---
title: "Google AI Overviews Aren’t Ad Slots: Six Beginner Beliefs That Waste Budget"
description: "A PPC operator breaks down six paid-search assumptions that fail in Google AI Overviews, what each one costs, and where to start instead."
url: "https://groas.com/post/what-business-owners-get-wrong-when-they"
image: "https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/9f0e1b35-adf6-4a10-a148-d2bb6ff87bab.png"
published: "2026-10-11T05:36:47.636Z"
modified: "2026-10-11T05:36:47.705Z"
---

October 11, 2026 · 9 min read

# Google AI Overviews Aren’t Ad Slots: Six Beginner Beliefs That Waste Budget

[Alexander PerelmanHead Of Product @ groas](https://groas.com/author/alexander-perelman)[LinkedIn](https://www.linkedin.com/in/alexander-433793253/)

![Cartoon of a roofer offering cash at a library answers desk while the librarian ignores him, clipping the useful paragraph that actually earns the spot.](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/9f0e1b35-adf6-4a10-a148-d2bb6ff87bab.png)

In this article

1. [“I just need to target the right keyword”](#i-just-need-to-target-the-right-keyword)
2. [“One comprehensive article will capture the category”](#one-comprehensive-article-will-capture-the-category)
3. [“Schema markup is the technical shortcut”](#schema-markup-is-the-technical-shortcut)
4. [“If I rank on page three, I can skip ahead”](#if-i-rank-on-page-three-i-can-skip-ahead)
5. [“It’s like buying an ad: pay once and you’re set”](#its-like-buying-an-ad-pay-once-and-youre-set)
6. [“I’ll know it worked when organic traffic jumps”](#ill-know-it-worked-when-organic-traffic-jumps)
7. [The order I’d actually start in](#the-order-id-actually-start-in)
8. [“If our answer is right, Google will eventually cite us”](#if-our-answer-is-right-google-will-eventually-cite-us)

Someone asks which keyword will get their business into a Google AI Overview. Someone else asks what an agency charges to put them there by next month. I know the mental model behind both questions: they are treating an AI citation like an ad placement.

In paid search, that model has a logic. You enter an auction, target a query, set a bid, pass editorial review, and pay to show a message. An AI Overview is not another slot in that auction. Google retrieves passages relevant to parts of a user’s question, then synthesizes an answer from them. You cannot bid on the summary box or buy your way into it with an _AI-optimized_ article.

That distinction sounds simple until it hits a budget. Teams commission generic posts, add schema, and watch a handful of prompts for the citation they expected to appear within weeks. Each move borrows a habit from paid search. Each fails for a reason a PPC person should recognize. These are the six beliefs I would drop first.

## “I just need to target the right keyword”

This is the most common starting point. In paid search, you can target a term like _commercial flat roof repair_, write matched ad copy, and measure what that traffic does. It is tempting to put the same phrase in a page title, headers, and body copy, then wait for an AI Overview to pick it up.

But a buyer may ask whether a cold-applied membrane holds up better than heat-welded TPO on an uninsulated warehouse roof. That is not one tidy keyword. As [Google Search Central’s guidance on generative AI search](https://developers.google.com/search/docs/fundamentals/generative-ai-search) explains, AI search builds on Google’s existing search systems rather than requiring a special keyword tag. The system can break a complex question into narrower searches and retrieve passages that address its parts. To understand [why AI assistants cite some businesses and skip others](https://groas.com/post/how-ai-assistants-decide-which-business), look at whether a passage answers a useful question, not how often it repeats a target phrase.

**The cost is paying for pages built around the wrong unit of work.** A high-volume head term may tell you what market a buyer is in, but not which technical question the overview needs answered. Map the factual questions customers ask before they sign a contract. Then make each relevant section answer one clearly, with the useful detail near the top. If your page buries that detail, a competitor’s passage is easier to use.

## “One comprehensive article will capture the category”

I understand the appeal of the _ultimate guide_. Pay for one big asset, cover every angle, and let it collect visibility. The same instinct can produce a sprawling industry hub or sixty programmatic listicles. Either way, volume becomes a substitute for an answer.

![An oversized leather-bound guide sitting bypassed on a desk while a mechanical scanner reads only a concise index card pinned to its front cover.](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/082cec94-f841-467d-a2cc-bee81888f639.png)

An AI system does not have to read your page like a patient human editor. It can retrieve a short passage that answers a narrower question. In an audit of 100 Google AI Overview citations, [CXL found that 55% of cited passages appeared in the top 30% of the source page](https://cxl.com/blog/google-ai-overviews-sources-study/); 48 of the 100 came from the top 20%. Kevin Indig’s analysis of 18,012 ChatGPT citations found that 44.2% came from the first third of a document. Those findings do not mean everything below the fold is worthless. They do give you a good reason to stop hiding the answer behind five paragraphs of introduction.

**The cost is burying your best facts inside the asset you paid to create.** If your team has benchmark data on membrane tensile strength, a reader should not need to reach paragraph eighteen to find it. Use the documentation you already have. Put the conclusion, number, pricing range, or operational limit early in the relevant section, under a heading that says what the section answers. Let the longer explanation earn its space after that.

## “Schema markup is the technical shortcut”

Every new search feature seems to attract someone selling a snippet of code as an express ticket in. For AI Overviews, the pitch is that JSON-LD around pricing tables, author bios, or product specs will get the page cited. I see why it sells: a contained development task feels cheaper and more predictable than rewriting weak documentation.

The evidence in the draft cuts against treating schema as a citation switch. [Ahrefs tracked 1,885 pages that added JSON-LD schema against 4,000 control pages](https://ahrefs.com/blog/schema-ai-citations/) and found no statistically significant citation uplift. Its reported changes were -4.6% in Google AI Overviews, +2.4% in AI Mode, and +2.2% in ChatGPT. Pages that earn citations may also tend to have schema, but that overlap is not proof that the markup caused the citations. Testing reported by [Search Engine Land](https://searchengineland.com/schema-and-ai-overviews-does-structured-data-improve-visibility-446754) likewise found that structured data without external entity authority did not earn them.

**The cost is spending an engineering sprint on labels while the answers remain missing.** Schema can help maintain the structured information used for traditional search features. Keep it clean. But if the visible page hides turnaround times, pricing baselines, and technical tolerances, markup cannot make that page a useful answer. Fix the plain-language content before polishing the container.

## “If I rank on page three, I can skip ahead”

This one is tempting because the ordinary ranking climb can feel brutal. If established competitors dominate your commercial terms, an AI Overview above the blue links looks like a way around them. A consultant offering a shortcut will happily agree.

There is a grain of truth in the pitch. In an analysis of 863,000 keyword results pages and 4 million AI Overview URLs, [Ahrefs found that only 37.9% of cited pages ranked in the organic top 10 for the primary query](https://ahrefs.com/blog/ai-overview-citations/). Another 31.2% ranked in positions 11–100, and 31.0% sat outside the top 100. But the primary query is not the whole retrieval process. A page absent from its top results can still be a strong result for a narrower background question. That is different from Google plucking an unknown site out of thin air.

**The cost is neglecting pages that already have a foothold.** If a relevant URL sits around position 8 or on page two, I would inspect it before commissioning a new batch of speculative pages. [Fix the page already at position 8](https://groas.com/post/stop-writing-ai-overview-content-your-fa): cut the corporate opening, surface the numbers and constraints, and make the specific answer easy to find. Do not confuse a route around one head-term ranking with a route around being findable at all.

## “It’s like buying an ad: pay once and you’re set”

A setup invoice encourages a satisfying story: optimize the site, get a citation, take a screenshot, mark the project complete. Traditional SEO has sold versions of that story for years. Even a profitable paid-search structure can feel stable until the auction changes.

AI citations are retrieval events, not permanent billboards. The answer can change when the prompt changes, when available sources change, or when competitors publish more useful information. In testing discussed by [SparkToro and Gumshoe](https://www.reddit.com/r/SEO/comments/1fq8m2k/how_are_you_guys_actually_measuring_ai_visibility_vs_normal_rankings/), researchers ran prompt variations repeatedly and found that identical brand citations were rare. A screenshot establishes that your brand appeared in one result. It does not establish that buyers will keep seeing it.

![A heavy brass plaque engraved 'PERMANENT CITATION 1ST POSITION' crumbling into drifting sand on a slate table, illustrating citation decay.](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/fc843301-05f4-48bf-81e6-9ba7206625ff.png)

**The cost is mistaking one appearance for durable visibility.** A company can celebrate its first citation, stop maintaining the underlying pages, and miss the change when buyer questions or competing answers move on. Treat this as an ongoing operating task: keep crawl paths clear, check the questions buyers ask, and refresh factual answers when your market changes. A static visibility report will not do that work for you.

## “I’ll know it worked when organic traffic jumps”

Traffic is the familiar scoreboard. Win an ad auction or improve an organic ranking, and you expect more sessions. It is natural to assume a citation in the prominent AI box should make the analytics chart climb too. If sessions stay flat, the work must have failed. Right?

Not necessarily. An AI Overview can answer the question on the results page. In behavioral data covering 68,879 Google searches, [Pew Research Center found that users clicked a traditional result 8% of the time when an AI Overview appeared, versus 15% when one did not](https://www.searchenginejournal.com/google-ai-overview-citations-from-top-ranking-pages-drop-sharply/540502/). That comparison does not tell you what any one citation is worth. It does show why a citation and a jump in organic sessions need not travel together.

![An analytical comparison diagram showing click-through behavior dropping from 15 percent down to 8 percent when an AI summary appears.](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/3f92cec0-a47d-44bc-86e2-7a463f63243d.png)

**The cost is cutting work because the wrong chart stayed flat.** If Google uses your engineering data to answer a buyer’s question, that buyer may encounter your name without visiting the source page then and there. I would track citation presence across relevant prompts alongside branded search and conversions on commercial pages, rather than declaring victory or failure from raw session counts. The goal is qualified demand, not a prettier traffic report.

## The order I’d actually start in

Once I stop treating AI Overviews as placements to buy, the work gets less glamorous and more useful. I would not start with sixty new articles or a screenshot-filled retainer report. I would start here, in order:

1. **Audit pages already within reach.** Find relevant commercial and technical URLs around positions 8–20. They are existing assets worth inspecting before you fund new ones.
2. **Make the useful passages extractable.** Go section by section. Put the direct answer, price range, technical limit, or steps near the beginning, then supply the context a human reader needs.
3. **Check external corroboration.** Compare the operational details on your site with your Google Business Profile, relevant directory listings, reviews, and industry discussions. Contradictory or missing information is worth fixing; markup alone will not settle it.
4. **Track a consistent set of buyer prompts.** Check citation presence across the same kinds of questions over time. One favorable screenshot is a moment, not a baseline.

![A wooden workbench showing an orderly step-by-step restoration setup with labelled components neatly arranged.](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/8669f598-c702-4612-9448-341549480c4a.png)

## “If our answer is right, Google will eventually cite us”

This is the belief I am still not sure about. Technical founders have a good reason to hold it: if their answer is demonstrably better than the industry’s rule of thumb, it seems reasonable to expect the better answer to win.

But retrieval is not a laboratory bench test. If affiliate sites, trade directories, and old forum threads all repeat one claim, while your page alone presents contradictory data, the broader web may be easier for a system to corroborate than your correction. **Being right may not be enough to get cited.** That does not make accuracy optional. It means the work might also include getting sound findings into trade publications, community discussions, and other places buyers encounter them, rather than waiting for one page to overturn the consensus by itself.

I am not ready to say that the best-supported answer always wins. I am even less willing to sell the opposite promise: that paying for one page, one schema sprint, or one agency invoice can make it win by next month.

## Frequently Asked Questions

### Can you pay Google to get your page cited in an AI Overview?

No. An AI Overview is not part of the ad auction: Google retrieves passages relevant to parts of a user's question and synthesizes an answer from them. There is no bid, keyword target, or editorial review process that buys placement in the summary box.

### Is it enough to optimize a page for one target keyword to get an AI Overview citation?

Not really. AI Overviews can break a complex question into narrower searches and retrieve passages that address those parts, so a single head term does not tell the system which technical question to pull from your page. Map the factual questions customers ask before signing, and have each section of the page answer one of them clearly, with the useful detail near the top.

### Where should important facts and answers appear on a page for AI citations?

Early. A CXL audit of 100 Google AI Overview citations found 55% of cited passages appeared in the top 30% of the source page, and Kevin Indig's analysis of 18,012 ChatGPT citations found 44.2% came from the first third of a document. Put the conclusion, number, pricing range, or operational limit at the start of the relevant section, under a heading that states what it answers.

### Does adding JSON-LD schema markup get a page cited in AI Overviews?

Not on its own. Ahrefs compared 1,885 pages that added JSON-LD schema with 4,000 control pages and found no statistically significant citation uplift, reporting -4.6% in Google AI Overviews and +2.4% in AI Mode. Fix the visible plain-language content first; schema can remain for traditional search features, but markup cannot make a page a useful answer.

### Do you need to rank in Google's top 10 for a page to be cited in an AI Overview?

No. In an Ahrefs analysis of 863,000 keyword results pages, only 37.9% of cited pages ranked in the organic top 10 for the primary query, while 31.2% ranked in positions 11-100 and 31.0% sat outside the top 100. Before commissioning new pages, inspect relevant URLs already around position 8 or page two and make their specific answers easy to find.

### Once a page earns an AI Overview citation, does it keep appearing there permanently?

No. AI citations are retrieval events, not permanent billboards: the answer can change when the prompt changes, when available sources change, or when competitors publish more useful information. In testing discussed by SparkToro and Gumshoe, researchers ran repeated prompt variations and found identical brand citations were rare. Treat visibility as an ongoing operating task rather than a one-time project.

### Will an AI Overview citation increase my organic traffic?

Not necessarily. Pew Research Center data covering 68,879 Google searches found users clicked a traditional result 8% of the time when an AI Overview appeared, versus 15% when one did not, because the overview can answer the question on the results page. Track citation presence across relevant prompts alongside branded search and conversions instead of judging success by session counts.

### If my answer is factually more accurate than what's already online, will Google eventually cite it?

Not automatically. If affiliate sites, trade directories, and old forum threads all repeat one claim while a single page contradicts it, the broader web may be easier for a retrieval system to corroborate than the correction. Accuracy stays essential, but the work may include getting sound findings into trade publications and community discussions where buyers encounter them.

## Pay For Results, Not For Hours

Businesses buy the outcome, agencies resell it, and groas answers for it either way.

[See If You Qualify](https://groas.typeform.com/to/xC1bQNUT)

## Structured data

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://graph.groas.com/entity/groas","name":"groas","url":"https://groas.com/","logo":{"@type":"ImageObject","url":"https://groas.com/icon-512.png","width":512,"height":512},"sameAs":["https://www.linkedin.com/company/groas/"]},{"@type":"WebSite","@id":"https://groas.com/#website","name":"groas","url":"https://groas.com/","inLanguage":"en-US","publisher":{"@id":"https://graph.groas.com/entity/groas"}},{"@type":"WebPage","@id":"https://groas.com/post/what-business-owners-get-wrong-when-they#webpage","url":"https://groas.com/post/what-business-owners-get-wrong-when-they","name":"Google AI Overviews Aren’t Ad Slots: Six Beginner Beliefs That Waste Budget","description":"A PPC operator breaks down six paid-search assumptions that fail in Google AI Overviews, what each one costs, and where to start instead.","inLanguage":"en-US","isPartOf":{"@id":"https://groas.com/#website"},"about":{"@id":"https://graph.groas.com/entity/groas"},"primaryImageOfPage":{"@type":"ImageObject","url":"https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/9f0e1b35-adf6-4a10-a148-d2bb6ff87bab.png"},"breadcrumb":{"@id":"https://groas.com/post/what-business-owners-get-wrong-when-they#breadcrumb"},"dateModified":"2026-10-11T05:36:47.705Z","mainEntity":{"@id":"https://groas.com/post/what-business-owners-get-wrong-when-they#article"}},{"@type":"BreadcrumbList","@id":"https://groas.com/post/what-business-owners-get-wrong-when-they#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://groas.com/"},{"@type":"ListItem","position":2,"name":"Blog","item":"https://groas.com/blog"},{"@type":"ListItem","position":3,"name":"Google AI Overviews Aren’t Ad Slots: Six Beginner Beliefs That Waste Budget","item":"https://groas.com/post/what-business-owners-get-wrong-when-they"}]},{"@type":"BlogPosting","@id":"https://groas.com/post/what-business-owners-get-wrong-when-they#article","url":"https://groas.com/post/what-business-owners-get-wrong-when-they","mainEntityOfPage":{"@id":"https://groas.com/post/what-business-owners-get-wrong-when-they#webpage"},"isPartOf":{"@id":"https://groas.com/blog#blog"},"headline":"Google AI Overviews Aren’t Ad Slots: Six Beginner Beliefs That Waste Budget","description":"A PPC operator breaks down six paid-search assumptions that fail in Google AI Overviews, what each one costs, and where to start instead.","datePublished":"2026-10-11T05:36:47.636Z","dateModified":"2026-10-11T05:36:47.705Z","image":{"@type":"ImageObject","url":"https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/9f0e1b35-adf6-4a10-a148-d2bb6ff87bab.png"},"author":{"@id":"https://groas.com/author/alexander-perelman#person"},"publisher":{"@id":"https://graph.groas.com/entity/groas"},"keywords":"google, page, overviews, answer, six, citation, search, pages","wordCount":1805,"inLanguage":"en-US"},{"@type":"Person","@id":"https://groas.com/author/alexander-perelman#person","name":"Alexander Perelman","jobTitle":"Head Of Product @ groas","description":"Ex Goldman Sachs and Ex Stanford Computer Science","image":"https://groas.com/media/blog/4f17cac3a81acc224cc1b2cfab8f94bbc295a3089d13e8421125771ef8d4064f.jpg","worksFor":{"@id":"https://graph.groas.com/entity/groas"},"sameAs":["https://www.linkedin.com/in/alexander-433793253/"]},{"@type":"FAQPage","@id":"https://groas.com/post/what-business-owners-get-wrong-when-they#faq","mainEntity":[{"@type":"Question","name":"Can you pay Google to get your page cited in an AI Overview?","acceptedAnswer":{"@type":"Answer","text":"No. An AI Overview is not part of the ad auction: Google retrieves passages relevant to parts of a user's question and synthesizes an answer from them. There is no bid, keyword target, or editorial review process that buys placement in the summary box."}},{"@type":"Question","name":"Is it enough to optimize a page for one target keyword to get an AI Overview citation?","acceptedAnswer":{"@type":"Answer","text":"Not really. AI Overviews can break a complex question into narrower searches and retrieve passages that address those parts, so a single head term does not tell the system which technical question to pull from your page. Map the factual questions customers ask before signing, and have each section of the page answer one of them clearly, with the useful detail near the top."}},{"@type":"Question","name":"Where should important facts and answers appear on a page for AI citations?","acceptedAnswer":{"@type":"Answer","text":"Early. A CXL audit of 100 Google AI Overview citations found 55% of cited passages appeared in the top 30% of the source page, and Kevin Indig's analysis of 18,012 ChatGPT citations found 44.2% came from the first third of a document. Put the conclusion, number, pricing range, or operational limit at the start of the relevant section, under a heading that states what it answers."}},{"@type":"Question","name":"Does adding JSON-LD schema markup get a page cited in AI Overviews?","acceptedAnswer":{"@type":"Answer","text":"Not on its own. Ahrefs compared 1,885 pages that added JSON-LD schema with 4,000 control pages and found no statistically significant citation uplift, reporting -4.6% in Google AI Overviews and +2.4% in AI Mode. Fix the visible plain-language content first; schema can remain for traditional search features, but markup cannot make a page a useful answer."}},{"@type":"Question","name":"Do you need to rank in Google's top 10 for a page to be cited in an AI Overview?","acceptedAnswer":{"@type":"Answer","text":"No. In an Ahrefs analysis of 863,000 keyword results pages, only 37.9% of cited pages ranked in the organic top 10 for the primary query, while 31.2% ranked in positions 11-100 and 31.0% sat outside the top 100. Before commissioning new pages, inspect relevant URLs already around position 8 or page two and make their specific answers easy to find."}},{"@type":"Question","name":"Once a page earns an AI Overview citation, does it keep appearing there permanently?","acceptedAnswer":{"@type":"Answer","text":"No. AI citations are retrieval events, not permanent billboards: the answer can change when the prompt changes, when available sources change, or when competitors publish more useful information. In testing discussed by SparkToro and Gumshoe, researchers ran repeated prompt variations and found identical brand citations were rare. Treat visibility as an ongoing operating task rather than a one-time project."}},{"@type":"Question","name":"Will an AI Overview citation increase my organic traffic?","acceptedAnswer":{"@type":"Answer","text":"Not necessarily. Pew Research Center data covering 68,879 Google searches found users clicked a traditional result 8% of the time when an AI Overview appeared, versus 15% when one did not, because the overview can answer the question on the results page. Track citation presence across relevant prompts alongside branded search and conversions instead of judging success by session counts."}},{"@type":"Question","name":"If my answer is factually more accurate than what's already online, will Google eventually cite it?","acceptedAnswer":{"@type":"Answer","text":"Not automatically. If affiliate sites, trade directories, and old forum threads all repeat one claim while a single page contradicts it, the broader web may be easier for a retrieval system to corroborate than the correction. Accuracy stays essential, but the work may include getting sound findings into trade publications and community discussions where buyers encounter them."}}]}]}
```