Issue · June 27, 2026
You Rank #1 on Google and Show Up in 3% of the AI Answer
This was the week AI search stopped being a frontier and started behaving like infrastructure. Semrush, now owned by Adobe, shipped an index built on 126 million real US AI prompts. Walker Sands studied 828 enterprise companies across 45 million queries and...
This was the week AI search stopped being a frontier and started behaving like infrastructure. Semrush, now owned by Adobe, shipped an index built on 126 million real US AI prompts. Walker Sands studied 828 enterprise companies across 45 million queries and found the median B2B brand shows up in just 3% of the AI Overviews where it already ranks on Google. ChatGPT's share of the AI assistant market fell below half for the first time, to 46.4%. And the fight over who gets cited forked three ways in five days: Getty signed a deal with OpenAI and its stock jumped about 90%, nearly 400 US newspapers sued OpenAI and Microsoft, and Google quietly told 3,000-plus publishers to hand over AI-training rights or lose their fee.
Here is the uncomfortable part. Most marketing teams are still measuring the wrong thing. They watch their Google rank go up and assume the AI answer follows. It does not. You can sit at position one and appear in the answer box almost never, because the engine is choosing sources by a different rulebook than the one you have optimised for the last fifteen years. This week the data finally made that gap impossible to argue with.
TL;DR
Walker Sands found the median B2B brand that ranks on Google appears in just 3% of the AI Overviews where it already shows up.
Semrush's expanded index analyses 126 million real US AI prompts across 22 industries, the clearest sign yet that AI visibility is now a measured category, not a theory.
ChatGPT dropped below 50% of the AI assistant market for the first time at 46.4%, with Gemini at 27.7% and Claude at 10.3%.
Getty signed an OpenAI display deal and jumped about 90% in a day, while nearly 400 US newspapers sued OpenAI and Microsoft the same week.
Half of US shoppers now fact-check AI product recommendations on Reddit before buying, and Google says gaming AI citations is officially spam.
1. Semrush Builds an Index on 126 Million AI Prompts. Visibility Is Now Measured
June 26, 2026 | Semrush, now an Adobe company, released its expanded 2026 AI Visibility Index built on 126 million real US AI search prompts across 22 industries and four AI platforms. That sits inside a global database of 261 million prompts, including 58.4 million from ChatGPT alone.
What I Think: This is the moment AI visibility stops being a vibe and becomes a number on a dashboard. Adobe did not pay $1.9 billion for Semrush to count blue links, they paid for this. When a tool can show a CMO exactly how often their brand appears in AI answers across 22 industries, the conversation in the boardroom changes from "is GEO real" to "why is our number lower than our competitor's." The measurement layer is here, and measurement is what turns a curiosity into a budget line.
Why It Matters: Once something is measured, it gets managed, and then it gets funded. If your rival can pull a report showing they are cited twice as often as you in your own category, you have a problem you can no longer wave away. Get your own baseline now, because the day your CEO sees a competitor's number, "we hadn't measured it" is not an answer that keeps your job.
2. B2B Brands Rank on Google and Appear in 3% of AI Overviews
June 24, 2026 | A Walker Sands study of 828 enterprise companies across 45 million March search queries found the median B2B brand is cited in only 3% of the AI Overviews where it already ranks organically. AI Overviews now appear on roughly half of those searches, and even top-quartile brands hit just a 4.5% citation rate.
Source: https://searchengineland.com/b2b-brands-rank-google-appear-ai-overviews-480954
What I Think: This is the single most important number of the week, and it should scare every head of marketing reading it. You have spent years and real money earning that first-page ranking, and the AI answer ignores it 97 times out of 100. Ranking and being cited are now two separate games with two separate rulebooks. The brands that keep reporting "we rank for our keywords" are reporting on a race that fewer and fewer customers are watching.
Why It Matters: Half your buyers are getting an answer before they ever see your link, and your link is not in that answer. This is not a B2B quirk, it is what happens in every category once AI Overviews show up on most searches. Audit your top commercial queries by the answer, not the ranking, and you will likely find the same 3% staring back at you.
3. Google Says Gaming AI Citations Is Now Spam
June 26, 2026 | Google finished rolling out its June 2026 spam update over roughly two days, its second of the year, applied globally and across all languages. The rollout added no new policy text, but Google's spam policies now explicitly cover attempts to manipulate generative AI responses, including buying or altering citations to game AI Overviews and AI Mode.
Source: https://www.searchenginejournal.com/google-begins-rolling-out-the-june-2026-spam-update/580424/
What I Think: This is Google drawing a line before the shortcut industry gets going. The second a category gets measured, a grey market appears to fake the metric, and Google is telling everyone the citation-buying playbook is a fast way to get burned. I read this as confirmation, not warning. If gaming AI citations were not already a real and growing problem, Google would not bother writing the rule. The honest path and the safe path are the same one here: be the best source, do not rent your way into the answer.
Why It Matters: Some vendor is going to pitch you "guaranteed AI citations" in the next quarter. This update is your reason to say no. The brands that win AI visibility the durable way do it by being clean, well-structured, and genuinely useful, not by buying mentions that one algorithm update can wipe out overnight.
4. The Content War Forks Three Ways: Sign, Sue, or Get Strong-Armed
June 24, 2026 | Nearly 400 US local and regional newspapers sued OpenAI and Microsoft for allegedly crawling their sites to train ChatGPT and Copilot without paying. The same week, Getty signed a multi-year display deal to show its licensed photos inside ChatGPT and its stock jumped about 90% in a day, while Google told its 3,000-plus publisher partners to grant AI-training rights or lose their annual fee.
What I Think: In one week the entire content-versus-AI fight settled into three lanes. You sign, like Getty, and the market rewards you. You sue, like the 400 papers, and you bet on the courts. Or you get strong-armed, like Google's publishers, told to give up your content or give up your cheque. There is no neutral lane left. The era of AI quietly taking your words for free is closing, and what replaces it is a negotiation where the side with cleaner rights and the stronger hand wins.
Why It Matters: If your brand publishes anything of value, original research, data, guides, you are now a supplier in a market you did not know you were in. Decide your stance before someone decides it for you. Know what AI engines are already taking from your site, and know whether you want to block it, license it, or feed it deliberately to show up in the answer.
5. ChatGPT Falls Below 50% Market Share. The One-Engine Strategy Just Died
June 16, 2026 | Sensor Tower's 2026 State of AI report shows ChatGPT's share of the global AI assistant market fell to 46.4% by the end of May, below half for the first time, down from over 50% in January. Gemini reached 27.7% with 662 million users and Claude hit 10.3% with 245 million.
Source: https://techcrunch.com/2026/06/16/chatgpts-market-share-slips-below-50-for-first-time/
What I Think: For two years "AI search" basically meant "ChatGPT," and you could almost get away with optimising for one engine. That shortcut is gone. When the leader is under half the market and Gemini is pulling 662 million users, optimising for one assistant means ignoring more than half your audience. The engines also cite wildly different sources, so showing up in ChatGPT does almost nothing for your Gemini or Perplexity visibility. This is fragmentation, and fragmentation is harder and more expensive to win.
Why It Matters: Your customers are spreading across at least three assistants that each answer the same question with different sources. A single-engine strategy now leaves the majority of buyers seeing an answer you are not in. Treat ChatGPT, Gemini, and Perplexity as three separate channels with three separate scorecards, the same way you would never collapse Google and YouTube into one number.
6. Half of US Shoppers Fact-Check AI on Reddit Before They Buy
June 22, 2026 | A Reddit survey of 13,956 US and 6,485 UK adults found half of US shoppers now verify AI-generated product recommendations on Reddit before completing a purchase. More than one in five shoppers add the word "Reddit" directly into their searches.
Source: https://ppc.land/half-of-us-shoppers-fact-check-ai-on-reddit-before-buying-survey-finds/
What I Think: People do not blindly trust the AI answer, and that is the most useful thing in this whole newsletter. They read the AI recommendation, then go to a human forum to check whether it is real. That means your AI visibility and your reputation on Reddit, reviews, and third-party sites are now one connected system. If AI recommends you but the forums quietly trash you, the sale dies at the verification step. The buying journey now has a built-in fact-check, and most brands are not watching it.
Why It Matters: Winning the AI answer is only half the job. The other half is making sure the places people verify that answer agree with it. Get your facts consistent across the review sites and communities your buyers actually trust, because the gap between what AI says about you and what humans find is where deals quietly fall apart.
7. AI Ads Go Mainstream: 2,000 Brands Now Buy Inside ChatGPT
June 22, 2026 | Criteo said more than 2,000 brands now run ads inside ChatGPT through its platform, double its early-May figure, with its Prompt Smart Ads driving roughly 4x higher spend and click-through rates two to three times higher than comparable formats. At Cannes the same week, OpenAI expanded ChatGPT ads to the UK, Japan and South Korea and opened a self-serve Ads Manager beta to UK advertisers.
Source: https://ppc.land/criteo-hits-2-000-brands-on-chatgpt-ads-as-prompt-smart-ads-show-4x-spend-lift/
What I Think: The free organic era of AI search is getting a paid sibling, fast. Two thousand brands in ChatGPT ads, doubling in under two months, is not an experiment anymore, it is a land grab. What strikes me is the claim that more than 80% of ChatGPT ad-driven traffic comes from new customers. That is the number that pulls budget out of tired retargeting and into the answer box. Paid AI visibility is arriving on exactly the timeline organic AI visibility did, and the early movers are quietly learning the format while everyone else waits for a case study.
Why It Matters: You now have two roads into the AI answer, earned and paid, and your competitors are already testing both. You do not have to spend heavily, but you do have to decide. Run a small ChatGPT ads test this quarter so you understand the format before it becomes a line item you are forced to defend without any first-hand data.
Travel Desk
Travel showed the rest of us where AI search is heading this week, because travel feels the pain first. Amadeus revealed that a single AI coding agent pointed at one airline's site pulled 881,076 fare combinations for one trip, the real cost of agents that never stop searching, and its fix is to precompute likely answers before the agent ever asks (Skift, June 23). A new Aven Hospitality white paper with Skift found 44% of travelers who use AI search now treat it as their primary source of travel insight, ahead of search engines and brand sites, yet only 11% of hotel organisations have deployed advanced AI agents to answer them. That gap, demand moving into the answer while supply stands still, is the exact pattern every CMO in every industry is about to live through. The travel brands that win will be the ones whose facts are clean enough for an agent to quote without double-checking. Full travel breakdown in this week's GeoTravel newsletter: Geotravel.ai.
The Strategic Reality Check
Step back from the individual stories and the week tells one story. AI search grew up. It got measured, with Semrush analysing 126 million prompts. It got audited, with Walker Sands proving a 3% citation rate and Google widening its Search Console AI reports. It got policed, with Google calling citation-gaming spam. It got litigated, with 400 newspapers in court. And it got monetised, with 2,000 brands buying ads inside ChatGPT. Every one of those is a sign of a category leaving its experimental phase and entering its accountable one. The window where you could call AI search "too early to measure" closed this week.
The trap is comfort. Your Google rankings still look fine, your traffic dashboard is still green, and nothing in your old reports is screaming. That is exactly the problem. The old reports are measuring a race fewer customers are running, while the new answer surfaces, where buyers actually get their recommendation, stay invisible to you. Comfort with the old number is the most expensive position in the room right now.
Three truths that should keep you up this week:
1. Ranking on Google no longer means appearing in the answer. The median B2B brand that ranks shows up in just 3% of its own AI Overviews.
2. There is no single AI search to win. With ChatGPT below 50%, a one-engine strategy now misses most of your audience by definition.
3. Visibility is becoming a market you pay into, not just earn. Getty signs, 400 papers sue, Criteo sells ChatGPT ads to 2,000 brands. Free organic citation is no longer the only road.
What You Should Do This Week
1. Measure the gap, not the rank. Take your top 20 commercial queries and check by hand whether you appear in the actual AI answer, not just the blue links below it. The space between "we rank" and "we are cited" is your real visibility problem, and right now you almost certainly cannot see it.
2. Split your scorecard by engine. Test ChatGPT, Gemini, and Perplexity separately. They cite different sources, so one combined "AI visibility" number hides where you are winning and losing. With no engine above 50%, you cannot afford to optimise for one and hope.
3. Feed the verifiers. Half your shoppers fact-check AI answers on Reddit and reviews before buying. Make your facts consistent on the third-party places both AI and humans cross-check, because that is where the sale is won or quietly lost.
4. Refuse the citation-gaming pitch. Someone will sell you "guaranteed AI citations" this quarter. Google now calls that spam. Win the answer by being the cleanest, best-structured, genuinely useful source, not by renting mentions one update can erase.
5. Run one paid-AI test. ChatGPT ads are live in more markets with 2,000 brands already in. You do not have to commit a big budget, but you do have to learn the format first-hand before it becomes a line you are asked to defend with no data of your own.
AI search did not arrive this week. It grew up this week. Measured, policed, litigated, and for sale.
Still treating AI search as one channel you can rank your way into? Still calling it the winner by your Google position while you vanish from the answer? Hit reply and tell me why.
Want to see exactly where you appear in AI answers before your competitors check theirs? Let's talk: https://calendly.com/vincent-getinference/
See you next week.
Vincent Chief AI Enthusiast, SimplyAI: AI SEO (GEO)