A rank tracker says Google has been slipping wrong results to scraping tools since September, and Google has not said a word.
- ReportedGoogle is blocking scrapers and, by one account, feeding them wrong results
- ReportedSearch Console's Performance report is back to normal
- ReportedStanding items: ChatGPT, Perplexity, Claude
Read the issue in full → A quiet Sunday. Google's Search Console report is more than 64 hours behind, and a count of 1,851 local firms finds fewer than half named by AI.
- ReportedSearch Console's Performance report is more than 64 hours behind
- ReportedStanding items: ChatGPT, Perplexity, Claude
- ReportedZarla: fewer than half of top-listed local trade businesses were named
- ReportedInsites: AI knew most local businesses and got nearly half of them wrong
Read the issue in full → A quiet Saturday. A rank tracker admits a week of its "ranking drops" during Google's spam update were not real.
- ReportedA rank tracker says Google fed it incomplete results through the spam update
- ReportedStanding items: ChatGPT, Perplexity, Claude
- ReportedWellows: sites cited elsewhere are the ones cited again
- ReportediPullRank: models guess a cause about half the time when the evidence is missing
Read the issue in full → Anthropic's new usage rules, effective 12 November, bar using Claude to plant fake "independent" sources for AI answers.
- ConfirmedAnthropic's usage policy names source-seeding for AI answers as abuse
- ConfirmedStanding items: Perplexity, Claude, ChatGPT
Read the issue in full → Google's spam update finished this morning after two weeks, so Google AI Overviews readings can be treated as settled again.
- ConfirmedGoogle's September spam update is complete
- ReportedChatGPT's app moves to GPT-6, and free users switch today
- ReportedAnthropic releases Claude Haiku 5.5
- ReportedPeec researcher: ChatGPT still quotes 50,000 pages of invented statistics
Read the issue in full → Google's rules for a business's own Maps listing now bar prices, time-limited offers and calls to action.
- ConfirmedGoogle's Business Profile policy: no prices, offers or calls to action in the listing
- ConfirmedGoogle's crawl-rate guide: error codes for more than a day or two can cost you the page
- ConfirmedGoogle's spam update: day fourteen, one post
- ReportedStanding items: Perplexity and Claude
Read the issue in full → Google's John Mueller on llms.txt: "the hope is bigger than the reality". AI crawlers are reading plain sitemaps and RSS feeds instead.
- ReportedMueller: AI crawlers read sitemaps and RSS feeds, and llms.txt does nothing in Google's systems
- ReportedGoogle's spam update: day thirteen, one post, and a reported last phase
- ConfirmedGoogle's opt-out setting now warns the parent owner when a sub-site overrides it
- ConfirmedPerplexity: new October changelog entries, none about Sonar
- MeasuredAhrefs: Google AI Overviews on 83% of brand-name searches
Read the issue in full → Google put numbers on how slowly it notices a change: about 30 days to re-read a page it already knows.
- ReportedGoogle gave typical timings for crawling, indexing and recovery
- ReportedGoogle on spam updates: scaled content, and "months" as the slow case
- ReportedA second Claude model in the engine is near its earliest retirement date
- ConfirmedStanding items: Perplexity
- ReportedOtterly: German AI answers cite German sources about two times in three
Read the issue in full → The "Google is tagging AI Overviews links" story was a browser extension, not Google.
- ReportedThe AI Overviews tracking tag was a browser extension
- ReportedGoogle AI Overviews now appear on most tracked brand-name searches
- ConfirmedStanding items: Google's spam update, Perplexity, Claude
- ReportedBuzzStream and YouGov: most Americans who have used AI say it has made things up
Read the issue in full → Google now calls a made-up author name "a form of deception", and says it marks a page as low quality.
- ConfirmedGoogle: a made-up author is "a form of deception"
- ReportedGemini starts tagging the links it sends out
- ConfirmedStanding items: Google's spam update, Perplexity, Claude
Read the issue in full → Google now says it is "critical" to fact-check every AI-written page by hand before publishing, and that covers titles, descriptions and schema too.
- ConfirmedGoogle: fact-check AI-written content by hand, including the parts nobody reads
- ConfirmedGoogle's "helpful content" page now spells out effort, originality, skill and accuracy
- ConfirmedStanding items: Perplexity, Claude, Google's spam update
- ReportedOtterly: Reddit's share of ChatGPT citations fell 89% in August
- ReportedGreen Flag Digital: half of ChatGPT's most-cited finance pages now get no citations
Read the issue in full → Anthropic has now put the model behind our Claude column on a retirement clock: deprecated 30 September, switched off 30 November.
- ReportedThe Claude column's model is formally deprecated
- ConfirmedCloudflare publishes who blocks what, and opens "Pay Per Use"
- ConfirmedSonar re-routing, day four, and Google's spam update, day eight
- ReportedProfound: a cited page fades in days
- ReportedA law-firm benchmark: the assistants name directories, not firms
- ConfirmedCloudflare's traffic numbers
Read the issue in full → The Claude column runs on a model that passed its earliest retirement date yesterday, and when its web search fails the answer still looks normal.
- ReportedThe Claude column's model reached its earliest retirement date
- ReportedOpenAI DevDay: a new model, nothing on search
- ConfirmedSonar re-routing, day three, and Google's spam update, day seven
- ReportedPeec: why visibility scores jump around
- ReportedGoogle AI Overviews and the click
- ReportedGoogle says AI Mode has over 1 billion monthly users
Read the issue in full → Perplexity now says old Sonar requests are being quietly re-routed to its new API, so the Perplexity column can change without a single error.
- ConfirmedSonar requests are now being re-routed without notice
- ReportedGoogle opens AI Mode monitoring to everyone
- ConfirmedSpam update: day six, still open
- Reported"Users won't click": OpenAI's own words, via the publishers' lawsuit
- ReportedProfound: half of cited content is under 13 weeks old
Read the issue in full → OpenAI's own help page lists the reasons a website goes missing from ChatGPT's index, and most of them are things a coach's web host can do without the coach knowing.
- ConfirmedOpenAI says why a page is missing from its index
- ConfirmedSonar still answers a day after its retirement date
- ReportedGoogle's spam update: a heavy weekend, day five
- ConfirmedCloudflare: automated traffic passed human traffic in May 2026
Read the issue in full → Perplexity's Sonar API is still answering on its retirement day, but the replacement it points to changed twice this month, so the migration should pin a model and a search setting, not a preset.
- ConfirmedSonar still answers on its last supported day
- ConfirmedPerplexity's replacement presets moved twice this month
- ConfirmedGoogle's spam update: day four, still open
- ReportedBrightEdge: ChatGPT sends 95.1% of AI referral traffic
Read the issue in full → Perplexity's old API shuts off tomorrow, and a new measurement paper shows why asking an assistant once is not enough to know who it recommends.
- ConfirmedPerplexity Sonar retires tomorrow
- ReportedGoogle's September spam update: day three, volatility spiking
- MeasuredOne ask is not a full answer: a repeated-query audit of 50 buying questions
Read the issue in full → Google started a spam update that can run two weeks, and it overlaps the Perplexity cutoff on 27 September, so the next fortnight is a bad window to measure any before and after.
- ConfirmedGoogle's September 2026 spam update is rolling out, for up to two weeks
- ConfirmedPerplexity Sonar retires in two days; still no migration
- ReportedBing tests calling its Copilot search tab "AI Mode"
- ReportedPublishers lost 40% of their Google Search traffic in a year
- ReportedGoogle Research: a faster way to split one question into ten searches
Read the issue in full → A vendor changed only the "last updated" date on ten pages and AI citations rose 31.5%, and Google now spells out that strangers' edits to a business listing can go live after four days.
- ConfirmedGoogle: ignore a suggested edit for four days and it may go live
- ReportedGoogle tests moving AI Overviews into a side panel for hotel searches
- ConfirmedPerplexity Sonar retires in three days; still no migration
- ReportedChanging only the date lifted AI citations 31.5% on one vendor's blog
- MeasuredA small lab test: fabricated detail can outrank honest detail
- ReportedChatGPT ads in almost half of video-game chats
Read the issue in full → Four assistants, almost no shared sources, and Google tests links inside AI Overviews that lead back into Google.
- ReportedGoogle tests AI Overview links that send people to AI Mode, not to a website
- ConfirmedPerplexity Sonar retires in four days; the migration branch still does not exist
- MeasuredSame prompt, four engines, almost no shared sources
- MeasuredWhen your own site is in what the assistant retrieved, you get named far more often
- ReportedOtterly's September update: social sources reshuffle fast
Read the issue in full → Google's own tooling now audits a file that does nothing for a coach, and a controlled experiment finally puts a number on the disappearing click.
- ReportedGoogle's Lighthouse now audits a file that has nothing to do with being recommended
- ConfirmedPerplexity Sonar retires in five days and the migration still has not started
- MeasuredRouting searches to AI Mode cut click-through by 18.8 points and made people trust Google less
Read the issue in full → A genuinely quiet Monday, and a paper arguing that the question list is the measurement.
- ConfirmedPerplexity's Sonar retirement is six days out and the migration has not started
- MeasuredA prompt list is not a sample of demand, it is a definition of the market being measured
Read the issue in full → 94% of 28,488 Florida businesses never got named once, and the Perplexity cutover is seven days out.
- ReportedGoogle v. SerpApi: the hearing is 13 October, and SerpApi just lost the discovery fight that matters
- ReportedSISTRIX has been collecting at a reduced rate for four days with no resolution posted
- ReportedOnly 5.9% of 28,488 local businesses were ever named by an AI assistant
Read the issue in full → Google gets better at blocking the tools that read its results, and Perplexity's cutover is eight days out.
- ReportedGoogle is blocking search scrapers and rank trackers far more successfully since about 13 September
- MeasuredFormatting a page redistributes citations, it does not clearly get the page cited
- MeasuredReddit's AI search picks the formal, already-upvoted comments
Read the issue in full → A hidden price gets filled in by other sites, and the Perplexity cutover is nine days out.
- ConfirmedPerplexity Sonar retirement: nine days out, no migration started
- ReportedWhen a company hides its price, AI quotes someone else's
- ReportedOne agency's glossary: 0.10% click rate on Google, most AI crawler visits were training
Read the issue in full → Google tests AI summaries inside business knowledge panels, and a judge rules AI Overviews can defame.
- ReportedGoogle tests an AI Overview inside the local knowledge panel
- ReportedA federal judge lets a defamation claim over AI Overviews proceed
- ConfirmedOpenAI launches Sponsored Agents and puts ChatGPT ads inside HubSpot
- ReportedCompany-owned pages take over half of AI citations across 17 industries
- ReportedA company's name alone shifts how AI describes it, until the brand is known
Read the issue in full → Perplexity retires its Sonar API in 11 days, and Cloudflare's "Block" now shuts out Google too.
- ConfirmedPerplexity retires its Sonar API on 27 September
- ConfirmedCloudflare's "Block" now stops Googlebot, and most sites were moved to a softer setting
- ReportedBrand mentions are steady, cited sources churn daily
- ReportedChatGPT runs its own family of search indexes, including a local one
Read the issue in full → Most AI citations never say the brand's name, and Cloudflare's AI crawler default takes effect today.
- ConfirmedCloudflare's AI crawler default is live from today
- Measured62% of AI citations never name the brand they link to
- ReportedTwo experiments: other people's pages drove the mentions, and half stopped within a month
Read the issue in full → Google will pay some publishers when their page shapes an AI answer, and says a link added after the answer is written does not count.
- ReportedGoogle is piloting payments for content that shapes AI answers
- ReportedUnconfirmed Google ranking dip from 4 September reversed on 13 September
- ReportedAssistants change their brand picks depending on who is asking
Read the issue in full → Google says Search Console cannot usefully tell you where you sit inside an AI answer, which turns the "Google gives me this for free" objection into Google's own words.
- ReportedGoogle: AI feature position in Search Console is a block, not your link
- MeasuredThe most careful estimate of AI Overviews' traffic effect is about 5%
Read the issue in full → Google documented an official full-web search API and put it behind a partner door, which narrows the second of the two routes we have to AI Overviews data.
- ReportedGoogle has documented an official full-web search API, and it is partner-only
- ReportedGoogle will not backfill the June gap in Search Console's page indexing report
- ReportedChatGPT Shopping swapped the open web for a product feed in a single day
- ReportedBuyers ask longer questions, and our question set already sits at the long end
Read the issue in full → Perplexity's share of AI chatbot referrals nearly halved in two months, and the one big assistant we do not audit now sends slightly more traffic than one we do.
- ReportedWhich assistants actually send visitors, August 2026
- ReportedLe Monde says AI Overviews have not hurt its audience so far
- MeasuredThe only systematic review of the field says almost nothing is proven
- ReportedOnly 14% of content marketers say their blog works, and AI made no difference
Read the issue in full → The last question box on Google that still handed out links is now an AI answer 97% of the time, measured across 19.2 million queries.
- ReportedPeople Also Ask is now 97% AI Overviews
- ConfirmedGoogle is now telling retailers the actual questions people ask its AI
- ReportedApple's rebuilt Siri ships Monday, into 2.5 billion requests a day
- ReportedPublishers spent $113 million in one month buying back the traffic AI answers took
- ReportedWhat content actually gets cited, by format, in July
- ReportedOlder, and only worth one line: who is actually blocking AI crawlers
Read the issue in full → A monitoring vendor shipped a free, no-login AI visibility report yesterday that shows the customer the real answer text across three engines, which is three quarters of our free check given away instantly.
- ReportedGoogle is testing a thinner citations panel on AI Overviews
- ReportedGoogle's John Mueller on mass-produced pages and site-level trust
- MeasuredOnly 15.2% of categories in ChatGPT have an owner, and small categories are more likely to have one than big ones
Read the issue in full → Apple has now written down that its AI opt-out cannot cost you search ranking, eight days before Cloudflare's new default makes it easy to block the wrong thing entirely.
- ReportedApple has written down that its AI training opt-out does not affect search ranking
- ConfirmedChecked, and nothing else landed
- ReportedBeing cited in a Google AI Overview is worth about twice the clicks of sitting uncited on the same page
- ReportedThe number that would price the whole product is published without a method
Read the issue in full → The category spent three weeks explaining Reddit's citation collapse with a robots.txt change that never happened, and the vendor who collected the data blames a ChatGPT change this radar already carried on 21 August.
- ReportedReddit's collapse in ChatGPT citations was not a robots.txt block, and I checked
- ReportedFive engines agree on brands about three times more often than on pages
- ReportedUS traffic, the year to July 2026
Read the issue in full → Google called yesterday's missing citations a bug and fixed it inside 36 hours, and in the same 48 hours OpenAI shipped a new top model, which means two of our five engines changed the software behind their answers this week and our pipeline recorded neither.
- ReportedGoogle says the missing AI Mode citations were a bug, and they came back the next morning
- ConfirmedOpenAI shipped GPT-6 Astra, and it reaches the API within days
Read the issue in full → Google put a new model inside AI Mode on Tuesday and by Wednesday morning practitioners were posting answers with no links and no citations at all.
- ReportedGoogle's newest AI Mode model is answering without citing sources
- ReportedGoogle's free AI shopping reporting reached four more countries
- ReportedWeek-to-week citation churn measured across 1.5 million snapshots
- ReportedThe intent shape of local AI Overviews, four months of caveat included
Read the issue in full → Google published the terms, the eligibility test and the dates for handing its European search data to rival answer engines, two days after the EU formally classified ChatGPT as one.
- ConfirmedGoogle published the rules and the calendar for licensing its European search data
- ReportedBing is testing result snippets with no site name
Read the issue in full → Google shipped the AI opt-out switch and the AI impressions report to every website in the world on 31 August, and on the same day the European Commission formally classified ChatGPT as a search engine.
- ConfirmedGoogle's AI opt-out switch and AI impressions report went global
- ReportedThe EU designated ChatGPT a Very Large Online Search Engine
Read the issue in full → Ahrefs measured 75,000 brands and found that what gets you named by an AI is being mentioned elsewhere, not publishing more pages, which is the closest thing to independent proof of this business's core method that has ever appeared in this brief.
- ReportedMicrosoft put a login wall in front of Copilot, and left Bing's answers open
- ReportedGoogle moved shopping ads inside the AI Mode answer
Read the issue in full → There is a switch inside Google Search Console that removes a site from AI Overviews and AI Mode without touching its rankings, which is a client-side failure mode we had not accounted for and a correction to something this radar said on 10 August.
- ConfirmedGoogle shipped an opt-out from AI Overviews and AI Mode, and we told you it did not exist
- ConfirmedThe free Google report cannot tell AI Overviews from AI Mode
- ConfirmedMicrosoft is giving away a share-of-citations metric
- ReportedA small correction to yesterday's AI Overviews item
Read the issue in full → The google.com/goto link wrapper does reach AI Overviews, and AI Overviews is where it bites hardest, which turns yesterday's open question into a defect to fix rather than a thing to watch.
- ReportedThe goto wrapper reaches AI Overviews, and that is where it is worst
- ReportedGoogle is expanding AI Overviews to full answers by default
- ReportedAI Overviews and AI Mode credit local businesses in nearly opposite ways
- ReportedCommunity content is about half the top citations, in three industries that are not ours
Read the issue in full → Google confirmed it is rewriting search result links through google.com/goto, an anti-scraping measure that breaks domain matching for every tool that reads its results, and Google's Search Central blog broke its August silence with a site reputation policy update that names the four criteria it uses to judge third-party content.
- ReportedGoogle confirmed it rewrites search result links through google.com/goto
- ConfirmedGoogle's site reputation policy update names the four criteria it judges third-party content by
- ReportedForty-five percent of consumers now ask an AI assistant for a local business, and AI Mode is second among the tools they use
- ReportedEarned media and third-party validation are what move a local business into an AI recommendation
Read the issue in full → Google started putting rows of clickable article cards inside AI Mode answers, and the only study that splits AI citation rates by company size puts a solo operator at 1.4 percent.
- ReportedGoogle puts developing-topic link carousels inside AI Mode
- ConfirmedCloudflare's crawler deal has four conditions, and we recorded the post without them
- ReportedCitation rates by company size, and the number for a solo operator is 1.4 percent
Read the issue in full → The largest dataset on what earns an AI Mode citation puts domain traffic above every other factor, and large sites earn roughly three times the citations of small ones.
- ConfirmedGoogle starts rolling generative UI into AI Overviews
- ReportedWhat actually predicts an AI Mode citation, across 2.3 million pages
- Reportedllms.txt now has a number attached to it, and the number is zero
Read the issue in full → Google's John Mueller said yesterday that there is nothing special to do for AI answers, which is the single objection our sales page has to beat. It is also a claim about one engine out of five.
- ReportedGoogle says there is nothing special to do for AI answers
- ReportedAI answers about small local businesses are frequently wrong, not merely missing
- ReportedBeing cited is not being recommended, now with a number on it
Read the issue in full → Cloudflare will now rewrite a website's robots.txt for it, automatically and on by default for new customers, which turns the crawler-access file into a moving part of an audit.
- ConfirmedCloudflare now writes a site's AI crawler policy into robots.txt for it
- ConfirmedCorrection: the 15 September Cloudflare default is scoped to pages that carry ads
- Reported78% of independent local operators have effectively no citation share, per a PR agency's own research
Read the issue in full → The ad industry's trade body has written down what counts as a real measurement of AI visibility, and a 25-question board sits one rung below its bottom rung.
- ReportedThe IAB published measurement guidance that grades every AI visibility claim, including ours
- Reported57% of domains cited once are never cited again, in one narrow slice
Read the issue in full → ChatGPT does not run on Bing. It runs on its own index, and in the mode nine out of ten people use it reads only your page title and about 200 characters starting at your headline.
- ConfirmedGoogle's documentation now names AI Mode and AI Overviews for Preferred Sources
- ReportedSomebody instrumented ChatGPT's retrieval and wrote down what it stores about a page
- ReportedThe average local business scores 8.24% on AI visibility
- ReportedAI crawlers are getting blocked roughly twice as often as a year ago
Read the issue in full → On 8 August ChatGPT stopped only searching the open web and started sending one in six of its background searches straight at a named site. A separate study of 129 million citations found trade publications out-cited mainstream media by 213%.
- ReportedChatGPT started aiming its background searches at named sites
- ConfirmedGoogle's August 2026 spam update finished, and it was the longest of the three
- ConfirmedThe Search Console generative AI bug now has an end date
- ReportedPublishers who signed with OpenAI get cited more on ChatGPT, and trade press beats national press almost everywhere
Read the issue in full → Ask Gemini the same local service question a day apart and roughly three quarters of the sources it quotes have changed. Google's ordinary local pack returned the same top business 90.2% of the time in the same test.
- ReportedJavaScript-only navigation left AI crawlers finding nothing
- ReportedThe same question, asked twice, mostly changes its sources
Read the issue in full → ChatGPT's citations to Reddit fell 86% on a single day, 14 August, with no announcement from OpenAI and no explanation since. A source that was the most-cited domain on the platform in April is now a rounding error.
- ReportedChatGPT's citations to Reddit collapsed 86% on 14 August, unannounced
- ConfirmedGoogle released the August 2026 spam update, and those policies now govern AI Overviews
- Reported22.4% of AI Overview clicks arrive in analytics as Direct, not Organic
Read the issue in full → A study of 4,320 AI answers found that 77% of brands were named by only one of the three models tested, and just 11% were named by all three. That is the strongest public number yet for why one engine is not a measurement.
- ReportedGoogle confirmed a Search Console logging bug that understates generative AI impressions from 13 August
- Reported77% of brands were named by only one of three models
- MeasuredGoogle Research: the facts are in there, the models just cannot get them out
- ReportediPullRank, 13.1 billion search events, mostly not about us
Read the issue in full → 2,202 US adults were asked about AI search. Sixty-five percent use it, fifteen percent trust it a lot, and seventy-six percent say seeing where the information came from matters.
- ReportedAnthropic will watermark Claude's text, and has published how the watermark breaks
- ReportedMost people use AI search. Almost nobody trusts it.
- ReportedChatGPT referrals to B2B sites tripled in a year, from a very small base
Read the issue in full → ChatGPT appears to decide who it will recommend before it searches. In one practitioner's network trace, brands that appeared in ChatGPT's own self-written search query were cited 68.9% of the time, against 2.1% for brands that merely turned up in the results.
- ConfirmedOpenAI: robots.txt "may not apply" to the bot that fetches pages mid-conversation
- ReportedA measurement to go with it: about 15% of page fetchers in Europe reached disallowed URLs
- ReportedMeta is building a sixth engine, and it is crawling hard
- ReportedChatGPT names the brands before it looks them up
- ReportedThe category's problem is belief, not price
Read the issue in full → a second independent dataset says the AI referral is almost invisible. Scrunch found 1.1% of post-chat visits to news sites carried an AI referrer, while about three quarters arrived as direct navigation.
- ConfirmedGoogle's AI opt-out control is real, documented, and defaults to "included"
- ReportedMicrosoft Clarity added a scrape-to-referral ratio, per site and free
- ReportedThe AI referral barely exists in the data, again
- MeasuredShopify's own merchants: AI referrals are small, growing fast, and convert well
Read the issue in full → ChatGPT is moving the "Sources" button out of sight, which means the link a citation earns is worth less and the name in the answer is worth more.
- ReportedOpenAI is making ChatGPT's sources harder to find
- ReportedGoogle's AI Overviews for local results are citing self-promoting listicles
- ReportedGoogle amended its scraping lawsuit against SerpApi, this time around licensed content
- ReportedThe Search Console generative AI report reached many more sites, but not all of them
- ReportedBeing recommended shows up as branded search, not as AI referral traffic
- ReportedSocial citations, and YouTube overtaking Reddit
Read the issue in full → when a free ChatGPT account is asked about a business, the answer is not built from Google. It is built from OpenAI's own index, and the model sees about 200 characters of your page.
- ReportedChatGPT's retrieval was measured pipeline by pipeline, and the source engine changes with the tier
- ReportedBlocking the AI crawlers does not reliably stop the citation
- ReportedLocal visibility does not carry over from Google
Read the issue in full → Cloudflare is now selling AI visibility from inside the pipes, and a 50,000-prompt study shows the ad slot and the recommendation are two different things.
- ConfirmedCloudflare launched an AEO suite: Agent Readiness plus an AEO Visibility dashboard
- ReportedYelp reservations and waitlists now run inside ChatGPT
- ReportedChatGPT shows ads on a quarter of commercial prompts, and almost never cites the advertiser
- ReportedThe academic audit of ChatGPT ads
Read the issue in full → a site can be invisible to ChatGPT without anyone ever choosing to block it.
- ReportedAI Mode reached English speakers in 180+ countries this month
- ReportedSearch Console appears to be surfacing AI Mode queries
- ReportedPublishers still cannot opt out of AI Overviews cleanly
- Reported40.9% of the top 1,000 sites are unreadable to GPTBot, and a sixth of them block it by accident (baseline)
- Reported86% of AI citations come from brand-controlled sources (baseline)
- MeasuredOnly 38% of AI Overview citations come from pages ranking in the top 10 (baseline)
Read the issue in full → the assistants recommend whoever the referees rate, not whoever is biggest.
- ReportedChatGPT Ads adds conversion bidding, and sponsored carousels appear in answers
- ReportedThe Ninth Circuit says an AI agent browsing on your behalf is you
- ConfirmedGoogle's FAQ rich result retirement finishes this month
- ReportedA 2% broadband provider beat BT, Sky and Virgin Media combined
Read the issue in full → getting cited and getting recommended are two different things, and only one of them sells a coach's service.
- ConfirmedGoogle's generative AI report is impressions only, and there is now an opt-out toggle
- ConfirmedCloudflare's default AI crawler blocking starts 15 September 2026
- MeasuredCited is not the same as recommended (Semrush, with Kevin Indig)
- Measured97% of llms.txt files were never fetched by anything (older, June)
- ReportedAdding schema did not move AI citations (older, May)
Read the issue in full →