Sited
AI Visibility Checker
A 6.77m-session study shows ChatGPT sends 92.4% of standalone AI referral traffic, up from 84%. Here's what it means for your GEO strategy.
ChatGPT now accounts for 92.4% of all trackable referral traffic from standalone AI platforms, up from roughly 84% seven months ago, according to a new analysis of 6.77 million AI-driven sessions. Claude has overtaken Perplexity as the fastest-growing challenger, while Microsoft Copilot has collapsed by 96% from its peak. The takeaway for anyone tracking AI search visibility is blunt: a GEO strategy that isn't built around ChatGPT first is optimising for the wrong platform.
What changed?
Previsible, a GEO agency, published the third edition of its AI Traffic Study on 6 July 2026, expanding its dataset to 6.77 million LLM-driven sessions across 166 websites spanning SaaS, e-commerce, finance, legal, health, insurance, education and publishing, according to Search Engine Land.
The headline finding is consolidation, not diversification. ChatGPT commands 92.4% of trackable LLM referral traffic, growing 12.8 times over 19 months with no sign of slowing, per Previsible's report. That is a sharp jump from the same study's previous edition, when ChatGPT's share stood at about 84% in December 2025, with Perplexity at 8.9%, Gemini at 4.5%, Copilot at 2.1% and Claude at just 0.6%.
The most interesting movement is happening below the top spot. Claude grew 64 times, from 133 sessions in November 2024 to 8,528 in May 2026, and overtook Perplexity in March 2026 for the first time, Previsible reports. It stayed ahead after that, with growth accelerating fourfold in two months as Claude's agentic tools and enterprise integrations gained adoption.
Perplexity's trajectory has gone the other way. Monthly sessions fell by 61% since March 2025, partly because Perplexity has been keeping users inside its own browser and agent tools rather than sending them onward, according to the report cited by Marketing Tech News. Copilot fared worse still, dropping 96% from a peak of 8,651 monthly sessions in August 2025 to just 339 in May 2026.
Gemini has quietly become the steady number two, growing 3.2 times to become the second most visible model behind ChatGPT, per Business Wire's coverage of the report.
The study also flags a November 2025 wobble worth knowing about if traffic dipped unexpectedly around then: sessions fell 50% in one month, driven almost entirely by ChatGPT referrals dropping from 448,412 to 213,345 while other platforms held steady. Previsible attributes this to a model-level change, noting that modest product tweaks have previously halved some sites' ChatGPT traffic when the model shifted to favouring sources like Wikipedia and Reddit.
Crucially, the study is upfront about what it doesn't measure. It covers standalone LLM referral traffic only. AI discovery inside Google's own results, including AI Overviews, almost certainly drives more AI-related traffic than all standalone platforms combined, but it operates on a different measurement paradigm and is excluded here, as ALM Corp's summary of the report notes.
Why does this matter for your business?
If you've been spreading AI visibility efforts evenly across ChatGPT, Perplexity, Gemini and Copilot, this data suggests you're misallocating effort. ChatGPT is the only standalone LLM sending meaningful referral volume at scale, so optimising for "AI visibility" in the abstract, without prioritising ChatGPT specifically, means chasing a target that barely exists in the traffic data.
The study also reveals a landing-page problem many brands haven't noticed. For SaaS sites, search results pages capture 34.6% of ChatGPT referrals. ChatGPT trusts the domain but can't identify the right page, so it routes the user to an internal search result instead. Sites with strong search UX convert these sessions; sites without one lose high-intent visitors at the front door.
Publishers face a particularly stark imbalance. News pages capture 54% of LLM referrals, yet against more than 120 million organic sessions, penetration is just 0.08%, according to Previsible's report. Publishers produce much of the content LLMs train on and cite, but capture almost none of the resulting traffic.
There's also a behavioural split worth building strategy around. ChatGPT and Gemini behave as search-pattern models, leaning on domain trust with page-level uncertainty. Perplexity and Claude behave as content-selection models, picking specific pages and over-indexing on long-form content. If editorial content drives your qualified traffic, Perplexity and Claude matter disproportionately to their current market share.
E-commerce brands should note that product pages are the primary AI landing surface, with traffic almost entirely from ChatGPT and users arriving with purchase intent already formed. Structured, comparable product data is becoming an AI discoverability requirement, not just a conversion optimisation nice-to-have.
What should you do now?
Previsible's report author frames the priority order clearly: build the foundation in search first, becoming a source Google's AI results want to cite through solid site architecture and content signals, then focus on winning ChatGPT as the leading standalone surface, according to Search Engine Land.
Concretely, the report recommends building citation-worthy evidence, improving authority across trusted third-party sources, making websites easier for AI systems to read and extract, optimising for answer journeys, and measuring business impact rather than broad visibility alone.
Three practical actions follow from the data. First, audit your internal site search: if ChatGPT is sending a large share of its referrals to search results pages rather than specific answers, a weak search experience is quietly costing you conversions. Second, don't ignore Claude just because its raw numbers are still small; its 64-times growth and pull with developers and technical buyers make it worth monitoring, particularly for B2B and professional-services brands.
Third, treat Google's AI Overviews and AI Mode as the bigger prize, even though this study excludes them. That surface likely dwarfs standalone chatbot referrals combined, so ChatGPT optimisation should sit alongside, not instead of, classic search visibility work. If you haven't checked where your brand currently stands across these different AI surfaces, running a free audit with a tool like Sited's at https://sited.online is a fast way to spot the gaps before rebuilding your GEO priorities.
Frequently asked questions
Does this study include Google's AI Overviews?
No. The report measures referral traffic from standalone chatbot platforms only. AI discovery inside Google's own results, including AI Overviews, almost certainly represents a larger volume of AI-driven traffic than all standalone LLM platforms combined, but it doesn't generate trackable referral sessions in the same way.
Why did Perplexity's traffic fall so much?
The study attributes part of the decline to product changes rather than a pure loss of usage. Monthly sessions fell 61% since March 2025, partly because Perplexity has been keeping users inside its own browser and agent tools instead of sending them to external sites.
Is Claude worth optimising for even though its overall share is small?
Yes, according to the data. Claude grew 64 times over the tracked period and overtook Perplexity in March 2026, showing particular strength among developers, technical buyers and professional services. Its growth rate outpaces every platform except ChatGPT.
What should e-commerce sites prioritise based on this data?
Structured product information appears critical. Product pages are the primary landing surface, traffic is almost entirely from ChatGPT, and users typically arrive with purchase intent already formed. Structured, comparable product data is becoming an AI discoverability requirement, not just a conversion optimisation play.
How big was the underlying dataset?
Previsible analysed 6.77 million AI-driven sessions across 166 websites spanning SaaS, e-commerce, finance, legal, health, insurance, education and publishing, tracked from November 2024 to May 2026.
Sources
Google is piloting an AI Performance report in Merchant Centre, showing ecommerce brands their share of voice in AI Mode, AI Overviews and Gemini.
Google has begun rolling out an AI performance insights report inside Merchant Centre, giving retailers their first native look at how products are discovered across AI Mode, AI Overviews and the Gemini app. It includes a "share of voice" metric benchmarked against similar brands. For any brand selling online, this is the clearest signal yet that AI visibility is becoming a measurable, reportable line item rather than a guess.
What changed?
On 14 July 2026, search marketers spotted the pilot live inside real Merchant Centre accounts. Search Engine Roundtable's Barry Schwartz confirmed that Google Merchant Centre is rolling out a beta AI Performance report that shares how well products are performing within AI Mode and AI Overviews, with Google describing it as revealing performance specifically on those two surfaces.
Google first announced the feature back in May, but access only started reaching real accounts this week. SEO consultant Brodie Clark, who gained access for a client account, wrote on X that "Google Merchant Centre is (surprisingly) the first Google product to give query data for AI Overviews and AI Mode," adding that the report can be found under Analytics > Products > AI Performance and is currently available to a limited number of accounts in the US.
According to Google's own help documentation, the tool exists because consumers are increasingly using conversational surfaces to shop, so Google is introducing AI performance insights in Merchant Centre to help brands understand how their products are being discovered.
The report has four components. As Google's business blog explains, brands can compare their share of voice, meaning how often they're linked in relevant results, against shopper demand, alongside a breakdown of journey stages such as discovery, evaluation or purchase, the product terms shoppers search for, and structured attributes such as dimensions, weight, materials and colours.
Crucially, share of voice is relative rather than absolute. Semrush's analysis notes that the metric shows how often a brand surfaces in AI-driven experiences across Search and Gemini, benchmarked against similar brands, and is a relative figure, not a traffic or ranking number.
The rollout remains narrow. Google's help centre confirms that AI performance insights is currently in a pilot with a limited number of Merchant Centre accounts in the US, with plans to expand to Australia, Canada, India and New Zealand in the coming months.
Why does this matter for your business?
For years, brands have had almost no direct way to check whether their products even appear inside AI Overviews or AI Mode, let alone how they compare to competitors. This report changes that, at least for Google's own surfaces.
Industry analysis frames this as a genuine milestone. Semrush states plainly that this is the first time Google has offered native reporting on AI search visibility metrics for ecommerce, signalling that AI performance is now its own measurement category.
That matters because until now, most AI visibility measurement for ecommerce has relied on third-party tools guessing at Google's black box. Now Google is handing over some of its own numbers, even if the picture is partial. PPC Land's breakdown notes that benchmarking against similar brands has existed in paid search tools before, such as auction insights in Google Ads, but its application to AI surface performance is new. That shifts the question from "are my products appearing?" to "how often do my products appear relative to competing brands in AI conversations?"
There's a catch worth flagging for anyone budgeting around this data. The report only counts unpaid visibility. Google's help documentation is explicit that current data is strictly limited to organic AI traffic, meaning cost-free listings, and paid ads traffic isn't included.
It's also worth being realistic about how actionable the data is right now. Brodie Clark, who has hands-on access, cautioned that "in its current form, similar to the recent rollout of AI reporting in Search Console, there isn't a great deal of actionability behind the data, though it is good to see at least some form of query data being included."
This launch also sits inside a wider pattern. Google isn't acting alone, and it isn't even first. Semrush points out that Bing rolled out its own AI performance report in Webmaster Tools earlier this year. AI visibility reporting is quickly becoming standard across major search platforms, not a one-off experiment.
What should you do now?
If you sell through Google Merchant Centre, check your Analytics tab now. Google's guidance is to log in to Merchant Centre, navigate to Analytics, select Products, then select the AI performance tab to load the dashboard. Most accounts won't see it yet, since the pilot is limited to a small group of US advertisers, but it's worth checking regularly as the rollout expands.
While you wait for access, focus on the fundamentals the report is built to measure: clean, complete product attributes, clear specifications and terms that match how shoppers actually phrase queries during discovery, evaluation and purchase. Google's own framing notes that the goal is to give brands a view into their performance on AI surfaces and how to optimise product data to improve results there.
Don't treat this as your whole AI visibility picture, either. It only covers Google's surfaces, and only organic listings. If you want to understand how your brand shows up across ChatGPT, Perplexity, Gemini and Google's AI features together, running a free audit with a tool built specifically for cross-platform AI visibility, such as Sited's free audit, can give you a fuller picture than any single platform's native reporting offers on its own.
Finally, treat this launch as a preview of where measurement is heading generally. Share of voice, funnel-stage visibility and attribute completeness are likely to become standard metrics across AI search reporting tools over the next year, not just inside Merchant Centre.
Frequently asked questions
What is Google's new AI performance insights report?
It's a beta reporting feature inside Google Merchant Centre that shows retailers how their products are discovered across AI Mode, AI Overviews in Search, and the Gemini app, including a share of voice metric benchmarked against similar brands.
Who can access it right now?
Access is limited to a pilot group of accounts in the United States. Google has said it plans to expand the rollout to Australia, Canada, India and New Zealand in the coming months, with no exact dates confirmed.
Does the report include paid advertising data?
No. The data is limited to organic, cost-free listing visibility within AI experiences; paid ads traffic is not part of this report.
How is this different from existing Google Search Console reporting?
Search Console has recently added some AI reporting, but Merchant Centre's new tool is the first to include query-level data, such as query type and frequency, specifically for AI Mode and AI Overviews, giving ecommerce sellers more granular insight than currently exists in GSC.
Should I stop using third-party AI visibility tools now?
No. Google's report only covers its own AI surfaces and organic listings, so it won't show you how your brand appears in ChatGPT, Perplexity or other AI platforms. A broader audit tool is still needed for the full picture.
Sources
Google Merchant Center AI Performance Insights - Search Engine Roundtable
Daily Search Forum Recap: July 14, 2026 - Search Engine Roundtable
Insights for AI-powered shopping experiences coming soon - Google Merchant Center Help
Google's new Merchant Center report tracks your brand in AI Mode - PPC Land
Google to add AI performance report to Merchant Center - Semrush
Google Adds AI Performance Insights to Merchant Center - The Keyword
Google's John Mueller says Cloudflare's Content Signals robots.txt tag isn't honoured by any crawler or LLM. Here's what actually works instead.
Google's John Mueller has confirmed that Cloudflare's Content Signals directive, a robots.txt addition meant to let sites state preferences on AI training and AI-generated answers, has no effect on any crawler or large language model. If your site relies on this tag to keep content out of AI training or AI answers, it isn't doing that job, and you need network-level blocking instead.
What changed?
Mueller confirmed on Reddit that Cloudflare's Content Signals field is not read or respected by any crawler or LLM, according to Search Engine Roundtable, which first reported the exchange on 6 July 2026. He said the directive, "invented by Cloudflare last year", has "no effects whatsoever for any crawler or LLM."
Mueller went further: "none of the crawlers / llms use the 'content-signal' robots.txt directives. It was made up by a CDN... Using it just adds bloat & future maintenance to your robots.txt file," he wrote, per the same report.
Cloudflare launched Content Signals in late 2025 so site owners could state, in machine-readable form, whether crawlers could use their content for search indexing, AI training, or as live input to AI-generated answers. It sits inside the same robots.txt file that has governed basic crawl access for decades, but it's a preference, not a technical block. It does not depend on any bot actually reading the line, it depends on that bot choosing to comply.
Mueller's comments were prompted by a question about whether Content Signals and llms.txt files help disambiguate a person as a distinct entity in search. His answer covered both: Google does not recognise or use llms.txt or llms-author.txt files, and no other known crawlers or LLMs are confirmed to use them either, aside from some SEO tools, according to the same Search Engine Roundtable report.
This follows a related finding just days earlier: Google also confirmed that llms.txt files and content "chunking" don't influence AI search visibility, according to Adapt Worldwide's roundup of June-July 2026 AI search news. Several of the community's favourite technical fixes are turning out to be inert.
Why does this matter for your business?
Many publishers and marketing teams spent real engineering time over the past year adding Content Signals to robots.txt, believing it gave them licensing leverage over how AI systems reuse their content. That belief was reasonable given how the tag was marketed, but Mueller's comments confirm the leverage doesn't currently exist in practice.
The scale involved is significant. Cloudflare sits in front of roughly 21.3% of all websites as of January 2026, per usage data cited in the reporting. A meaningful share of the web has this directive live right now, under the impression it does something.
It's not only about training data. Content Signals cover three categories: search (building a search index), ai-input (feeding content into AI models for real-time answers), and ai-train (training or fine-tuning AI models), as Search Engine Roundtable notes. If none of these are honoured, brands have no reliable way to say "index me for search, but keep my pages out of an AI Overview or a ChatGPT answer" using robots.txt alone.
There's a deadline worth tracking too. Cloudflare has set 15 September 2026 as the date it will apply new defaults across the three categories: for new domains onboarding to Cloudflare, Training and Agent access will be blocked by default on pages that display ads, while Search stays allowed by default. That default blocking happens at the network edge through Cloudflare's bot management, a separate and enforceable mechanism from the robots.txt signal Mueller was dismissing.
For brands managing AI visibility, the takeaway is uncomfortable: the tools that feel like they give you control over how AI systems use your content mostly don't. Google's own AI features already ignore the signal, and no other major crawler operator has confirmed support for it either.
What should you do now?
First, don't rip anything out. Mueller described the directive as inert, not harmful, so teams that already added Content Signals entries don't need to remove them.
Second, separate preference from enforcement. A line in robots.txt is a polite request that depends on voluntary compliance, which Mueller says is currently observed by no one. Edge-level blocking is enforceable because it sits between the bot and the content. If you genuinely want to stop specific AI crawlers, you need Cloudflare's WAF rules, bot management, or equivalent server-side controls, not a robots.txt comment.
Third, treat this as a reminder that visibility, not restriction, is where most brands should focus their GEO and AEO effort. If your goal is to be cited and recommended by AI systems rather than excluded from them, the controls that matter are original research, clear entity signals, first-hand expertise, and pages that are genuinely easy to extract and quote. This is a good moment to check how your brand actually shows up across ChatGPT, Google AI Overviews, Gemini and Perplexity rather than assuming a robots.txt tweak has settled the question either way, and Sited's free audit at sited.online can give you that baseline quickly.
Finally, watch the standards process rather than betting on vendor-specific tags. The IETF's AIPREF group is working on standardising AI usage preferences, so it's fair to treat Cloudflare's approach as pragmatic, interim signalling. Until a standard has buy-in from OpenAI, Google, Anthropic and other major crawler operators, any single vendor's directive risks the same fate as Content Signals.
Frequently asked questions
Does this mean robots.txt is useless for controlling AI crawlers?
No. Standard robots.txt directives like Disallow still work for crawlers that respect the Robots Exclusion Protocol, including many AI crawlers. It's specifically the newer Content Signals field, which expresses usage preferences rather than access rules, that Google says goes unread.
Should I remove Cloudflare's Content Signals from my site?
There's no need. Mueller described the tag as inert rather than damaging, so leaving it in place costs you nothing beyond a slightly longer robots.txt file. Just don't rely on it as your only protection.
What should I use instead if I want to block AI training on my content?
Use enforceable controls such as Cloudflare's WAF rules, Bot Management, or AI Crawl Control, which operate at the network edge and can block or challenge specific crawlers by user agent, rather than depending on a crawler choosing to honour a preference.
Is llms.txt affected by this too?
Yes. Google does not recognise or use llms.txt or llms-author.txt files, and no other known crawlers or LLMs are confirmed to use them either, aside from some SEO tools. Both llms.txt and Content Signals fall into the same category of well-intentioned but currently unsupported technical fixes.
What does this mean for AI search visibility strategy generally?
It reinforces that visibility comes from content quality, structure and trust signals that AI systems can actually extract and cite, not from technical directives asking to be excluded or included. Brands are better served focusing on being genuinely citable than on trying to game inclusion through robots.txt add-ons.
Sources
Arc XP's Ask The News lets publishers embed their own AI answer layer. Here's what it means for brand visibility across AI search and answer engines.
The Washington Post's technology arm, Arc XP, has launched "Ask The News", a tool that lets publishers embed their own AI-powered answer engine directly on their sites. It exists because only 4% of AI chatbot users click through to original news sources, compared with 19% from search, according to the Reuters Institute Digital News Report 2026. For any brand relying on being found and cited by AI, this is a signal that the entire discovery layer is being renegotiated, not just optimised.
What changed?
Arc XP, the media platform built by The Washington Post, announced on 13 July 2026 that it is launching Ask The News, an AI-powered answer layer that publishers embed directly on their own digital properties. Millions of readers who once visited a news site now get their answers from ChatGPT, Perplexity or Google's AI Overviews without the publisher seeing a pageview, collecting a data point, or earning a cent, and Arc XP built this product to try to reverse that.
The mechanics matter. When a reader has a question, Ask The News answers it using the publisher's own reporting, with attribution and editorial guardrails against the hallucination risk that defines generic AI tools. The publisher owns the interaction, the intent data it generates, and any business opportunity that follows. That's the crux of it: this isn't just another chatbot, it's an attempt to reclaim ownership of the "answer moment" that ChatGPT, Perplexity and Google currently dominate.
Arc XP's Vice President of Content Intelligence, Joey Marburger, framed the shift bluntly: "The industry has spent years worrying about Google and social platforms. The next threat is quieter and faster: readers learning to ask AI instead of visiting a news site. Ask The News puts publishers back in that conversation." He drew a sharper line between the three models competing for the same query: "Search shows links. Generic AI gives answers from the open web. Ask The News answers with a publisher's journalism, rules, and business model."
The product also has a built-in refusal mechanism: it automatically declines to answer when there isn't enough source reporting to do so confidently, a guardrail that deliberately contrasts with the hallucination risk in open-web AI answers.
The rollout isn't limited to Washington Post-owned properties. Ask The News is available to Arc XP publisher partners, with a standalone deployment for non-Arc XP publishers via a JavaScript embed. Arc XP's reach is significant: the platform is used by more than 2,500 sites globally, according to Arc XP's own materials. One early adopter, Martin Kautz, Head of Media Technologies at RND, RedaktionsNetzwerk Deutschland, said the tool "has the potential to change how audiences experience our journalism... It gives readers a more intuitive way to explore our reporting while helping us preserve control of the reader relationship and better understand what people want to know."
Arc XP isn't stopping at answers, either. It plans to expand the product into a broader intelligence platform, with future capabilities including personalised briefings, saved conversations, topic tracking, editorial intelligence dashboards, and proactive reader experiences.
Why does this matter for your business?
If you run a brand, this news isn't really about journalism, it's about where the next wave of "answer infrastructure" gets built, and who controls it. Right now, the dominant AI systems, ChatGPT, Perplexity and Google AI Overviews, decide which sources get cited, summarised or ignored. Ask The News is the first serious attempt by a major publisher network to build a parallel, first-party answer layer that sits outside that system entirely.
This has two direct implications for AI search visibility strategy.
First, the "answer moment" is becoming contested territory, not a fixed pipeline. If large publisher networks succeed in keeping readers on their own properties by answering questions directly, rather than losing them to ChatGPT or Google, some of the traffic and attention that brands currently chase through third-party AI citations may migrate to these first-party answer layers instead. Being cited well inside a publisher's own AI answer tool could become a new visibility channel worth tracking, separate from being cited in ChatGPT or AI Overviews.
Second, it confirms just how severe the click-through collapse already is. The 4% versus 19% click-through gap isn't a hypothetical risk, it's measured behaviour now driving product decisions at one of the world's largest media technology providers. If a brand's visibility strategy still assumes users will click through from an AI answer to verify a source, that assumption is increasingly shaky, whether the AI in question is generic or a trusted publisher's own tool.
There's also a licensing and monetisation angle worth watching. Arc XP has already been building infrastructure to control and charge for AI bot access to publisher content, including a March 2026 integration with TollBit, which offers AI agents and bots an easy and compliant way to compensate websites directly for content access. Ask The News extends that same instinct from bot access to reader-facing answers. Publishers are no longer treating AI purely as a threat to fend off with blocking rules, they're building competing answer products.
What should you do now?
Brands that rely on being cited inside third-party AI tools should start treating publisher-run answer layers as a new, emerging surface, not a footnote. A mention or citation inside a major outlet's own AI tool could carry weight with a reader who trusts that publisher more than they trust a generic chatbot response.
Practically, this means:
Keep monitoring where your brand shows up, across all AI surfaces, not just the big three. As more publishers launch first-party answer tools, the number of places your brand can be cited, or missed, keeps growing.
Make sure your public-facing content is unambiguous and well-sourced. Any answer engine, whether it's ChatGPT or a publisher's in-house tool, favours content that's easy to extract and attribute cleanly.
Don't assume click-through will recover. Build your content and PR strategy around being useful within the answer itself, since a visit to your website is no longer guaranteed, even from trusted, brand-safe sources.
If you're not sure how your brand currently appears across AI-driven answers, a quick way to get a baseline is by running your own domain through Sited's free audit, which shows how AI systems are currently representing your brand right now.
Frequently asked questions
What is Ask The News?
It's a product from Arc XP, The Washington Post's media technology arm, that lets publishers embed an AI-powered answer tool on their own sites. When a reader has a question, Ask The News answers it using the publisher's own reporting, with attribution and editorial guardrails against the hallucination risk common in generic AI tools.
Why are publishers building their own AI answer tools instead of just blocking AI crawlers?
Blocking alone doesn't solve the underlying problem: readers are increasingly getting answers without ever visiting a publisher's site. Only 4% of AI chatbot users click through to news sources, versus 19% from search, according to the Reuters Institute Digital News Report 2026. Publishers are responding by trying to own the answer experience themselves rather than ceding it entirely to third-party AI tools.
Does this affect brands outside the news industry?
Indirectly, yes. As more publishers, review sites and industry authorities build their own first-party AI answer layers, the places where your brand can be surfaced, cited or ignored will multiply. Tracking visibility purely inside ChatGPT, Perplexity and Google AI Overviews may no longer give you the full picture.
Is Ask The News available to publishers outside Arc XP's existing network?
Yes. Ask The News is available to Arc XP publisher partners, with a standalone deployment for non-Arc XP publishers via a JavaScript embed.
How is this different from a publisher simply being cited in ChatGPT or Google AI Overviews?
Ask The News is built and controlled entirely by the publisher, using only their own reporting, whereas citations in ChatGPT or AI Overviews are decided by an external company's algorithm using open-web sources. As Arc XP's Joey Marburger put it, this is about building a durable reader habit: when someone has a question, they go to a trusted news organisation first, rather than a generic AI tool.
Sources
75% of PR pros say GEO matters, but 29% have no owner and 39% don't measure it. Here's what the ownership gap means for your brand and what to do next.
Most PR teams agree that AI search visibility matters, but most haven't decided who is responsible for it. A new Muck Rack study finds 75% of PR professionals see Generative Engine Optimisation (GEO) as important to their strategy, yet 29% say no one at their organisation owns it and 39% aren't measuring it at all. That gap between recognising the problem and managing it is now one of the biggest blind spots in marketing.
What did the new study find?
Muck Rack's latest State of PR report surveyed 1,115 PR professionals between 14 May and 12 June 2026, with 971 qualified responses making the final cut, according to Muck Rack's press release. Nearly three-quarters of PR professionals say GEO is at least somewhat important to their communications strategy, yet 29% say no one at their organisation owns it.
The measurement gap is just as stark. 39% of professionals say they aren't measuring GEO success at all, despite 45% citing media measurement as a major part of their work. That mismatch matters because you cannot improve what you refuse to track, and right now nearly four in ten teams have chosen not to look.
With 61% expecting AI and automation to grow over the next five years, AI search visibility has quickly become a strategic priority without a clear plan behind it. Muck Rack cofounder and CEO Greg Galant put it plainly: "Most PR professionals recognise that AI visibility matters, but many companies still haven't decided who owns it or how to measure it. This is a critical moment where better data and better tools can make all the difference in how we approach this shift in the PR workflow."
Why does this ownership gap matter for your business?
If nobody owns AI visibility inside your organisation, it falls between departments, and that is expensive. Marketing assumes it's a search engine optimisation job. SEO assumes it's a PR job. Communications assumes it's a technical, website-level job. Everyone assumes someone else is watching what ChatGPT, Perplexity, Gemini and Google's AI Overviews say about the brand, and often no one is.
This is not a small channel to ignore. Semrush's expanded 2026 AI Visibility Index, which analysed 126 million US AI search prompts between January and April 2026, found a clear performance split. Among organisations that fully integrate SEO and AI visibility into a unified workflow, 81% reported increased traffic or leads from AI platforms, compared with only 36% among organisations managing the two areas separately.
That 45-point gap is the practical cost of the ownership problem Muck Rack has just documented. Companies that treat AI visibility as somebody's side project are leaving growth on the table compared with those that have built it into how they already run SEO and content.
The measurement side is just as concerning. Semrush's research adds another layer: 45% of marketing leaders cannot accurately measure their brand visibility within AI-generated answers, while only 9% have the tools to track all relevant metrics across platforms. Put the two studies together and a pattern emerges: most organisations neither own AI visibility internally nor have the tools to measure it externally. That's a double blind spot, not a single one.
It also matters because AI-driven discovery is growing fast, not slowly. Adobe data cited in Semrush's report shows AI traffic to US retail sites surged 1,324% between October 2024 and May 2026, while in the travel sector AI traffic rose 2,215% over the same period. Every month that ownership stays unresolved, more prospective customers are forming an opinion of your brand through a channel nobody at your company is actively managing.
What's driving the confusion?
Part of the problem is that AI visibility doesn't fit neatly into any one team's existing remit. It touches technical SEO (can AI crawlers reach and parse your content), PR and communications (are you being mentioned and cited by trusted third parties), and product marketing (does the language AI systems use to describe you match what you actually offer).
Muck Rack's study also found that 51% of PR professionals say thought leadership has become increasingly vital to their jobs, up 5% from last year, with the rising emphasis on thought leadership tracking closely alongside LinkedIn's growing importance. That's a clue about where GEO ownership is naturally drifting: towards communications teams building authority and citations, even without a formal mandate.
Meanwhile, the underlying platforms keep shifting. Analysis of citation patterns across major AI platforms shows that only 11% of domains are cited by both ChatGPT and Perplexity, because each platform operates on fundamentally different citation logic. A brand that appears frequently in one AI tool can be invisible in another, which makes "who is responsible" an even harder question to answer with a single team or a single dashboard.
What should you do now?
Start by naming an owner, even an interim one. It doesn't need to be a new hire. It can be an existing SEO lead, a comms director or a marketing operations manager, but someone needs explicit responsibility for tracking how AI systems describe, cite and recommend your brand.
Second, put a measurement baseline in place. You don't need every metric on day one. Start by checking whether your brand appears at all when AI tools answer questions relevant to your category, then track that consistently over time. Checking your own AI visibility with a free audit, such as the one Sited offers at sited.online, is a straightforward way to get that baseline without building a measurement system from scratch.
Third, treat AI visibility as a shared workflow rather than a handoff. Semrush's research is explicit that the biggest gains come from integration, not separation. That means SEO, PR and content teams need a shared view of what AI systems are saying, not three separate reports that never get compared.
Finally, revisit ownership regularly. Because AI platforms and their citation behaviour shift quickly, a static "job done" mentality won't hold. The Muck Rack findings suggest most organisations are only at the start of working out how GEO fits their structure, so expect this to be revisited as budgets and priorities firm up over the rest of 2026.
Frequently asked questions
What is GEO and how is it different from SEO?
GEO, or Generative Engine Optimisation, is the practice of improving how often and how favourably your brand is mentioned, cited or recommended by AI systems such as ChatGPT, Gemini and Google AI Overviews. Traditional SEO focuses on ranking in search results, while GEO focuses on being the source AI tools pull from and quote.
Why don't more companies have someone owning AI visibility?
According to Muck Rack's research, 75% of PR professionals see GEO as at least somewhat important, yet 29% say no one at their organisation owns it, largely because the discipline cuts across SEO, PR and marketing without fitting cleanly into any one existing role.
How do I know if my brand is currently visible in AI answers?
The simplest approach is to run a set of realistic buyer questions through ChatGPT, Perplexity, Gemini and Google AI Overviews and check whether your brand appears, and how it's described. Free tools such as Sited's audit at sited.online can also give you a quick read on your current standing.
Does integrating SEO and AI visibility actually make a measurable difference?
Yes. Semrush's 2026 research found that 81% of organisations that fully integrate SEO and AI visibility into a unified workflow reported increased traffic or leads from AI platforms, compared with only 36% among those managing the two separately.
Is AI search traffic actually worth chasing compared with traditional search?
Adobe data cited in Semrush's report shows AI traffic to US retail sites surged 1,324% between October 2024 and May 2026, while travel sector AI traffic rose 2,215% over the same period, suggesting the channel is growing fast enough that most businesses cannot afford to leave it unmanaged.
Sources
ChatGPT routes queries through four hidden retrieval systems, new research shows, which explains why the same question can produce different brand citations.
ChatGPT does not run one search system behind its answers. Independent analyses of its raw network traffic found four hidden retrieval pipelines, internally labelled Labrador, Bright, Oxylabs and SERP, each pulling from a different pool of sources. When ChatGPT switches between them, the citations it shows can change even though your content hasn't.
What changed?
Two SEO researchers inspected the raw network traffic ChatGPT sends behind its answers, rather than relying on the polished citation cards users see. According to Search Engine Land, they found internal source-selection labels sitting behind the answer: Labrador, Bright, Oxylabs and SERP.
Chris Green ran the larger test. He tested 1,000 prompts up to 10 times each and captured 9,946 completed search runs. Most prompts stayed on one retrieval source: Labrador accounted for 88.1% of primary search sources, Bright for 9.9%, Oxylabs for 1.7%, and SERP for 0.3%. But 11.6% of prompts changed primary search source across repeated runs.
The consequence of that switching is the real story. URL overlap dropped from 0.273 to 0.149 when the search source changed, and domain overlap fell from 0.265 to 0.155, roughly 45% lower URL overlap and 42% lower domain overlap. In plain terms, when ChatGPT switches pipelines, nearly half the sources it would have cited disappear from the answer, replaced by different ones.
Suganthan Mohanadasan, co-founder of Snippet Digital, ran a separate analysis of his own account's traffic. Writing on Search Engine Journal, he found the labels point to Bright Data and Oxylabs, two commercial scraping firms that are direct rivals, handling ChatGPT's open web fetching. Bright did the bulk of the fetching, especially on commercial, shopping, finance and weather queries, while Oxylabs skewed regional and local, Labrador stayed on news and reference, and SERP mostly turned up on news.
He also found that not every query reaches these pipelines at all. ChatGPT classifies some queries with a field called turn_use_case before deciding whether to search, meaning some prompts are filed as text and skip web search entirely, even when they sound current. That routing decision determines which pages ever get a chance to be read, let alone cited.
Mohanadasan's traffic also separated three outcomes that are often lumped together in GEO advice: fetched, cited and mentioned. A page can be fetched into ChatGPT's context without being shown to users. It can be cited as the source behind a specific sentence. Or a brand can simply be mentioned without being the source of the claim. In his sample, Reddit and YouTube were both fetched often, but Reddit was cited and YouTube was not, which he attributed to text availability: Reddit threads expose text, while YouTube search results often provide metadata rather than transcripts.
Both researchers are careful about the limits of their samples. As Search Engine Journal notes, this is one person, one logged-in Pro account, a few days of traffic, not a population study, logging around 1,240 source records across a few dozen searches. The structural finding, that these pipelines exist and behave differently, holds up. The exact percentages are a snapshot, not a fixed rule.
Why does this matter for your business?
If you've tracked your brand's ChatGPT citations and watched them swing week to week with no obvious cause, this is likely part of the explanation. You're not being tracked by one ChatGPT. You're being tracked by several retrieval systems that share an interface but pull from different pools of sources.
This has direct implications for pricing pages, product specs and comparison content. Mohanadasan's traffic showed vendor pages were cited for their own facts, such as prices and specs, while third-party pages were more likely to support broader recommendation claims. That splits your visibility strategy into two jobs: making your own site machine-readable for facts, and making sure independent sites carry the opinion-based claims that win recommendations.
Readability matters more than most brands assume. Both analyses showed ChatGPT's source selection depended partly on what it could actually retrieve and read. Mohanadasan found cases where ChatGPT appeared to prefer official pricing pages, then fell back to third-party sources when prices were hidden behind JavaScript or otherwise hard to parse. A price buried in a JavaScript widget isn't just a UX flaw. It can be the reason a competitor's third-party review gets cited instead of your own product page.
This lands at a moment when ChatGPT's referral traffic is climbing fast. Weekly analysis from Anicca reports that SE Ranking analysed traffic data from 101,574 websites across 250 countries and territories and found ChatGPT's share of referral traffic rose from 0.23% in April to 0.32% in May, its highest level on record and 8% above the previous peak from October 2025. More traffic is riding on a citation system that is proving less predictable than most brands assumed.
What should you do now?
Stop treating "ranking in ChatGPT" as a single target. Multiple pipelines with different source pools mean consistency, not a one-off fix, is the real challenge.
Prioritise plain, crawlable HTML over content locked behind scripts. The research is consistent here: plain HTML, crawlable facts, clear pricing and specs, strong third-party coverage, and text-heavy pages all became more important once source selection depended on retrieval and readability.
Build presence on the third-party sites that actually get cited for opinions and recommendations, not just your own domain. Track outcomes separately too, because being fetched, mentioned and cited are three different things, and only the last one drives the specific sentence a customer reads.
Because AI Overviews, ChatGPT and other assistants change their retrieval behaviour from week to week, it's worth checking where your brand currently stands with a free audit at Sited rather than assuming last month's snapshot still holds.
Finally, treat any single vendor's percentages as directional rather than definitive. As the researchers themselves caution, the numbers move even when the structure holds, so re-check your visibility regularly rather than optimising once and walking away.
Frequently asked questions
What are Labrador, Bright, Oxylabs and SERP?
They are internal labels found in ChatGPT's network traffic marking which retrieval provider fetched a given web result. A result_source field is attached to web results, with Labrador covering established publishers and reference sites, Bright tied to Bright Data, Oxylabs tied to Oxylabs, and SERP an open-web baseline that appeared mostly in news-style results.
Does this mean my SEO work is wasted?
No, but it means SEO alone isn't sufficient. Readable, well-structured pages still need to win the fetch-and-read step before they can even be considered for citation, and third-party coverage remains critical for recommendation-type claims.
Why do my ChatGPT citations change from week to week?
Partly because your query may be routed to a different retrieval pipeline than last time, each with a different pool of sources. Green's research found that URL overlap dropped from 0.273 to 0.149 when the search source changed, and domain overlap fell from 0.265 to 0.155 between runs.
Is this the same for every ChatGPT user?
Not necessarily. Both studies were based on limited, individual-account samples, and the researchers themselves warn against treating specific percentages as universal, since the numbers come from a relatively small set of queries on single accounts.
Should brands try to target a specific pipeline like Bright or Oxylabs?
You can't choose which pipeline handles a query, so the more useful approach is making your content easy for any scraper to read: plain HTML, visible pricing, clear specs, and strong coverage on third-party sites that these systems already trust.
Sources
Google's updated AI search guide says llms.txt, chunking and AI schema don't affect Google visibility. Here's what actually matters instead.
Google has updated its official guide on optimising for generative AI search, stating plainly that llms.txt files, content "chunking" and bespoke AI schema do nothing for visibility in AI Overviews or AI Mode. The refreshed guidance says AEO and GEO aren't separate disciplines from SEO, just new labels for the same work. If your business is paying for these tactics, it's a clear signal to redirect that budget.
What changed?
Google's developer documentation page, "Optimising your website for generative AI features on Google Search," was first published on 15 May 2026 and has now been refreshed with more detail, according to Search Engine Journal. The guidance is explicit: for Google Search, you can ignore tactics like "chunking" content, creating unnecessary AI text files such as llms.txt, or pursuing inauthentic mentions.
The page sits inside a "Mythbusting generative AI search" section that names specific tactics directly, something Google's guide has rarely done before. Google states that, from its perspective, optimising for generative AI search is optimising for the search experience, and is therefore still SEO. This echoes positions Google employees have taken at conferences, but now it's in published documentation as an official reference to cite, as Semrush notes.
On llms.txt specifically, Google says Search doesn't use machine-readable files, AI text files, markup or Markdown to determine visibility, and adding one will neither harm nor help a site's rankings, since Google Search simply ignores them. On chunking, the guide is equally direct: there's no need to break content into tiny pieces, because Google's systems can already understand nuance across multiple topics on a page and surface the relevant part to users.
Independent data broadly backs this up. A 90-day log study by OtterlyAI, cited by CXL, found that only 0.1% of AI crawler requests ever touched /llms.txt, with most crawlers ignoring the file entirely. Separately, Ahrefs tracked 1,885 pages that added JSON-LD schema against a matched control group and found no citation lift across AI Overviews, AI Mode or ChatGPT, per the same CXL analysis.
Not everything has landed cleanly. Google's warning against "inauthentic" mentions has drawn pushback from well-known SEO voices, with Lily Ray calling it "classic Googlespeak" on X, arguing the language is vague enough to mean whatever Google needs it to mean later, as flagged by LinkedIn coverage of the guide.
Why does this matter for your business?
An entire cottage industry has grown around "AI search optimisation", selling llms.txt generation, chunking services and bespoke AI schema as must-have infrastructure. Google's guidance suggests much of that spend, for the Google channel specifically, has been wasted.
That matters because Google still dominates how people find things through AI. Nearly 40% of Google's AI Overviews rank in the top 10 organic search results, and nearly 70% rank in the top 100, according to CXL. That means the foundation for showing up in Google's AI features is still classic ranking performance, not a separate AI-specific playbook.
The guide also reinforces why grounding matters. Google's generative AI features rely on retrieval-augmented generation, a technique that improves accuracy and freshness by pulling relevant, up-to-date pages from Google's Search index and generating a response with clickable links, as described in Google's own guide. If your page isn't retrievable through normal indexing, no amount of AI-specific markup will fix that.
However, the guidance is scoped narrowly to Google. It doesn't necessarily apply to how ChatGPT, Perplexity or other assistants source information, and that distinction matters given how much traffic those platforms now send. ChatGPT led the AI chatbot market as of July 2026 with 53.9% of worldwide web visits across the seven largest generative AI chatbots, ahead of Google Gemini at 27.9% and Anthropic's Claude at 9.2%, according to Momentic. A tactic Google says is pointless might still carry weight elsewhere, even if the evidence for that is thin so far.
What should you do now?
Stop paying for llms.txt generation, chunking services or bespoke AI schema as if they were guaranteed levers for Google visibility. The data and Google's own documentation agree these are largely wasted effort for that channel.
Instead, focus on fundamentals Google explicitly says still matter: solid technical SEO, clean indexing and genuinely useful content. Google draws a sharp line between generic "commodity content" and original, first-hand material, contrasting a generic listicle with a non-commodity alternative such as "Why We Waived the Inspection & Saved Money: A Look Inside the Sewer Line", where the distinction is whether the content offers unique insight beyond common knowledge, per Google's guide.
Don't treat Google's word as the whole picture, though. As one analysis from Jasper put it, core SEO fundamentals apply everywhere, but the off-site visibility layer that Google's guide de-emphasises is likely more important for AI platforms that aren't anchored to Google's own index. That means checking your presence on the review sites, forums and comparison pages that ChatGPT and Perplexity actually cite is still worthwhile, even if it isn't what Google's guide is talking about.
Given how differently each platform sources and cites information, it's worth actually checking where your brand does and doesn't show up across these systems rather than assuming. Running a free audit, such as the one at Sited, is a quick way to see your current AI visibility before deciding where to spend time and budget.
Finally, keep watching for divergence between what Google's documentation says and what its own product teams do in practice, since even Google's teams haven't been fully aligned on every point of this guidance.
Frequently asked questions
Do I still need an llms.txt file?
Not for Google Search. Google Search doesn't use these files to determine visibility, and creating one will neither help nor harm your rankings there. It may still be worth maintaining one if you specifically want to communicate with other AI platforms that read it, though evidence on actual usage remains limited.
Is content chunking still worth doing for AI search?
According to Google, no. There's no requirement to break content into tiny pieces, since Google's systems can understand nuance across multiple topics on a page and surface the relevant part to users. Independent schema testing found similar results, with no measurable citation benefit from the tactic.
Are AEO and GEO different from SEO?
Not according to Google. From Google Search's perspective, optimising for generative AI search is optimising for the search experience, and is therefore still SEO. The terms describe a focus area rather than a separate technical discipline, at least as far as Google's own systems are concerned.
Does this guidance apply to ChatGPT and Perplexity too?
No, it's scoped specifically to Google Search. Other platforms have different retrieval and citation behaviours, and tactics Google dismisses may still carry some weight on platforms that don't share Google's index or ranking systems.
What should I actually prioritise instead?
Solid technical SEO, including crawlability, indexing and page experience, combined with original, first-hand content that goes beyond generic summaries. Google's own framing is to create content people find genuinely useful, since that is what its systems are ultimately trying to surface.
Sources
Google's Guide to Optimizing for Generative AI Features on Google Search
Google's New AI Search Guide Calls AEO And GEO 'Still SEO' - Search Engine Journal
Google's AI Search Optimization Guide: What It Gets Right (and What's Missing) - CXL
Google publishes guide to optimizing for generative AI search - Semrush
Google Just Published Its Official AI Optimization Guide. Here's Everything That Changed - LinkedIn
July 2026 Top Generative AI Chatbots & LLMs by Market Share - Momentic
What is Generative Engine Optimization? GEO vs AEO vs SEO Guide 2026 - Jasper
From 15 September 2026, Cloudflare blocks mixed-use AI crawlers by default on ad-supported pages, unless site owners opt back in. Here's what it means.
Cloudflare will block "mixed-use" AI crawlers by default from any page that carries adverts, starting 15 September 2026. In practice, bots that blend search, agent use and AI training will lose automatic access to monetised web pages unless the site owner opts back in, forcing AI companies to separate search crawling from training and pay publishers when their content actually gets used.
For anyone tracking their brand's visibility in ChatGPT, Gemini or AI Overviews, this is the biggest structural change to how AI systems access web content since crawling began at scale. It will not show up as a ranking drop or a citation league table, but it will quietly determine which sites AI models can even see.
What changed?
Cloudflare announced the policy on 1 July 2026, giving the AI industry a hard deadline to separate crawlers used for traditional search from those used for AI agents and model training. As TechCrunch reported, starting on 15 September 2026, Cloudflare's default settings will block mixed-use crawlers from any pages that host ads, unless the site owner adjusts the settings otherwise.
The new defaults will not touch every site overnight. According to TechCrunch, the changes will apply to new Cloudflare customers, new sites set up by existing customers, and all existing free customers. Existing paying customers who have already configured their own crawler settings will not be switched over automatically, according to AI Chat Daily.
Alongside the block, Cloudflare is relaunching its bot payment scheme. As TechCrunch explains, the company's earlier Pay Per Crawl marketplace, which let websites charge AI bots for scraping, is evolving into "Pay Per Use," aiming to pay publishers when their content creates value rather than simply when it's fetched. The first two partners testing this model are Ceramic.ai and You.com: when a publisher opts in, they're paid when their content appears in Ceramic's AI search results or when You.com accesses a piece of their premium content.
The move follows clear frustration inside Cloudflare about how little traffic AI companies send back in exchange for what they crawl. Data cited by Let's Data Science shows how lopsided the exchange has become: Cloudflare's 2025 crawl-to-referral ratios put Google at 14:1, OpenAI at 1,700:1 and Anthropic at 73,000:1, arguing the historic crawl-for-traffic bargain has broken down for AI crawlers.
Cloudflare CEO Matthew Prince framed the change as overdue given how the internet's traffic mix has shifted. Per Technology.org, Prince said: "Now that the majority of traffic on the Internet is non-human, we must go further and act faster so that a sustainable ecosystem can emerge," referring to the recent milestone where bots surpassed human traffic online for the first time.
Google is the elephant in the room. Its crawler blends search indexing with AI training, and Cloudflare's messaging clearly targets that setup without naming it outright. Technology.org reports that Cloudflare calls out the world's largest search engine for holding roughly twice the information access of other AI companies, since staying discoverable often means accepting AI use too. Google does offer a workaround: Google-Extended lets site owners opt out of having content used for training and AI products such as Gemini Apps and Vertex AI, without affecting inclusion in Google Search, though its flagship Googlebot still crawls for Search, including AI features like AI Overviews and AI Mode.
Why does this matter for your business?
This is not a niche infrastructure story. Cloudflare sits in front of a substantial slice of the web, so its defaults effectively become policy for a huge number of sites without their owners lifting a finger. As The Next Web puts it, the company sits in front of a large share of the world's web traffic.
If your website runs display advertising and sits on Cloudflare, the crawlers that currently feed answers into ChatGPT, Perplexity or AI browsers may simply stop reaching your ad-supported pages from September, unless you or your web team actively re-enable access. That's a direct threat to AI visibility for exactly the kind of publisher-style content, guides, comparisons and explainers, that brands rely on to get cited in generative answers.
The flip side is leverage. Cloudflare is giving publishers a genuine negotiating position for the first time, rather than a binary choice between blocking everything or giving content away for free. Let's Data Science notes that major outlets are already backing the shift, with reporting naming organisations including The Associated Press, Time, The Atlantic and Reddit as participants in publisher advocacy.
There's also a data-quality upside buried in the announcement. Cloudflare says much of the current crawling activity is wasted anyway. Per TechCrunch, Cloudflare's data suggests over 50% of crawl traffic from AI crawlers is spent re-fetching unchanged pages, meaning a chunk of the current crawl-to-referral imbalance is inefficiency rather than genuine value extraction.
None of this happens in isolation. It lands amid a wider standoff between publishers, regulators and AI platforms over who gets to read the web for free. The Next Web reports that the UK is forcing Google to let publishers opt out of AI search without losing their ranking, and news publishers are suing OpenAI over training.
What should you do now?
First, find out whether you're on Cloudflare and what your current bot settings actually allow. Many site owners have never touched these settings and won't know they're about to be switched from "open by default" to "closed by default" on any page carrying ads.
Second, treat 15 September as a real planning date, not background noise. Decide deliberately which AI crawlers you want reaching your content, rather than letting a default setting make that call for you. If AI citations already drive meaningful referral traffic or brand mentions, blocking indiscriminately could cut off a channel you didn't realise you depended on.
Third, separate your thinking about "search visibility" from "AI training exposure." A crawler that helps you get cited in a live AI answer is doing something different from one that hoovers up your content to train a future model with no attribution at all. Cloudflare's whole policy is built around forcing that distinction into the open, and your content strategy should make the same distinction.
Finally, this is a good moment to check where your brand currently stands in AI-generated answers, since the crawler landscape you're being cited from is about to shift under your feet. Running a free check, such as the one available through Sited's audit at https://sited.online, gives you a baseline before the September changes take effect, so you can tell whether any drop in visibility later in the year is down to Cloudflare's new defaults or something else entirely.
Frequently asked questions
What is a "mixed-use" crawler?
It's Cloudflare's term for a bot that combines multiple jobs at once, typically search indexing, AI agent retrieval and model training, under a single crawler identity. As Proxycove explains, this is how Cloudflare refers to bots that collect data for two purposes simultaneously: search indexing and training AI models.
Will this block ChatGPT, Gemini or Perplexity from citing my site?
Not automatically for pure search-and-answer crawlers, but any crawler that Cloudflare classifies as mixed-use will lose default access to your ad-supported pages unless you opt back in. The practical effect depends heavily on how each AI company's crawlers are classified.
Does this affect Google Search rankings?
No. Cloudflare's plan keeps standard search indexing working as before; the change specifically targets crawlers that also harvest content for AI training or agent use. Google's own opt-out tool, Google-Extended, already lets sites block AI training use without losing Search inclusion.
What is "Pay Per Use" and how is it different from "Pay Per Crawl"?
Pay Per Crawl let publishers charge a fee every time a bot fetched a page. Pay Per Use goes further, aiming to pay publishers when their content creates value, not just when it's fetched, tying payment to actual use in an AI answer rather than a simple fetch.
Should small businesses worry about this if they're not big publishers?
Yes, if you run a content-heavy site with advertising and rely on AI citations for discovery. Even without ad revenue, it's worth checking your Cloudflare crawler settings now, since the classification of "mixed-use" bots could affect how future AI products access your content regardless of your monetisation model.
Sources
Cloudflare's new policy pushes AI companies to pay for publishers' content - TechCrunch
Cloudflare to Block Mixed-Use AI Crawlers on Ad Pages From September 15 - Technology.org
Cloudflare will block mixed-use AI crawlers by default from September 2026 - AI Chat Daily
Cloudflare gives AI crawlers a September deadline: pay publishers or get blocked - The Next Web
Cloudflare Blocks Mixed-Use Crawlers on Monetized Pages - Let's Data Science
Cloudflare blocks AI crawlers on advertising sites (2026) - Proxycove
A new study finds Google's AI Overviews and AI Mode drive more AI-influenced traffic than ChatGPT, Gemini, Claude and Perplexity combined. Here's what it means for brands.
Google's AI Overviews and AI Mode generate more AI-influenced traffic than every standalone chatbot combined, according to a new report from AI discovery agency Previsible. ChatGPT remains the runaway leader among standalone assistants people query directly, but Gemini and Claude are closing the gap fast. The practical takeaway for brands chasing AI mentions: win inside Google's ecosystem first, then chase the standalone chatbots.
What changed?
Previsible released the third edition of its AI Traffic Study on 6 July 2026, described as the industry's earliest and longest-running look at GEO trends, tracking data from November 2024 to May 2026 (Previsible via Businesswire). The study analysed 6.77 million AI-driven sessions across 166 websites spanning SaaS, e-commerce, finance, legal, health, insurance, education and publishing.
The headline finding concerns where AI discovery actually happens. AI discovery inside Google, through AI Overviews and AI Mode, represents more AI-influenced traffic than every LLM assistant combined, and Previsible argues it should remain the priority surface for marketers (Previsible via Businesswire).
Among standalone AI platforms people visit directly, ChatGPT is in a league of its own, carrying 92.4% of trackable standalone referral traffic and still climbing. But the report flags real movement beneath that number: Gemini grew 3.2x with steady consistency and now ranks second behind ChatGPT, while Claude grew 64x over the tracked period and overtook Perplexity in March 2026, showing particular strength among developers, technical buyers and professional services (Previsible via Businesswire).
Ecommerce brands should take particular note. Across models, e-commerce content saw AI referral traffic rise 37x, suggesting shoppers increasingly arrive on product pages with intent already formed rather than just browsing.
The study lands the same week Google clarified its own position on this terminology. On 15 May 2026, Google published an official guide stating that AEO and GEO are still SEO, and that four widely promoted AI optimisation tactics, llms.txt files, content chunking, rewriting content specifically for AI, and overfocusing on structured data, are unnecessary for Google Search (Bring). Crucially, Google did not claim this applies to ChatGPT, Perplexity, Claude or other AI search systems, since those retrieve content differently. That distinction matters if you're trying to build one playbook for every AI surface: what works for AI Overviews won't necessarily work for ChatGPT search or Perplexity.
Why does this matter for your business?
If you've concentrated your GEO effort on standalone chatbots, this data suggests you may be optimising for the smaller slice of the pie. Google's AI Overviews and AI Mode sit inside the search engine practically everyone already uses, so the volume of AI-influenced queries flowing through Google dwarfs anything ChatGPT, Gemini, Claude or Perplexity currently generate on their own.
That doesn't make standalone assistants irrelevant. Cloudflare Radar data tracked through early July shows chatgpt.com remains one of the busiest sites on the internet, and Sensor Tower data cited by Reuters shows ChatGPT's app crossed 1 billion global monthly active users in June 2026, the fastest app in history to reach that milestone, ahead of TikTok, Instagram and Google Maps (TechnologyChecker.io).
There's a catch worth noting, though: this traffic converts brilliantly despite still being small in volume. One case study found traffic referred from ChatGPT converts at 16%, compared with 1.8% for Google Organic, nearly nine times higher, even though AI traffic still accounts for less than 1% of overall organic traffic (fatjoe).
Put together, the picture is this: Google is the volume play, standalone chatbots are the high-intent, high-conversion play, and both need attention, just not equal amounts of it.
There's also a structural challenge worth flagging. AI answers aren't stable in the way Google search rankings are. Between 40% and 60% of cited sources change month-to-month across Google AI Mode and ChatGPT, making visibility far less predictable than organic rankings (EMARKETER). As one analyst put it, almost every GEO response is different from every other GEO response, unlike Google search, where you'd get a consistent answer if you queried it ten times.
What should you do now?
Previsible's own recommendations offer a sensible starting checklist. The firm and its Chief Product Officer, David Bell, suggest businesses start by becoming a source Google's AI results want to cite, building the site architecture and content signals AI systems rely on, then work to win ChatGPT as the leading standalone surface (Previsible via Businesswire).
More broadly, the report recommends five core efforts to win in AI search: create citation-worthy evidence, build authority across trusted third-party sources, make websites easy for AI systems to read and extract, optimise for answer journeys, and measure business impact rather than site-wide visibility alone.
In practice, that means:
Auditing whether your key pages are technically easy for AI crawlers to read and extract, not just indexable by traditional search bots.
Building genuine third-party authority, such as press coverage, reviews and expert citations, rather than relying purely on owned content, since AI systems weight external validation heavily.
Tracking citation appearances across Google's AI features and standalone assistants separately, since the same tactics don't transfer cleanly between them.
Measuring conversions and revenue from AI referrals, not just raw visibility, given how disproportionately well AI-referred visitors tend to convert.
Because AI citations shift so quickly and unevenly across platforms, it's worth regularly checking how your own brand actually appears in AI answers rather than assuming last month's visibility still holds. Sited's free audit at sited.online lets you check your current AI visibility across these different surfaces in one place, a useful starting point before deciding where to invest.
Frequently asked questions
What is the difference between GEO and AEO?
The terms are used almost interchangeably in practice. As one analyst put it, AEO and GEO describe the same underlying approach, and no common taxonomy exists for the category (EMARKETER). Broadly, GEO refers to making your brand more likely to be mentioned in AI-generated answers, while AEO focuses on getting content extracted as a direct answer.
Does Google's AI Overviews really matter more than ChatGPT for AI visibility?
According to Previsible's latest study, yes, in terms of overall volume. AI discovery inside Google, through AI Overviews and AI Mode, represents more AI-influenced traffic than every LLM assistant combined (Previsible via Businesswire). ChatGPT still dominates among standalone assistants, so both surfaces deserve attention.
Do I need to rewrite my whole site for AI crawlers?
Not according to Google's own guidance. Google's May 2026 guide explicitly calls out llms.txt files, content chunking, rewriting content specifically for AI systems, and overfocusing on structured data as unnecessary for Google Search (Bring). This guidance is specific to Google, however; other AI platforms retrieve and rank content differently, so a one-size-fits-all approach isn't advisable.
How stable is AI search visibility once you achieve it?
Not very. Between 40% and 60% of cited sources change month-to-month across Google AI Mode and ChatGPT (EMARKETER), so brands need to treat AI visibility as an ongoing effort rather than a one-off project.
Is Claude or Gemini worth optimising for specifically?
Increasingly, yes, particularly for certain audiences. Previsible's data shows Claude grew 64x over the tracked period and overtook Perplexity in March 2026, with particular strength among developers, technical buyers and professional services, while Gemini has shown steady, consistent growth as the second most visible standalone model (Previsible via Businesswire).
Sources
Semrush analysed 126 million AI search prompts and found only 36 brands consistently visible across ChatGPT, Gemini and Google AI. Here's what it means for you.
Semrush has published its expanded 2026 AI Visibility Index, analysing 126 million US AI search prompts across ChatGPT, Gemini, Google AI Mode and Google AI Overviews. The headline finding is that visibility is now highly concentrated in a handful of categories, with only 36 brands appearing in the top 100 on every platform every month, and that being mentioned by an AI system is not the same as being cited as its source. For most businesses, this confirms that AI search visibility needs active management, not an assumption that good old SEO will carry over automatically.
What changed?
Semrush launched the original AI Visibility Index in September 2025 with a sample of 2,500 prompts. The new 2026 edition is a different order of magnitude entirely.
The expanded study analysed 126 million US AI search prompts from January through April 2026, making it one of the largest datasets yet assembled on how brands are mentioned, cited and represented inside AI-powered discovery tools. That scale matters because it moves the conversation from anecdote to pattern: this is close to a census of how real people ask AI systems about brands, not a handful of test prompts.
The report arrives at a moment when, as HubSpot notes, ChatGPT now processes over 2.5 billion prompts per day, and industry analysts expect AI-driven search traffic to overtake traditional search by 2028.
What does the data actually show?
The most striking finding is how unevenly visibility is spread across industries. Semrush found that in News and Media, the three most visible brands accounted for 82.9% of total category visibility, while in Consumer Electronics the top three represented 76.9%. By contrast, visibility was far more distributed in Finance, where the top three brands accounted for 41.4%, and Industrial, where they represented 42.2%.
That gap matters for strategy. Semrush's own analysis suggests these less concentrated categories may offer greater opportunities for brands to gain visibility over time, because the field hasn't already been locked up by three dominant names.
Consistency across platforms is rare. Only 36 global brands maintained top-100 visibility across all four platforms during every month of the study, a group Semrush calls the "Universal 36", including YouTube, Google, Reddit, Amazon, Facebook, Apple, Walmart, Disney and Nintendo. Everyone else appears inconsistently: strong on one platform in one month, absent the next.
The platforms themselves behave very differently when it comes to sourcing their answers. Semrush found that ChatGPT cites an average of 15 sources per response and frequently relies on community and reference platforms such as Reddit and Wikipedia, while Gemini cites an average of just 3 sources per response, drawing on a smaller pool that includes Wikipedia, Reddit and YouTube. A brand optimised for one platform's citation habits could easily be invisible on another.
Perhaps the most important nuance for marketers is the distinction between being talked about and being credited. Semrush is explicit that brand mentions and citations measure different forms of visibility, and being mentioned in an AI-generated answer does not necessarily mean a brand's own website is cited as the supporting source. An AI system can describe your product accurately while sending all the credibility, and the click, to a third-party review site or forum thread instead of you.
Why does this matter for your business?
This is no longer a niche concern for SEO specialists. As Pew Research data highlighted by HubSpot shows, Google's AI Overviews already appeared in 18% of US searches, and that share is climbing. If your brand isn't part of the source mix behind those answers, you are effectively absent from a growing share of the discovery journey, even if your traditional rankings look healthy.
The concentration data is both a warning and an opportunity depending on your sector. If you sell consumer electronics or operate in news and media, Semrush's findings suggest the top three brands already dominate AI answers in your category, which makes displacement hard but not impossible. If you're in finance or industrial sectors, the more distributed picture means there's genuine room to build visibility before a small group of incumbents locks it down.
The mentions-versus-citations gap is arguably the most actionable insight here. A brand can be name-checked constantly by AI tools while getting none of the traffic, trust transfer or link equity that comes from being the cited source. As Andrew Warden, Vice President of Marketing at Adobe and former CMO of Semrush, put it in Semrush's announcement, the foundations of SEO remain critically important for creating trust signals, but visibility now depends on how consistently a brand reinforces its narrative across every digital channel, not just its own website.
What should you do now?
Start by finding out where you currently stand. Before you can fix a visibility gap, you need to know whether AI systems mention you at all, and separately, whether they cite you as the source when they do. This is exactly the kind of check-up Sited's free audit at sited.online is built for, giving you a quick read on how your brand currently shows up across AI answer engines.
Once you know your baseline, treat each platform separately rather than assuming one strategy fits all. Given that ChatGPT draws on roughly 15 sources per answer including community platforms, while Gemini draws on around 3, a presence on Reddit and Wikipedia may do more for ChatGPT visibility than a polished product page ever will, while Gemini visibility depends on a narrower, more concentrated set of trusted sources.
Prioritise being cited, not just mentioned. That means making sure your own pages contain the specific facts, figures and direct answers that AI systems can lift and attribute, rather than generic marketing copy that gets paraphrased and credited to someone else.
Finally, revisit your category's concentration level. If you're in a fragmented category like finance or industrial, act now while the field is still open. If you're in a concentrated category like consumer electronics or media, focus on niche queries and specific use cases where the dominant three brands are less likely to be the default answer.
Frequently asked questions
What is the Semrush 2026 AI Visibility Index?
It's a flagship study from Semrush that analysed 126 million US AI search prompts from January to April 2026 across ChatGPT, Gemini, Google AI Mode and Google AI Overviews, expanding on an original 2,500-prompt study launched in September 2025.
What is the "Universal 36"?
It's Semrush's name for the group of 36 global brands that maintained top-100 visibility across all four AI platforms studied, every month, including names like YouTube, Google, Amazon and Apple. Almost every other brand's visibility fluctuates month to month and platform to platform.
Is being mentioned by ChatGPT or Gemini the same as being cited?
No. Semrush's data shows brand mentions and citations measure different forms of visibility, and an AI-generated answer can mention a brand without citing that brand's own website as the source. You want both, but citation is what typically drives referral traffic and trust.
Which industries have the most room to gain AI visibility right now?
According to Semrush, Finance and Industrial are the least concentrated categories, with the top three brands accounting for only 41.4% and 42.2% of visibility respectively, compared with over 80% in News and Media. That makes them more accessible for brands trying to build AI visibility from scratch.
Why does Gemini cite fewer sources than ChatGPT?
Semrush's study found that Gemini cites an average of just 3 sources per response compared with ChatGPT's average of 15, suggesting it draws on a smaller, more curated pool. The exact reasons aren't disclosed by Google, but it means brands need a more concentrated, high-authority footprint to be picked up there.
Comment