GEO, AEO and SEO: What Each One Optimizes For and Where They Overlap

Three acronyms now compete for the same budget line: SEO, AEO and GEO. They overlap heavily, but each targets a different surface and is judged by a different result, and knowing which is which keeps you from paying twice for the same work.

What SEO still covers

SEO (search engine optimization): the work of making pages crawlable, relevant and trustworthy enough to rank in a search engine’s results. Its success measures are rankings, clicks and the traffic and enquiries that follow.

SEO also remains the entry ticket for Google’s AI features. Google’s documentation on AI features states that to be shown as a supporting link, “a page must be indexed and eligible to be shown in Google Search with a snippet,” and that there are no additional technical requirements.

The implication is practical. If your pages are not indexed, no amount of AI-specific tactics will place them in Google’s AI answers. Fix crawling, indexing and snippet eligibility first.

What AEO adds

AEO (answer engine optimization): shaping content so a system can lift a short, direct answer from it. The older form is the featured snippet, and the newer form includes answers spoken or summarized by AI assistants.

Google’s featured snippets documentation describes snippets as special boxes where the format of a regular result is reversed, and says site owners cannot mark a page as a featured snippet. Google’s systems decide whether a page makes a good one for a given query.

So AEO is less a technique you switch on than a writing habit. Put the direct answer first, define terms in one sentence, and make each section understandable when quoted alone. Our article on query fan-out shows why answering the related sub-questions matters as well.

What GEO research found

GEO (generative engine optimization): the practice of increasing how often and how prominently your content appears in responses from generative search engines. The term comes from a paper by Pranjal Aggarwal and colleagues, accepted to KDD 2024.

The authors built GEO-bench, a benchmark of 10,000 queries, and tested nine content methods. Their abstract states that GEO “can boost visibility by up to 40% in generative engine responses.” The strongest methods were adding citations to sources, adding quotations and adding statistics, while keyword stuffing performed poorly and measured about 10% below baseline on Perplexity.

The paper also reported that results varied by domain and that lower-ranked sites gained more than top-ranked ones. These are results from the authors’ own tests, not guarantees for live engines. They do support a simple habit: make claims specific, sourced and quotable.

The three disciplines side by side

The table summarizes how the disciplines differ in target, measure and levers. The final column notes how strong the public evidence is for each.

Discipline	Target surface	Typical success measure	Main levers	Evidence base
SEO	Ranked results pages	Rankings, clicks, conversions	Crawlability, relevance, links, page quality	Mature, documented by search engines
AEO	One extracted answer, such as a featured snippet	Snippet or answer appearances	Direct answers, clear definitions, structure	Documented for snippets; engine-selected
GEO	Sources cited inside AI responses	Citation frequency and position	Sourced claims, quotations, statistics, clarity	Early; one peer-reviewed benchmark plus vendor studies

Notice that Google’s AI features page does not use the terms GEO or AEO at all. The labels come from the industry, which is why they are defined differently from vendor to vendor.

Where to start, in order

Because the disciplines share most of their foundations, sequence matters more than choosing a label. A workable order for a small team follows.

1. Confirm important pages are indexed, snippet-eligible and not blocked from the crawlers you want, using our guide to robots.txt for AI crawlers.
2. Rewrite key sections so the direct answer comes first and each passage stands alone.
3. Add attributed statistics and named sources inline, following the pattern the GEO paper tested.
4. Treat schema as supportive only, as outlined in what the evidence supports for schema markup.
5. Measure citations over repeated runs, since answers vary, as described in our guide to measuring AI search visibility.

What to ask a vendor who sells GEO

A vendor selling GEO as a separate product should be able to explain what it does that good SEO does not. Google states there are no special optimizations necessary for its AI features, so any claim of a secret technique needs evidence.

Ask which engine the work targets, how citations are measured across repeated runs, and what happens to results when the engine changes. Ask for a before-and-after example with the method stated. If the answer is a proprietary score with no method, treat it as marketing, not measurement.

Our piece on how to choose an AI marketing agency adds general questions that apply here too.

Frequently Asked Questions

Is GEO different from SEO?

GEO targets citations inside AI-generated answers, while SEO targets ranked results. They share foundations such as indexing and content quality, so most of the work overlaps.

Does Google require special optimization for AI Overviews?

No. Google’s documentation says there are no additional requirements or special optimizations needed beyond a page being indexed and eligible to appear with a snippet.

Which discipline should a small business prioritize first?

Start with SEO fundamentals, because indexing and quality gate everything else. Then structure content for direct answers and add sourced evidence, which serves AEO and GEO together.

Separating these labels makes budgets easier to defend and vendor claims easier to test. At Grow with Hafsa we help businesses plan AI search and AI marketing work around evidence instead of acronyms, and we are happy to review where your current effort overlaps.

Google’s Business Agent for Leads: What a Chat Inside the Ad Means for Lead Generation

On May 20, 2026, Google announced a Business Agent for leads: a chat agent inside a search ad that answers questions from the advertiser’s own website instead of sending people to a static form. For lead-generation advertisers, the first sales conversation moves into the ad itself, before anyone reaches your site.

What Google announced

Google’s Ads blog post, “A new generation of ads for the AI era of Search,” published May 20, 2026, says the Business Agent for leads “puts a smart brand agent right inside your ad.” Built with Gemini, it lets a person click “Chat” and get answers rather than fill in a static form.

Google’s example is a student researching universities who receives instant answers based on the university’s website. The post says the format is coming “in the coming months,” so its availability in your account may differ from what this article describes.

The shift matters because a form captures contact details and little else. A conversation can answer objections, narrow intent and pass a warmer lead to your team, if the agent is set up well.

Your website becomes the knowledge base

The detail that deserves the most attention is the source of the answers. Google states that the agent’s answers are based on the advertiser’s website. Whatever your pages say, or fail to say, is what a prospect will hear in the chat.

This is a claim-then-consequence situation. A page that hides pricing ranges, leaves out who the service is not for, or contradicts another page gives the agent nothing reliable to repeat. Pages that state scope, eligibility, timelines and exclusions plainly give it material it can quote accurately.

For service businesses, that means the audit starts on the site, not in the ad account. Our guide to AI crawler access covers the related question of which automated systems can read your pages at all.

Static form versus chat agent

The table below compares the two formats. The first column reflects Google’s description; the remaining rows are our reading of what each format implies for a lead-generation team.

Question	Static lead form	Business Agent chat
Where do answers come from?	No answers; the form only collects details	The advertiser’s website, per Google
What does the prospect do?	Fills fields and waits	Asks questions and gets replies in the ad
What can go wrong?	Low-intent submissions and slow follow-up	Inaccurate or off-brand answers drawn from unclear pages
What your team must own	Follow-up speed	Page accuracy, qualification rules and handoff

The risk moves from the form to the content. A weak form wastes a lead; a weak knowledge base can misinform one.

What Direct Offers shows about Google’s direction

Direct Offers: a Google pilot, begun in January 2026 with brands including Chewy, Gap and L’Oreal, in which advertisers upload promotions and Gemini builds a deal for the shopper. In the May 20 post, Google says brands supply the promotions, the eligible products and guardrails, and that “we use Gemini to construct a deal.”

The pattern is the same as with the Business Agent. The advertiser provides inputs and limits, and Gemini composes the customer-facing exchange. Google also says that to take full advantage of the new formats, advertisers should build a strong foundation with AI Max for Search, AI Max for Shopping and Performance Max.

Google cites a December 2025 Ipsos Global Consumer Journeys survey it commissioned, covering online shoppers aged 18 and over in 25 countries who use AI Overviews or AI Mode for shopping. In it, 75% said they make faster, more confident decisions. The sample is shoppers, not B2B buyers, so treat it as directional. For background on the campaign type involved, see our article on Google AI Max for Search.

How to prepare before it reaches your account

Because Google has not described advertiser controls, preparation should focus on what you can govern yourself. Work through these steps in order.

1. Audit the pages an agent would draw on, and fix contradictions in pricing, service area, timelines and eligibility.
2. Write down what a qualified lead looks like in plain language: budget range, location, company size and urgency.
3. Decide the handoff: who receives a chat that turns serious, how fast, and through which channel.
4. Choose a lead-quality measure beyond form fills, such as qualified meetings, and connect it to your CRM.
5. Start with a capped test budget and review chat transcripts weekly, if the account exposes them.

The same discipline applies to any agent that speaks or spends for you. Our piece on approval gates for agentic ad spend explains how to decide where human sign-off belongs.

What remains unconfirmed

Several practical questions are unanswered in the announcement. Google does not state which industries will receive the format first, whether advertisers can edit or restrict what the agent says, how transcripts are shared, or how it is billed.

That gap is the reason to treat vendor pitches carefully. Anyone claiming to know the controls, the conversion lift or the pricing before Google has published them is guessing. Our overview of what is actually autonomous in agentic advertising offers a way to test such claims.

Frequently Asked Questions

What is Google’s Business Agent for leads?

It is a Gemini-built chat agent inside a search ad. Instead of filling in a form, a person can click “Chat” and receive answers based on the advertiser’s website, according to Google’s May 20, 2026 announcement.

Can advertisers control what the agent says?

Google has not described advertiser controls in the announcement. Until it does, the safest lever is the accuracy and clarity of your own website pages.

Is the format available now?

Google said it was coming “in the coming months” as of May 20, 2026. Check your Google Ads account or your account representative for current availability.

Preparing a site and a qualification process for agent-led conversations is practical groundwork, and it improves ordinary lead handling too. At Grow with Hafsa we help teams build AI marketing and performance-marketing workflows that keep a person in charge of the decisions that matter, and we are glad to talk through your setup.

Reddit and AI Citations: What Semrush’s Data Says About Community Content

AI citation share is the proportion of AI-generated answers that link to a given source. Reddit is one of the most cited sources in AI search, but Semrush’s own data shows that its share can swing sharply within weeks. That matters if you are deciding how much of a GEO (Generative Engine Optimization) budget to put into community content.

How much Reddit’s share moves

The first thing to know is that Reddit’s citation share is not stable. Semrush published a study on November 10, 2025 that tracked weekly snapshots from July 14 to October 12, 2025, covering more than 230,000 prompts and over 100 million AI citations across ChatGPT search, Google AI Mode and Perplexity.

In that data, Reddit appeared in close to 60% of ChatGPT prompt responses in early August, then dropped to around 10% by mid-September. Wikipedia’s ChatGPT share fell over the same period, from roughly 55% to under 20%. Even after the drop, Reddit and Wikipedia remained ChatGPT’s two most-cited domains. Semrush said the data did not establish a cause.

The practical reading is that a single platform’s citation pattern can change without notice. A strategy that depends on one source type being favored is exposed to the next shift, which is why we recommend tracking citations over time, as described in our guide to measuring AI search visibility.

Which Reddit content gets cited

Semrush’s second study, also published November 10, 2025, looked at 217,000 prompts on Google AI Mode, Perplexity and ChatGPT Search, and identified 248,000 unique Reddit URLs cited or mentioned in responses. It found that question-and-answer threads alone accounted for more than half of all Reddit citations, and that Q&A, comparison and discussion posts together made up nearly three-quarters.

Measure	Finding in the Semrush Reddit study
Q&A threads	More than half of Reddit citations
Q&A, comparison and discussion combined	Nearly three-quarters of cited content
Cited posts with fewer than 20 upvotes	80%
Average age of a cited post	About 900 days
Median length of a cited post	About 80 words

The pattern is that assistants favor short, direct answers to specific questions, not viral posts. That mirrors what we see elsewhere: clear answers to buyer questions are the content type AI systems quote, which is also the logic behind query fan-out.

What the studies do not show

These are descriptive studies by a company that sells visibility tracking, and they describe which Reddit pages were cited, not whether a brand’s participation produced a citation or a customer. Semrush’s own write-up gives slightly different platform shares in its body and its summary, so we have not quoted a platform-by-platform percentage.

The data also stops in October 2025. Other trade sources we reviewed report conflicting later figures for Reddit on ChatGPT, and we could not trace them to primary datasets, so we left them out. The age statistic is a caution, too: because cited posts average about 900 days old, a thread written today may take a long time to matter, if it ever does.

None of this shows that Reddit marketing drives revenue. For a B2B service business, the more reliable test is whether buyers in your category mention Reddit threads during evaluation, which you can check by asking them.

How to act on it

Reddit communities enforce rules against self-promotion, and moderators remove content that reads as advertising. The sensible approach is to treat Reddit as a research source and a place for genuine, disclosed participation, not as a channel to manufacture citations.

1. Run your ten most valuable buyer questions through ChatGPT, Perplexity and Google AI Mode and record whether Reddit threads are cited.
2. Read the cited threads to learn what buyers ask and which objections they raise.
3. Publish clear, sourced answers to those questions on your own site first, since that is the asset you control.
4. Where you have real expertise, answer questions in relevant communities openly, with your affiliation disclosed.
5. Re-run the same questions monthly and log changes, because the pattern may shift.

Our analysis of brand mentions versus backlinks covers the wider question of how third-party signals relate to AI visibility.

Frequently Asked Questions

Is Reddit the most cited source in AI search?

Semrush found Reddit among the top five most-cited domains on ChatGPT search, Google AI Mode and Perplexity in 2025, but its ChatGPT share fell from close to 60% to around 10% within weeks.

What type of Reddit post do AI tools cite most?

Question-and-answer threads. Semrush’s analysis of 248,000 Reddit URLs found they accounted for more than half of Reddit citations.

Should a business invest in Reddit for AI visibility?

Only modestly and honestly. The data shows Reddit is cited, but not that participation produces citations or leads, so your own site’s answers should come first.

Sorting out which third-party sources matter in your market is a measurement job before it is a content job. At Grow with Hafsa we build AI marketing and search-visibility systems around that evidence, and we are happy to talk through what your buyers are seeing.

Google’s Read-Only Ads MCP Server and the Case for Approval Gates on Agentic Ad Spend

An approval gate is a checkpoint where a human must confirm an AI agent’s proposed action before it changes live spend. Two recent developments show where the industry is placing that checkpoint: Google’s official Ads MCP server cannot change anything, and IAB Tech Lab’s latest agentic advertising release calls for human approvals on spend commitments.

What Google’s official server allows

MCP, the Model Context Protocol, is an open standard that lets an AI agent call tools in another system through a common interface. Google publishes an MCP server for Google Ads, and its design is conservative.

Google’s Ads API documentation, last updated October 16, 2026, lists three tools: list_accessible_customers, search for Google Ads Query Language queries, and get_resource_metadata. The page labels the current release as read-only and states that the implementation “cannot modify bids, pause campaigns, or create new assets.” The server’s GitHub repository adds a warning that connecting it exposes your account data to whichever agent or language model you attach.

For a buyer, the implication is practical. An agent wired to this server can audit, report on and diagnose an account, but it cannot spend. That makes it a low-risk starting point for analysis work, and a poor fit for any vendor claiming Google itself endorses autonomous campaign changes through this route.

Where write access comes from

If the official server only reads, agents that change campaigns must get write capability elsewhere. Google’s documentation points developers who need mutate operations to its Google Ads API Developer Assistant, which generates client-library code for those calls.

The market has also filled the gap. Markifact, for example, announced global availability of a hosted Google Ads MCP server on July 13, 2026, and described it as one with human-approved changes for use with Claude and ChatGPT. Other vendors advertise servers that adjust budgets and bids. We could not independently test these products, so treat their safeguards as claims to verify, not facts.

The consequence is that the safety model is no longer set by the platform. It is set by whoever builds the write layer, which is why an advertiser should ask what stops an agent from overspending before signing anything. Our guide to what is actually autonomous in agentic advertising covers how much of the market’s autonomy claims hold up.

What IAB Tech Lab’s AAMP 2.3 asks for

IAB Tech Lab, the industry standards body for digital advertising, announced AAMP 2.3 on July 30, 2026. AAMP is its framework for agent-to-agent advertising transactions between buyer agents and seller agents.

The release states that “any path that commits spend is deterministic and provable with human approvals outside of value-based thresholds.” In plain terms, small transactions may proceed on rules alone, while larger ones need a person to sign off, and the path to every spend decision must be reproducible. The release also adds a vendor approval gate to the Buyer Agent, tied to the IAB Diligence Platform.

PPC Land’s coverage of the release reports that the vendor gate is off by default and stays inactive without a verification key. We did not review the full specification, so the details of how thresholds are configured should be read from IAB Tech Lab’s own repository. The principle is what matters for buyers: autonomy is bounded by value, and the bound is a human decision.

Three levels of agent authority

The two developments point to a simple way of classifying any agentic ads setup by what the agent is permitted to do. The table below is our own framework, not an industry standard.

Level	What the agent can do	Typical control	Fits
Read-only	Query performance, flag anomalies, draft recommendations	Account permissions, as in Google’s official server	Audits, reporting, diagnosis
Write with approval	Propose changes that a person confirms before they apply	Per-change sign-off, change log	New campaigns, large budget moves, new audiences
Autonomous within caps	Apply small changes automatically	Hard daily and monthly spend ceilings, allowed-action list	Routine bid and budget adjustments

Most teams should start at the first level, earn their way to the second, and move to the third only for actions with a bounded downside. The AAMP wording maps onto the same logic: rules handle the routine, people handle what exceeds a value threshold.

Questions to ask an agentic ads vendor

Any vendor that touches your ad accounts should be able to answer the following in writing, in this order.

1. Which actions can the agent take without approval, and which require a named person to confirm?
2. Where are the spend ceilings enforced: in the vendor’s software, or in the ad platform’s own account settings you control?
3. Does every change appear in a log that records what the agent proposed, who approved it and what the account looked like before?
4. What happens when the approval service or the model is unavailable: does the agent stop, or continue?
5. Who is liable for a mistaken change, and is that stated in the contract?

Vendors with a clear design answer these quickly. Hesitation on the fourth question is particularly informative, because a safe system should stop when its control layer fails.

Frequently Asked Questions

Can an AI agent change my Google Ads campaigns through Google’s official MCP server?

No. Google’s documentation describes the server as strictly read-only, so it cannot modify bids, pause campaigns or create assets. Changes require the Google Ads API or a third-party server.

What spend threshold needs human approval?

No standard sets a number. IAB Tech Lab’s AAMP 2.3 refers to human approvals outside value-based thresholds, leaving each organization to decide where its own line sits.

Is a read-only agent still useful?

Yes. Reading performance data, finding anomalies and drafting recommendations are real efficiency gains, and they carry no risk to live spend.

Deciding where an agent should read, propose or act is a design question before it is a tooling question. At Grow with Hafsa we help teams set up AI agents and performance-marketing workflows with approval paths that match their risk, and we are glad to review a setup you are considering.

Content Freshness and AI Citations: Does Updating Old Pages Help?

Content freshness is how recently a page was published or substantively updated. Ahrefs found that AI assistants cite newer pages than Google’s organic results do. The gap is real but modest, and the same study warns against treating frequent edits as a strategy.

What the Ahrefs study measured

The question behind the study is a budget question: should a team spend hours refreshing old pages in the hope of earning more AI citations? Ahrefs tested it against data instead of opinion.

In an analysis published July 28, 2025, Ryan Law of Ahrefs examined about 16.975 million cited URLs, described as 17 million citations, drawn from Ahrefs’ Brand Radar tool. The sources were ChatGPT, Perplexity, Gemini, Copilot, Google AI Overviews and organic Google results. Page age was measured from the date Ahrefs’ crawler first saw the page, and last-updated age from the last published date recorded in its Content Explorer index.

The headline result was that AI assistants cited pages averaging 1,064 days since publication, compared with 1,432 days for URLs in organic results. Ahrefs presented the difference as about 25.7% fresher for assistants.

How platforms differ

The aggregate hides a spread between systems, and that spread matters because Google still handles most searches. In Ahrefs’ figures, ChatGPT and Copilot cited the newest pages, Perplexity and Gemini sat in the middle, and Google’s organic results and AI Overviews leaned toward older pages.

Source	Average days since publication
ChatGPT (references)	1,023
Copilot	1,056
Gemini	1,118
Perplexity (in-text order)	1,166
Organic Google results	About 1,432

Ahrefs also reported that Perplexity and ChatGPT tend to order their in-text references from newest to oldest. Ahrefs states that Google is the least influenced by freshness. The practical reading is that a refresh program helps most where your buyers use assistants, and least for classic Google rankings.

What the data does not show

This is a descriptive study, not an experiment. It shows which pages assistants cited, not that updating a page caused a citation. Newer pages may simply cover newer topics, which assistants are asked about more often.

Ahrefs lists its own caveats. Cited content still averaged about 2.9 years old, which means assistants continue to lean on durable pages. Freshness is one of many factors, and frequently updated low-quality content will not help much. The article also notes that Google’s John Mueller has warned against changing a publish date without changing the page itself.

The sample also describes citations in aggregate, not outcomes for any single site. We found no independent replication in this review, so treat the figures as one well-documented dataset from a vendor that sells visibility tracking.

Where this changes a content budget

The decision-relevant finding is the authors’ own conclusion: for most brands, new, high-quality content may return more than frequent updates, given the opportunity cost. Refreshing is justified where a page makes claims that expire, such as pricing, statistics, product capabilities, regulations or dated recommendations.

That points to a triage approach. Pages whose facts decay deserve a scheduled review. Evergreen explainers can be left alone until something material changes. A visible “last reviewed” date is useful to readers, but only if the review was real. Our guides to llms.txt and AI citations and schema markup for AI search apply the same evidence-first test to other popular tactics.

A practical refresh workflow

The sequence below keeps the effort tied to measurable change instead of cosmetic edits.

1. Export your pages that already earn traffic or appear in AI answers, and mark which contain time-sensitive facts.
2. Check each flagged claim against its primary source and replace anything outdated, with the source and year named in the sentence.
3. Add genuinely new material, such as a changed comparison, a new data point or a corrected recommendation, rather than rewording.
4. Update the displayed date only when the substance changed.
5. Re-test the buying questions in ChatGPT and Perplexity after a few weeks, using the approach in our guide to measuring AI search visibility.

Frequently Asked Questions

Should I update the publish date on every page?

No. Ahrefs notes that Google’s John Mueller has warned against changing publish dates without changing content. Update dates only when the page has materially changed.

Do older pages stop being cited by AI assistants?

No. Ahrefs found that cited content still averaged about 2.9 years old, so assistants continue to cite long-lived pages.

Is a refresh better than publishing something new?

Ahrefs suggests most brands may see better returns from new, high-quality content. Refreshes pay off mainly on pages with claims that expire.

Deciding which pages deserve a refresh and which deserve a replacement is an editorial and measurement problem together. At Grow with Hafsa we build AI marketing and content systems that keep that process evidence-led, and we are glad to talk through your content library.

Query Fan-Out Explained: How One AI Question Becomes Ten Searches

Query fan-out is the technique AI search systems use to split one question into several related searches and combine the results. Google confirms AI Overviews and AI Mode use it. A 2025 study shows how many searches one prompt can trigger, and that most of them are invisible to keyword tools.

What Google says query fan-out does

Query fan-out: a retrieval technique in which an AI system turns one user question into multiple related searches across subtopics and data sources, then assembles a single answer from the combined results.

Google’s own Search documentation on AI features states that both AI Overviews and AI Mode “may use a ‘query fan-out’ technique,” and that this helps its models identify more supporting web pages. Google adds that these features can show a wider and more diverse set of links than a classic search results page.

The same page is explicit about eligibility. A page must be indexed and eligible to appear in Search with a snippet, and Google says there are no additional technical requirements or special optimizations for AI features beyond standard SEO best practices. For a buyer, that is a useful anchor: fan-out changes which questions your pages must answer, not the technical rules for being included.

What the Seer Interactive study measured

Seer Interactive’s R&D lead, Nick Haigler, published an initial analysis on November 21, 2025. It ran 501 prompts that Seer tracks for itself and clients through the Gemini 3 API, across several industries, and recorded every search Gemini issued while answering.

Gemini 3 averaged 10.7 fan-out queries per prompt, in a range from 3 to 28. That was up 78% from an average of 6.01 for Gemini 2.5. The queries averaged 6.7 words, and 21.3% contained a year, mostly 2024 or 2025. Only about 1% of fan-out queries appeared more than once across the whole dataset, so almost every query was unique.

One caveat matters before you act on any of this. Seer forced Google Search grounding on every prompt, so the study does not measure how often Gemini searches on its own. Treat the figures as a picture of the mechanism, from one vendor and one model, not as a rate for AI Mode in general.

Why keyword tools cannot see it

The study’s most useful finding for planning is that 95% of fan-out queries had zero global monthly search volume. Keyword tools build volume figures from what people type into search engines. A machine-generated query such as a specific comparison of two vendors on one attribute was never typed by anyone, so it has no volume to report.

That has a direct consequence for how marketing teams prioritize topics. A volume-first keyword plan will systematically miss the long-tail questions an AI system asks on a buyer’s behalf. It will also overstate the value of a single head term when the answer is assembled from a dozen narrower lookups.

Observation	Figure	Source and scope
Average fan-out queries per prompt	10.7 (range 3 to 28)	Seer Interactive, Nov 2025, Gemini 3, 501 prompts
Fan-out queries with zero monthly search volume	95%	Seer Interactive, Nov 2025
Fan-out queries naming a brand	26.4%	Seer Interactive, Nov 2025
Fan-out queries containing a year	21.3%	Seer Interactive, Nov 2025
Fan-out queries shared across the dataset	About 1%	Seer Interactive, Nov 2025

What it changes about content planning

If one buyer question becomes about ten searches, then a page that answers only the headline question competes for one of ten slots. Haigler observed that fan-outs tend to expand a prompt into comparisons, attributes and deeper details, not simple rewordings. Pages that cover those angles clearly are the ones available to be retrieved.

Brand names add a second effect. In the Seer data, 26.4% of fan-out queries included a brand, and the author noted that in one payroll-software example four of sixteen fan-outs named a brand that then appeared in the answer. If systems already associate a vendor with a category, they may search for it directly. That fits the pattern in our review of brand mentions versus backlinks, although neither dataset proves cause.

The year pattern is a hypothesis only. Haigler suggests that years in fan-outs may be recency cues. Keeping dates and facts current on comparison pages costs little, and our piece on what AI search means for SEO covers the wider context.

A practical order of work

Start with the questions your buyers ask, because those are what an AI system will decompose. The steps below avoid guessing at hidden queries.

1. List five to ten real buying questions from sales calls and support tickets, phrased as a buyer would phrase them to an assistant.
2. For each, write out the comparisons, attributes, pricing, integrations and risks a careful researcher would check, as a stand-in for the likely sub-searches.
3. Check whether you have a clear, indexable page answering each sub-question, and fill the gaps with specific, dated, sourced content.
4. Run the original questions through AI Mode and ChatGPT and note which sources appear, following our guide to measuring AI search visibility.
5. Repeat quarterly, since the number and type of sub-queries differ by model and change as models change.

Frequently Asked Questions

Can I see the fan-out queries Google generates for my topic?

Not directly in Google’s own reporting. Researchers like Seer Interactive capture them by calling model APIs with search grounding enabled, which approximates but does not replicate what consumer AI Mode does.

Does query fan-out replace traditional SEO?

No. Google’s documentation says pages need to be indexed and eligible for snippets, with no extra technical requirements. Fan-out raises the value of topical coverage on top of that foundation.

Is 10.7 queries per prompt a number I can plan around?

Only as an indication. It comes from one model, 501 prompts and forced search. Real behavior will vary by question type, product and model version.

Mapping a buyer’s real questions to the sub-searches behind them is research work that pays off across both classic and AI search. At Grow with Hafsa we help teams build AI-ready content and search visibility systems, and we are glad to review your current coverage with you.

Fan-out queries often surface forum threads alongside publisher pages. Our analysis of Reddit and AI citations looks at which community content assistants actually cite.

Does AI Search Traffic Convert Better? What Three Studies Actually Found

Visitors who arrive from ChatGPT, Perplexity or Copilot are often described as high-intent. That claim drives budget decisions about GEO (Generative Engine Optimization), the practice of earning citations in AI answers. Before funding it, check what the underlying studies measured and on which kinds of sites.

Publisher data from Microsoft Clarity

Microsoft Clarity, a behavior analytics product, published a study in late 2025 covering 1,277 publisher and news domains. Conversion rates came from one month of data, while traffic growth covered eight months. Clarity detected conversions with what it calls smart events.

LLM referrals produced a 1.66% sign-up rate against 0.15% for search, 0.13% for direct and 0.46% for social. For subscriptions the rates were 1.34% for LLMs and 0.55% for search. Clarity also reported that AI referrals were still under 1% of overall traffic.

The implication is limited. The sample is publishers, the window is one month, and the base is tiny. A roughly 11x gap on sign-ups does not transfer to a B2B services site or a store, and Clarity sells analytics, which is worth noting when weighing the framing.

Retail data from Adobe Analytics

Adobe reported in April 2026 that AI traffic to U.S. retail sites rose 393% in the first quarter of 2026 year over year. Its analysis drew on online transactions covering more than 1 trillion visits to U.S. retail sites.

In March 2026, Adobe said AI traffic converted 42% better than non-AI traffic and generated 37% more revenue per visit. A year earlier, in March 2025, AI traffic converted 38% worse. The reported engagement gaps were 48% more time on site and 13% more pages per visit.

The reader implication is the direction of travel. The same dataset flipped from worse to better within twelve months, so any single-year benchmark ages quickly. Adobe does not publish a precise definition of AI traffic in this release, which limits how closely you can compare it with your own.

The academic counterpoint

Maximilian Kaiser of the University of Hamburg and Christian Schulze of Frankfurt School of Finance and Management studied 973 e-commerce sites with 12 months of first-party data. Organic search outperformed ChatGPT referrals by roughly 13%, and affiliate links were 86% more likely to convert.

The authors also found ChatGPT conversion rising month over month, describing gradual convergence with traditional channels. They use last-click attribution, which misses earlier ChatGPT influence, and the data covers under 0.2% of traffic. The paper was under review and not yet peer-reviewed when covered by Digiday.

Study	Sample	Finding	Main limit
Microsoft Clarity	1,277 publisher and news domains	1.66% sign-ups from LLMs vs. 0.15% from search	One month of conversion data; publishers only
Adobe Analytics	Over 1 trillion U.S. retail visits	AI traffic converted 42% better in March 2026	AI traffic not precisely defined
Kaiser and Schulze	973 e-commerce sites, 12 months	Organic search about 13% ahead of ChatGPT	Last-click attribution; not peer-reviewed

How to test it on your own site

Your analytics can settle this faster than a literature review, provided you separate AI referrals from the rest. Use the same conversion event for every channel so the comparison is fair.

1. Create a segment for referrers such as chatgpt.com, perplexity.ai, copilot.microsoft.com and gemini.google.com.
2. Pick one conversion that matters, such as a booked call or a qualified form submission, and not a page view.
3. Compare that segment against organic search over at least 90 days, and note the number of sessions behind each rate.
4. Add a self-reported “how did you hear about us” field, because many AI-influenced visits arrive later as direct traffic.

This is where attribution limits bite, a theme we covered in why marketing attribution models disagree with each other. For the citation side, see how to measure AI search visibility when answers change every run.

Frequently Asked Questions

Does AI search traffic convert better than organic search?

Studies disagree. Clarity and Adobe report higher conversion for AI referrals, while a 973-site academic working paper found organic search ahead by about 13%. Industry, conversion event and time period explain much of the gap.

How much traffic comes from AI search?

Clarity’s study of 1,277 domains found AI referrals still account for under 1% of overall traffic, although they grew 155.6% over eight months. Check your own referral report for the true share.

Should I invest in GEO based on conversion rates alone?

Conversion rate is one input. Volume, cost to earn citations and the effect on brand searches matter too. A small, well-measured pilot gives you your own numbers before a larger commitment.

Small traffic with strong intent can still be worth pursuing, but only if you can see it in your reporting. At Grow with Hafsa we build AI marketing and performance measurement setups that separate AI referrals from other channels, so you decide on your own data and not on a vendor’s headline.

AI Overviews and Click-Through Rate: What Pew and Ahrefs Actually Measured

AI Overviews are Google’s generated summaries at the top of some search results. Two studies, one from Pew Research Center and one from Ahrefs, are quoted most often when people ask what they do to clicks. Quoting them side by side without checking what each one measured leads to bad forecasts.

What Pew measured

Pew Research Center’s July 2025 analysis tracked the Google browsing of 900 U.S. adults in an online panel who had installed a tracking app. The data covers March 2025. Pew then reran each search in April 2025 to record whether an AI summary appeared.

Users clicked a traditional result link in 8% of visits to pages with an AI summary and in 15% of visits to pages without one. Only 1% of visits with a summary included a click on a link inside the summary itself. About 18% of the 68,879 unique searches in the sample produced a summary.

The implication is that Pew describes user behavior across everyday searching, not the performance of any ranking position. Pew also notes that summaries change over time, so the April rerun reflects how they looked when collected.

What Ahrefs measured

Ahrefs published its update on February 4, 2026, using aggregated Google Search Console data from its own tool, desktop only. It compared 150,000 keywords that triggered an AI Overview with 150,000 informational keywords that did not, using December 2023 as the pre-rollout baseline and December 2025 as the current period.

For position one, CTR on AI Overview keywords fell from 0.073 to 0.016. Ahrefs forecast what that CTR would have been without the feature, using the decline in the non-AI-Overview group, and put the gap at 58%. Its April 2025 version of the study had reported 34.5%.

The reduction was smaller further down the page, ranging from 50.8% at position two to 19.4% at position ten. Ahrefs describes the result as a correlation and acknowledges that zero-click trends such as featured snippets began before AI Overviews.

Why the numbers are not interchangeable

The two studies answer different questions. Pew asks how often searchers click anything when a summary is present. Ahrefs asks how much a top-ranked page’s CTR changed relative to a modeled counterfactual. A 58% relative drop and a drop from 15% to 8% of visits are different kinds of statements.

Dimension	Pew Research Center	Ahrefs
Published	July 2025	February 2026
Data period	March 2025	December 2023 vs. December 2025
Sample	900 U.S. adults, 68,879 unique searches	300,000 keywords
Source	Browsing-app panel data	Search Console data, desktop only
Headline	8% vs. 15% of visits with a result click	58% lower CTR at position one
Main limit	Google only; summaries rerun a few weeks later	Correlation based on a modeled baseline

The reader implication is practical. If you quote “58%” in a board deck as your expected traffic loss, you are applying a relative CTR change from one ranking position to your whole site. Your own query mix decides how many of your keywords trigger a summary at all.

How to re-baseline your own reporting

Your search console data already contains the answer for your site. The steps below use only data you own, so the result does not depend on whose study you trust.

1. Export query-level clicks, impressions and average position from Google Search Console for the last 16 months.
2. Split queries into informational and commercial intent, since Ahrefs’ comparison group was informational keywords.
3. Check a sample of your top queries in a logged-out browser to see which trigger an AI Overview, and tag them.
4. Compare CTR at the same average position before and after your tagged queries began showing summaries.
5. Report leads and revenue from organic search next to clicks, so a CTR decline is read against the outcome that pays.

Our guide on measuring AI search visibility when answers change every run covers the citation side of the same problem. Our overview of what AI search means for SEO in 2026 explains where these summaries fit among the other answer engines.

Frequently Asked Questions

Do AI Overviews reduce organic clicks?

Both Pew and Ahrefs report lower click rates when a summary is present. Pew found 8% of visits with a result click versus 15% without. Ahrefs found a 58% lower CTR at position one, which it labels a correlation.

Is the 58% figure the same as a 58% traffic loss?

No. It is the change in CTR for the top-ranking page on keywords that show an AI Overview, relative to a modeled baseline. Site-wide traffic also depends on how many of your keywords trigger a summary.

Do links inside an AI Overview get clicked?

Rarely, according to Pew’s March 2025 data, where 1% of visits to pages with a summary included a click on a link inside the summary. Being cited builds visibility, but it has not replaced the click.

Click data tells you where attention moved, not which visitors turned into customers. At Grow with Hafsa we help teams connect AI search visibility to AI marketing and performance measurement, so decisions follow revenue and not only a CTR chart.

Agentic SEO and Google’s Spam Policies: Where Automation Crosses the Line

Agentic SEO means AI agents carry out SEO tasks such as audits, content drafting and on-page changes with limited human input. Google’s spam policies judge the output, not the tool. That makes the approval process around an agent the thing a buyer should evaluate first.

What Google’s policy actually says

Google’s published spam policies define scaled content abuse as many pages generated “for the primary purpose of manipulating search rankings,” typically large amounts of unoriginal content that gives users little or no value. The policy states this applies “no matter how it’s created.”

Its example list includes using generative AI tools to generate many pages without adding value for users, scraping feeds or search results to create pages, and stitching together content from other pages without adding value. The list is described as not limited to these examples.

The reader implication is that the policy does not ban automation. It targets low-value volume. An agent that audits a site and proposes fixes sits far from the policy’s concern, while an agent told to publish hundreds of near-duplicate pages sits squarely inside it. Our explainer on what agentic SEO is separates those two modes.

Why the 2024 enforcement matters

Google’s March 5, 2024 announcement said the updated policy covers mass-produced content “whether automation, humans or a combination are involved,” widening a rule that had focused on clearly automated content. It said the combined spam and ranking changes were expected to reduce low-quality, unoriginal content in search results by 40%.

An April 26, 2024 update to the same post reported a 45% reduction after the rollout finished. These figures are Google’s own and describe results across its index, so they show the scale of enforcement rather than the risk to any single site.

For a buyer, the point is that enforcement is established policy and not a one-off event. A vendor that sells agentic content at volume should be able to explain how each page adds something a reader could not get from the pages it was built from.

What Google says about AI content

Google’s guidance on generative AI says the technology “can be particularly useful when researching a topic, and to add structure to original content.” The same page warns that using it to create many pages with no added value may violate the scaled content abuse policy.

It also says generative models predict likely word sequences rather than retrieving facts, so outputs can contain errors, and publishers should “manually factcheck and review all AI-generated content for accuracy.” That review extends to title elements, meta descriptions, structured data and image alt text, since these can appear in search results.

The practical reading is that Google expects a person to be accountable for what an agent publishes. This also bears on quality signals, which we covered in what AI content detection means for E-E-A-T.

Where to put human approval

Not every task carries the same risk. The table below is our own recommendation, based on the policy language above rather than on any Google rule about specific tasks.

Agent task	Policy risk	Suggested human role
Technical audits and issue lists	Low, since nothing publishes	Review findings before changes ship
Keyword and topic research	Low	Check for overlap with existing pages
Drafting new articles	Moderate	Fact-check claims and add original expertise before publishing
Bulk page generation	High	Approve a sample, and cap volume until quality is shown
Metadata and structured data changes	Moderate	Validate and review, as Google’s guidance advises

Our review of schema markup for AI search explains why structured data changes deserve the same scrutiny as copy.

Questions to ask an agentic SEO vendor

A vendor’s answers to the questions below show how much of the process is judgment and how much is volume. Ask for specifics, and treat vague answers as an answer.

1. Which tasks does the agent perform without a human reviewing the output first?
2. Who fact-checks claims in AI-drafted content, and how is that recorded?
3. How does the system prevent near-duplicate pages across a site?
4. What happens to content that fails review, and who can stop a publishing run?
5. Can you show a client site where agent-assisted content still ranks after a core update?

This pairs with the broader selection criteria in how to choose an AI marketing agency.

Frequently Asked Questions

Does Google penalize AI-generated content?

Google’s spam policy targets many pages made mainly to manipulate rankings, “no matter how it’s created.” Its guidance describes generative AI as useful for research and structure when content adds value.

Is a fully autonomous SEO agent safe to use?

It depends on the task. Audits and research carry little policy risk, while unreviewed bulk publishing is the pattern Google’s policy describes. Keep a human gate on anything that publishes.

Do I need to disclose that content was AI-assisted?

Google’s guidance says sites that generate content automatically should consider telling readers how it was made. It frames this as something to consider, not a ranking requirement.

Agents are most useful in SEO when a person stays accountable for what ships. At Grow with Hafsa we design AI agent and performance marketing workflows with review gates built in, so automation saves time without putting your site at risk.

Ads Inside AI Answers: What Google AI Mode and ChatGPT Ads Mean for Advertisers

Ads inside AI answers are advertising placed within a generated response, rather than beside a list of links. Google and OpenAI both moved in that direction in 2026. For a business buying search or social media ads today, the useful question is what to prepare, and what is still unconfirmed.

What Google announced for AI Mode

Google’s May 20, 2026 Marketing Live announcement said it was testing two new ad types built with Gemini, its AI model family. Conversational Discovery ads: Gemini builds the creative around the specific question a person asked. Highlighted Answers: when AI Mode returns a list of recommendations, Google says “highly relevant, high-quality ads are eligible to appear on that list.”

Google also said both formats will continue to be labeled “Sponsored” and will carry an independent AI-generated explainer alongside the advertiser’s creative. According to The Register’s coverage the next day, Google said Highlighted Answers use standards similar to its existing ad filtering and the same auction mechanics, but did not publicly define what counts as high quality.

The implication is that a strong account foundation matters more than a special AI Mode setup. Google’s announcement recommends building on AI Max for Search, AI Max for Shopping and Performance Max, which our guide to Google AI Max for Search explains.

What OpenAI has opened up in ChatGPT

OpenAI began a ChatGPT ads pilot in February 2026. According to Search Engine Roundtable, it announced a beta self-serve Ads Manager for U.S. advertisers on May 5, 2026, letting businesses register, add payment details, set budgets and bids, upload ads and view performance in a portal.

Marketing Dive reported on May 11, 2026 that OpenAI added cost-per-click bidding, after an early version limited to cost per thousand impressions, along with a Conversions API and pixel-based measurement. The same report said the pilot expanded on May 7 to the U.K., Mexico, Japan, Brazil and South Korea, and that access would open to more businesses gradually.

For a U.S. or U.K. buyer, the practical change is that ChatGPT can now be tested like any other paid channel, with conversion tracking you control. Marketing Dive also noted that a late-April OpenAI privacy policy update covers receiving purchase data from advertisers, which your legal team should read before you share customer data.

How the two compare

The two platforms are at different stages, and the gaps in public detail are as informative as the features. The table below uses only what the cited reports state.

Question	Google AI Mode ads	ChatGPT ads
Status	New formats in testing (Google, May 20, 2026)	Beta self-serve Ads Manager for U.S. advertisers (May 2026)
How you buy	Through existing Google Ads campaigns, with AI Max and Performance Max recommended	Through OpenAI’s own Ads Manager, or through ad-tech partners
Bidding	Same auction mechanics as existing ads, per Google via The Register	Cost per click and cost per thousand impressions
Labeling	“Sponsored,” plus an independent AI explainer	Not specified in the sources we reviewed
Measurement	Google Ads reporting	Conversions API and pixel

Both channels are early, so treat any benchmark you see for them as provisional until the platforms publish methodology.

What is still unconfirmed

Several claims circulate in trade coverage that we could not verify against a primary source, so we have left them out. These include pricing figures for ChatGPT ads, click-through comparisons and claims about how many AI Mode sessions end without a click. A figure without a named study and method is not a basis for a budget decision.

Google’s own announcement did not give a rollout date for the new formats beyond “testing,” and said AI-powered Shopping ads and Business Agent for Leads, a chat agent inside the ad, were expected in the coming months. The sources we read did not state where ChatGPT ads appear within a conversation or which user plans see them.

The implication is simple. Plan for these formats as a test line, not as a replacement for your core search and social spend, and recheck availability before each quarter’s planning, because the details are changing quickly.

How to prepare without overspending

The steps below fit a small team and rely on inputs you already control. They also overlap with the groundwork for agentic commerce, where product data quality decides what AI systems can use.

1. Audit your product or service data, because Gemini writes creative from it and a thin feed produces thin ads.
2. Confirm your conversion tracking records qualified leads or revenue, not only clicks, before any new channel goes live.
3. Run your Google campaigns on AI Max or Performance Max with a clear conversion goal, so you are eligible when formats open.
4. Set a small, capped ChatGPT test budget that you can lose, and judge it by cost per qualified result in your own CRM.
5. Review landing pages for clear answers, since the AI explainer will describe your offer in neutral terms next to your ad.

As we noted in our look at agentic advertising in 2026, platforms automate delivery, but the offer, the data and the measurement stay with you.

Frequently Asked Questions

Do I need a separate campaign to appear in Google AI Mode ads?

Google’s announcement does not describe a separate AI Mode campaign type. It recommends building on AI Max for Search, AI Max for Shopping and Performance Max, so check Google Ads for current eligibility.

Are ads in AI answers labeled?

Google said its two new AI Mode formats will continue to be labeled “Sponsored.” The sources we reviewed did not specify the labeling format for ChatGPT ads.

Is ChatGPT advertising available to U.K. businesses?

Marketing Dive reported that OpenAI’s ads pilot expanded to the U.K. on May 7, 2026, while the self-serve Ads Manager beta launched for U.S. advertisers. Confirm current access in your account.

Ads inside AI answers reward businesses with clean data, honest measurement and a clear offer, which is the work we focus on first. At Grow with Hafsa we build AI marketing and performance systems that make new channels like these easy to test and easy to stop.

Google has since announced a related format, a chat agent that answers from an advertiser’s website inside the ad. Our article on Google’s Business Agent for leads explains what it means for lead generation.