competitors. The searcher gets a synthesized answer for free, and if that answer endorses a rival, your paid impression has been neutralized or worse. You are still charged for the click when it happens, but the context around the click now includes an AI-generated second opinion that you did not write and cannot edit.
It matters because it inverts the usual logic of high-intent search. The whole premise of paid search for B2B SaaS is that you pay to be in front of a buyer at the moment of intent. When an AI Overview interjects a competitor recommendation into that moment, the intent is still there but it is being redirected. This is not the same as an ad simply being pushed below the fold. A pushed-down ad loses attention; a contradicted ad loses the argument. The distinction changes what you do about it: you cannot fix contradiction by bidding higher, because a higher position still sits above an AI Overview that names someone else.
How common this is now on B2B keywords
AI Overviews are no longer an edge case on B2B queries — they are the default. A 2026 study by Indexed found that roughly 84% of B2B keywords now trigger an AI Overview, and that coverage in the B2B Technology category jumped from 36% to 82% year over year, per its 2026 CTR study. When an AI Overview is present, the same research measured paid click-through rate falling from about 21% to about 10%. So contradiction is not a rare event you can ignore; it is a condition that applies to the large majority of your non-brand keywords.
For planning purposes, assume an AI Overview is present on most of your research-stage and category terms, and treat its absence as the exception to verify rather than the rule to expect. That single assumption reframes budget allocation: if the majority of your non-brand impressions are served alongside an AI answer, then the value of those impressions depends heavily on whether the AI answer helps or hurts you. The rising cost of clicks makes this sharper — with B2B SaaS click costs climbing, as we detail in why B2B SaaS CAC is rising, paying premium prices for impressions that a competitor-endorsing AI Overview undercuts is a compounding waste.
Which queries are most exposed
The exposure maps almost perfectly onto funnel stage. Informational and research-stage queries are the most contradicted, because AI Overviews exist precisely to answer those questions inline. A search like "best CRM for startups" or "what is a customer data platform" invites the AI to produce a ranked list of vendors, and your ad sits above a summary that may recommend three of them by name. These are the queries where the contradiction effect is most likely and most damaging, because the searcher has not yet decided and the AI is nudging that decision.
Bottom-of-funnel and branded queries are far more resilient. When someone searches "acme analytics pricing" or "acme analytics demo," they need to complete an action the AI cannot perform for them, so the paid ad retains its job. This is the same logic behind organizing accounts by intent in the first place, which we cover in B2B SaaS campaign structure by funnel intent tiers. The practical rule: the higher up the funnel a query sits, the more likely the AI Overview is to answer it with someone else's name, and the more carefully you should scrutinize the spend behind it.
A detection workflow you can run this week
Detection is manual but fast. Pull your top 20 to 30 non-brand keywords by spend, then check each one in a clean, logged-out or incognito browser session to avoid personalized results. For every query, record three fields: does an AI Overview appear, does it name specific vendors, and are any of those vendors your direct competitors. That third column is the one that matters — an AI Overview that summarizes a concept without endorsing a product is tolerable; one that ranks your rivals is not. Repeat the audit monthly, because AI Overview coverage is expanding quickly and a query that was clean last quarter may not be now.
Rank-tracking tools increasingly flag whether an AI Overview is present, but most still do not tell you which brands the summary cites, so the manual read of your highest-spend queries remains necessary. Concentrate the effort where money is at stake: the queries with the most spend and a named competitor are your priority list. This mirrors the discipline in our Google Ads optimization checklist — audit by dollars at risk, not by keyword count. Document the contested queries in a simple sheet so you can track whether your countermeasures move them over time.
How to route spend around the conflict
The response is reallocation, not wholesale cutting. First, separate the two cases you found in detection: queries where the AI Overview merely summarizes a topic, and queries where it explicitly recommends a competitor. Keep investing in the first if the clicks are qualified. For the second, shift budget down-funnel toward branded, comparison, and transactional terms the AI cannot resolve for the user — the queries where the searcher must click to act. This concentrates spend where paid search still owns the answer and pulls it away from impressions a competitor-endorsing AI Overview is neutralizing.
Second, strengthen what you can control on the contested queries you choose to keep. A sharper ad and a landing page that converts harder make the click you win more valuable, which partly offsets the AI Overview's drag. And feed your bidding the right signal so it does not chase cheap clicks on contradicted queries: optimizing toward qualified pipeline rather than raw form fills, as argued in optimizing for SQLs, not leads, naturally steers Smart Bidding away from the up-funnel terms where contradiction is worst. Treat a rising AI Overview presence on a query as a signal that it has drifted up-funnel, and let that reclassification drive where your money goes. The broader shift toward AI-mediated search — including how ads work in Google AI Mode — points the same direction: own the intent the machine cannot resolve, and stop overpaying for the intent it resolves for someone else.