OpenAI began serving ads inside ChatGPT across 31 European markets, France included, on 24 August. That follows roughly six months of testing ads with users in the United States. The answer box now has a paid lane, and for anyone who spent the past year rewriting pages to get quoted by AI, this is the moment the ground moves.

The details matter for planning, so here is what is actually live. Ads appear only for people on the free tier and the low-cost Go plan at 7.99 euros a month. Plus, Pro, Enterprise, Business and Education accounts stay ad-free. At launch the ads are not personalized: selection draws on the current conversation topic, approximate location, device type, time of day and language, and OpenAI says past chats and stored memories are not used yet. Booking runs through the OpenAI Ads Solutions team and agency partners for now, with self-service coming later. You can read the specifics in OpenAI’s own announcement on ChatGPT ads in Europe.

Why this changes the math for organic visibility

Picture a small B2B software founder who spent months getting cited when people asked ChatGPT for tools in her category. That mention was earned. It cost her time, not money, and a better-funded rival could not simply outbid her for it. Once a paid lane sits next to the organic answer, that protection weakens. Her competitor can now buy proximity to the exact question she worked to own.

The useful way to read this is not panic. It is timing. Right now the ad inventory is thin, unpersonalized, and locked to the free and Go tiers, while the paying, higher-intent users on Plus and Pro still see clean answers. That gap is the window. Earned visibility inside AI answers just shifted from a tactic to something closer to a moat, and moats are cheaper to dig before everyone agrees they matter.

What to actually do this month

Tighten the ground the models already read from. Make sure your brand, product names and category claims appear in clear, structured language on your own pages, not buried in marketing copy. This connects to a pattern we covered in why long, conversational queries changed which pages get quoted: the systems reward pages that answer a specific question plainly. Audit which questions in your category you already win in ChatGPT and Gemini, and write down where you appear and where you vanish. You cannot defend a position you have not measured.

Here is where honesty helps. The risk is over-reacting to a launch that is still small. Rushing to book ads through an agency before the self-service manager exists, with no personalization and limited reporting, may not be worth it if you run one product and a modest budget. Early ad inventory in a new format tends to be expensive attention with fuzzy attribution. This may not be worth it if you cannot yet tie a ChatGPT impression to anything you can measure.

What can go wrong is treating the paid lane as the whole story and quietly letting your organic footing rot. The brands that will struggle are the ones that stop investing in being genuinely quotable because they assume they can buy their way in later. By the time buying in is easy, it will also be crowded and priced accordingly.

The one move worth making today is unglamorous: pick your ten most valuable buyer questions, ask the major assistants each one, and record whether your brand shows up. That list is your baseline. Everything you decide about ads, content and budget over the next quarter should be judged against whether it moves those ten answers.

Editor’s note: This area changes quickly, so check the latest platform policy before making compliance decisions.

The post ChatGPT Just Added Ads. Your Free Ride in AI Answers Has a Deadline. appeared first on Mark8ng.com.

Here is the more useful one. The most common opening words in AI Mode queries are now what, how, I, is and can. Not nouns. Not product categories. Question words and pronouns. People are describing a situation and asking for help with it, the way they would ask a colleague who happens to be standing nearby.

Five verbs replaced the intent buckets

Google groups the new behaviour into five modes: explore, decide, learn, create and do. They are growing at noticeably different speeds. Explore queries, the open-ended brainstorming kind, are running around 30% ahead of AI Mode’s overall growth. Decide queries, built around comparison phrasing, about 40% ahead. Do queries, the planning kind covering everything from workout routines to household budgets, roughly 80% ahead over the past six months.

If you have ever built a content plan around the informational, commercial and transactional split, this is that framework being quietly replaced by something with more resolution. A comparison page and a getting-started guide can share a target keyword and still need completely different structures.

What most people will take from this, and why it is only half right

The advice going round is to put the answer first. That is correct, and it is not enough.

Answer-first at the page level is a 2015 idea in a new hat. Featured snippets taught everyone to front-load the intro paragraph, and plenty of sites did exactly that. What has changed is the unit of retrieval. AI Mode does not pull your page. It pulls a passage. Which means every H2 section needs to survive being lifted out on its own, with no surrounding context, and still make sense to somebody who never read your introduction.

Test this on your own best-performing article. Copy one middle section, paste it into a blank document, and read it cold. If it opens with something like “this is why the previous point matters”, or leans on a term you defined four hundred words earlier, it is not retrievable. It is prose that depends on being read in order, which is a fine thing to write and a bad thing to expect a machine to quote.

A before and after that takes two minutes

Take a small accountancy firm with a page called “Choosing a business bank account”. One section is headed “Fees and charges”, followed by three paragraphs building carefully towards a conclusion about monthly cost.

The retrievable version contains identical information with the order inverted. The heading stays plain. The first sentence states the actual finding: that most UK business accounts sit between zero and roughly fifteen pounds a month, and that the free tiers usually cap the number of free transactions. Then the nuance. Then the exception that catches people out.

Nothing was dumbed down and nothing was cut. The sentence order changed, and a section that previously needed the paragraphs above it now stands on its own feet.

The follow-up question is where the opportunity is

This is the part almost nobody is acting on. Follow-up questions in AI Mode are climbing quickly, and more than one in six searches now include something other than text, whether that is an image, voice, or a live back-and-forth exchange.

In practice that means your page should answer the second and third question, not only the one in the title. Somebody asking about business bank accounts asks next about switching, then about what happens to existing standing orders. If those three answers live on three separate thin pages built for three separate keywords, you are competing against yourself inside a system that would rather pull three passages from one source it already trusts.

One useful comparison doing the rounds is the telegraph, which pushed journalists into lead-first writing in the 1880s for reasons of cost and physical page constraints. The mechanism is different now. The pressure is the same.

Where this goes wrong

The failure mode is already visible: sites turning every article into a stack of disconnected answer blocks with no argument, no point of view, and nothing a person would choose to read. That is optimising for the retrieval system and losing the reader, which is a poor trade, because the reader is the one who buys something.

There is also a sourcing caveat worth stating plainly. This is Google’s own data, about Google’s own product, published by Google. The direction of travel is almost certainly right and matches what most people see in their own logs. The specific growth percentages are marketing, and should be treated that way.

And there is a cost question. Restructuring a two-hundred-page site for passage retrieval is a project, not an afternoon. For a small team I would take the ten pages that already earn something, fix those, and leave the rest alone until there is evidence it worked. Proving it worked is a separate problem, and counting mentions is not the same as counting outcomes.

Start here, not with a migration

Most sites do not have a content volume problem. They have a content order problem.

Pick your three highest-value pages. Read every H2 section in isolation and rewrite the opening sentence of any section that cannot stand alone. Nothing else. No migration, no new tooling, no rebuild, no consultant.

If it moves nothing in six weeks, you have lost an afternoon and learned something specific about your own site, which is more than most people get from reading another framework.

Editor’s note: This area changes quickly, so check the latest platform policy before making compliance decisions.

The post People Are Typing Sentences Now. Your Pages Still Answer Keywords. appeared first on Mark8ng.com.

That gap is the actual state of AI search measurement in 2026. Budget is moving. Proof is not.

Why the mentions are easy and the outcomes are hard

Mention counting is a solved problem. You run a prompt, you read the answer, you check whether your brand is in it. Vendors can do this across thousands of prompts and charge for the dashboard.

Outcomes are hard because the click that follows an AI answer usually carries no useful referrer. Someone reads a ChatGPT response naming three vendors, opens a new tab, types your brand name into Google, and lands on your homepage. Your analytics records that as branded organic or direct. The AI answer that created the demand is invisible, and no amount of prompt tracking will surface it. We wrote about the version of this problem where the citations exist but the traffic never arrives, and the measurement question is the other half of it.

So AI visibility tools stop at mentions. That is not laziness. It is the boundary of what the data allows.

The smallest measurement setup that is honest

You do not need a platform to start. You need three things running on a fixed weekly rhythm.

A frozen prompt set. Write twenty to forty prompts a real buyer would type: “best invoicing tool for freelance designers”, “alternatives to [competitor]”, “is [your category] worth it for a team of five”. Freeze the wording. Run them weekly across the two or three assistants your buyers actually use. Record which brands appear and which sources get cited. Freezing the wording matters because a changed prompt produces a changed answer, and you will not be able to tell whether you improved or just asked differently.

A citation log. When an AI answer cites a page, note the URL. After six weeks you will see which of your pages the models trust, and it is usually not the ones you would have guessed. Comparison pages, pricing pages and plain documentation tend to get cited more often than the thought leadership posts most teams keep publishing.

A weekly traffic baseline, split three ways. Track branded organic search, direct traffic and any identifiable AI referrers as three separate numbers. You are not trying to attribute anything. You are watching whether the weeks where your visibility improved are followed, two to four weeks later, by movement in branded demand. That is weak evidence. It is also more evidence than a mention count.

What the Shopify numbers actually suggest

Shopify released second-quarter commerce data this month showing AI-referred sessions to merchant storefronts up 197% year over year, with organic search traffic also up 12% on a much larger base. In categories where shoppers compare specifications before buying, AI-referred visitors converted at roughly twice the rate of organic visitors, and AI brought in about 1.3 times more first-time customers.

The detail worth stealing is quieter. Shopify found that AI-referred shoppers converted at twice the rate when the AI system used structured catalog data rather than scraped third-party product feeds. What moved the number was not clever prompt work. It was making the underlying product data machine readable and correct. Shopify’s breakdown of AI and organic search behavior by category is worth reading before you buy any AI visibility software, because it points at the unglamorous work first.

One caveat on that data. Shopify did not disclose how many merchants or transactions were in the analysis, and it is a platform reporting on its own ecosystem. Treat the direction as useful and the precise multiples as indicative.

When this is not worth doing

If your category attracts almost no AI query volume, a weekly prompt run is busywork. A local plumber covering three postcodes will learn more from a month of call tracking than from watching ChatGPT describe plumbing. The same applies to businesses whose buyers arrive through referral or repeat purchase rather than search.

There is a real failure mode in the other direction too. Teams that start measuring AI visibility often begin publishing content designed to be quoted by models, and quietly stop publishing anything a human would choose to read. The citation count rises. The pipeline does not. If your content calendar has drifted entirely toward listicles and comparison tables over the last quarter, you have already made that trade without deciding to.

Be careful with the tool category itself. Most AI visibility products are eighteen months old at most, their share-of-voice scores are not comparable between vendors, and switching resets your history. Run the manual version for a quarter first, so you can judge whether a tool is telling you something your spreadsheet did not.

The uncommon part

Most teams treat AI visibility as a ranking problem and buy a tool that reports a rank. The more useful framing is that it is a data quality problem wearing a marketing costume. The brands getting described accurately tend to be the ones whose pricing sits on a page instead of behind a form, whose specifications are structured, and whose category claims are stated plainly enough that a model can repeat them without inventing anything.

So start there. If you want to measure one thing this week, do not measure your rank in AI answers. Take the ten facts about your business that matter most commercially, ask three assistants to state each one, and count how many come back wrong. Fix those. The rank tends to follow, and unlike the rank, you can prove you changed it.

Editor’s note: This area changes quickly, so check the latest platform policy before making compliance decisions.

The post Your AI Visibility Tool Counts Mentions. It Cannot Tell You If They Worked. appeared first on Mark8ng.com.

This has a name now. Kevin Indig called it the ghost citation, and a study he ran with Semrush put a number on how common it is. Across 3,981 domain appearances from 115 prompts, run in 14 countries across ChatGPT, Gemini, Google AI Overviews and Google AI Mode, 61.7% of appearances were citations with no brand mention in the answer. Only 13.2% got both. Put differently: 74.9% of appearances included a citation, but only 38.3% included the brand name.

Your content can be doing all the work while your brand stays completely invisible to the person reading.

The engines are not doing the same job

The split between engines is severe enough that a single visibility score is close to meaningless.

• Gemini names the brand in 83.7% of appearances but generates a citation link only 21.4% of the time. It answers from what it already knows about you.
• ChatGPT does the reverse: cites 87% of the time, names the brand in 20.7%. It reads like a paper with footnotes.
• Google AI Overviews sit between the two and lean towards citing.

In the study’s data there was almost no overlap between the brands ChatGPT cited and the brands Gemini named for the same prompt. On 100 of 454 prompt and domain combinations, the engines disagreed on whether to name the brand at all. So if you are tracking one engine and reporting it as “AI visibility”, you are reporting one behaviour of one system.

The content type doing the damage is the one you produce most

This is the finding that should change what a small marketing team publishes next month.

Informational content, the “what is”, “how does”, “explained” format that most business blogs are built from, earned an 89.3% citation rate and an 18% mention rate. Comparative content, the “best”, “versus”, “recommended” format, produced a 43.3% mention rate, roughly 2.4 times more brand mentions. How-to content sat at 42.8%. Commercial queries reached 35.6% mentions against an 84.4% citation rate.

The logic is uncomfortable but obvious once you see it. Informational content is raw material. The model reads it, absorbs the fact, writes the fact in its own words, and links the source out of politeness. Comparative content forces the model to name things, because a comparison without names is not a comparison.

So the agency that spent eighteen months building a definitive glossary of industry terms has produced the single most citable and least brand-visible asset available. Meanwhile the competitor with six honest “X versus Y” pages, including the cases where they lose, gets named.

Why the model will not repeat your claim

There is a second layer under the format question, and Forbes framed it well this week: every brand claim an AI repeats without independent backing is a claim the AI is vouching for itself. Systems tuned against Google’s quality rater guidelines weight first-hand experience heavily, and a product page describing your own jacket as the warmest available has an obvious incentive problem. A forum thread describing a Colorado winter in that jacket does not.

This is where most GEO advice goes vague, so here is the concrete version. If your only evidence for a claim lives on your own domain, expect the model to use the surrounding information and skip the claim. Corroboration that actually shifts this looks like: a named customer willing to be quoted with specifics, a third-party review platform with enough volume to be more than a testimonial page, a trade publication write-up you did not pay for, forum and community threads where customers describe outcomes in their own words, and comparison content that names competitors honestly enough to be credible.

None of that is fast. That is the point. There is no schema markup fix for a claim nobody else has ever repeated.

What to actually do

A sequence that fits inside a normal fortnight:

1. Pull the ten prompts that matter most to your business. Run each one in ChatGPT and Gemini separately, and record two columns: cited, and named.
2. Find the pages that are cited but never named. These are your ghost pages, and they are usually your best informational content.
3. For the top three, add a comparison or evaluation section that names real alternatives, including where you are not the right choice.
4. Pick one claim you make constantly and find or build one piece of third-party evidence for it that is not a testimonial on your own site.
5. Re-run the ten prompts in six weeks. Compare the named column, not the cited column.

Query phrasing matters too, and it cuts against most keyword habits. Short conversational prompts produced brand mention rates near 100% in the study, while long structured prompts on the same topic produced 2% to 3%. Test the short version of your question, because that is how people actually ask.

What can go wrong here

The obvious risk is over-correcting. If you decide informational content is worthless and stop making it, you lose the citations that build the model’s familiarity with your domain in the first place. Citations and mentions come from different places: citations reflect domain authority and original work, mentions reflect brand familiarity and positioning. You need both, and killing one to chase the other is how sites lose ground in both.

The second risk is measuring this too often. Answers vary run to run. A single check on a single day tells you almost nothing, which is exactly the trap the sample-of-one approach falls into. Six-week intervals, ten fixed prompts, two engines minimum.

And this may not be worth your time at all. If you sell into a market where buyers do not use assistants for discovery, if your pipeline is referral and relationship driven, or if you are a publisher whose business model runs on the citation link rather than the name, then citation rate is the correct metric and the ghost citation is not your problem. Check who your buyers actually are before rebuilding a content plan around this.

The one thing worth taking away: stop treating a citation as proof of visibility. It is proof your content was useful to a machine. Whether it was useful to your business is a separate question, and it has a separate column in the spreadsheet.

The full dataset is in the ghost citations study, and the trust argument is set out in this piece on AI search and branded content. If you want the other half of the picture, what these systems can and cannot read on your site, we covered that in how AI builds a B2B shortlist before anyone calls you.

Editor’s note: This area changes quickly, so check the latest platform policy before making compliance decisions.

The post AI Search Is Citing Your Blog and Not Saying Your Name. Here Is What Changes That. appeared first on Mark8ng.com.

On 6 August OpenAI said it is removing limits on text chats and making GPT-5.6 Luna the default model for Free and Go accounts, replacing GPT-5.5 Instant. Free and Go users also get a Think button that routes harder questions through more reasoning. Plus and Pro accounts move to an updated GPT-5.6 Sol with a slider controlling how much thinking goes into an answer. Limits on files, images, voice and image generation stay in place. The announcement arrived shortly after ChatGPT passed a billion weekly users, with unlimited text chats for free accounts scheduled to land the following week.

The number worth holding onto is not the billion. It is the split.

Two tiers, two models, two answers about you

Until now, free accounts had a ceiling. Hit it and the conversation stopped, or dropped to a weaker fallback. That ceiling quietly rationed how many people used ChatGPT for the kind of open-ended question that mentions brands: which accounting tool suits a sole trader, who does commercial kitchen servicing near Leeds, what is the difference between two CRMs. Remove the ceiling and that behaviour has nowhere to stop.

So the model answering most brand questions is now GPT-5.6 Luna, and almost nobody in marketing has tested against it. OpenAI’s own evaluation puts factual errors 62% less common in Luna than in GPT-5.5 Instant, and 68% less common in Sol. Take those as directional rather than settled, since they are internal numbers with no published methodology. But the gap between the two figures is the point. Luna is better than what free users had. It is still not the model you are testing in.

A worked example. A regional B2B supplier checks how ChatGPT describes them, using the founder’s Plus account with the thinking slider left high. The answer is detailed, cites two trade publications, mentions a certification they hold. Reassuring. A buyer on a free account asks the same thing, gets a shorter reply from a different model, and the certification does not appear because the source it lives on was not retrieved. Nothing about the company changed between those two answers. Only the tier did.

What to actually do this week

Run the check twice, from two accounts, and write down which model produced each answer. That is the whole method. It takes about twenty minutes and it is the only way to see the gap.

Three questions are usually enough: your brand name alone, your main category with a location or qualifier attached, and a comparison question naming a competitor. Ask each in a fresh conversation. Log the date, the tier, the model, and whether any link was offered. Repeat monthly rather than daily, because the answers move around enough that a single day tells you very little.

If the free-tier answer is thinner, look at what the paid answer used that the free one did not. Usually it is a source that takes more retrieval effort to reach: a PDF, a page buried three clicks deep, a claim that only exists on a third-party directory. The fix is rarely clever. It is putting the fact somewhere shallower.

What can go wrong here

Three failure modes, and the first is the common one.

Treating a handful of answers as data. You are running a sample of two on one afternoon. Chatbot output varies between sessions for reasons that have nothing to do with your website. If you rewrite a page because of one bad answer, you will be rewriting it again next month. We covered the broader version of this problem in how AI describes your brand across many queries rather than one, and the same caution applies here.

Chasing model changes as a strategy. Default models will change again. Building a workflow around Luna specifically is building on something with a short shelf life. What lasts is the habit of checking both tiers, not the model name.

Assuming unlimited means used. Removing a cap raises the ceiling on usage. It does not prove your buyers were hitting that cap in the first place. If you sell into a market that researches through trade bodies, procurement portals or a rep they have known for years, none of this may show up in your pipeline for a long time.

When this is not worth your time

If your business runs on repeat customers and referrals, and search has never driven meaningful revenue, this is a twenty minute curiosity rather than a project. The teams that should take it seriously are the ones where a stranger comparing three options is a normal way to get discovered. For everyone else, knowing the gap exists is enough for now.

The uncomfortable part of this update is not the model change. It is that the AI answer most of your market sees has always been a different answer from the one you have been reading, and the tier that produces it just got a much bigger audience. Check both. Then decide whether the difference is worth fixing.

Editor’s note: This area changes quickly, so check the latest platform policy before making compliance decisions.

The post ChatGPT’s Free Tier Just Went Unlimited. You Have Been Checking the Wrong Model. appeared first on Mark8ng.com.

One survey put the share of UK senior decision makers using ChatGPT, Copilot or Perplexity somewhere in their procurement process at 66%. Treat the exact number with caution, because self reported tool usage surveys run high almost by design. The direction is harder to argue with. The stretch of the buying process that used to take a week of searching and comparison now happens in a single conversation, and it ends with a shortlist and a recommendation on who to call first.

The failure mode is not bad content

Suppliers that surface reliably are the ones whose credentials sit in text a machine can read. Suppliers that do not surface are often just as capable, with the same certifications and better outcomes, filed in PDFs, gated portals and slide decks. Analysts describing AI as the first gatekeeper in procurement keep landing on the same point: the gap is structural, not qualitative.

Take a mid sized industrial supplier with fourteen years of certified process and a customer list any competitor would envy, all of it inside a capability brochure behind an email form. To a model building a shortlist, that company has a homepage, a contact page and nothing to say. The competitor with a weaker product and a plain HTML certifications page takes the slot.

Four things to move into text this week

Most exclusions trace back to the same four gaps. Pricing, or at minimum a pricing model and a range. Integrations, named individually rather than shown as a logo wall image. Certifications and compliance, with the issuing body and the year. Case study outcomes, with the number written in the sentence instead of sitting in a downloadable file.

None of that needs new content. It needs existing content moved out of assets and onto pages. The test takes five minutes: open your site with images and JavaScript disabled and read what is left. That is roughly what the model gets.

The cost that comes with being findable

Being easier for AI to find also means being easier for AI generated noise to find. Suppliers are already reporting inbound inquiries routed automatically to every name on a category list, arriving in a form that looks like qualified interest and is not. A two person sales team can lose a week to that before anyone spots the pattern.

Decide the filter in advance rather than after. No company domain, no specific requirement, and a timestamp that matches several competitors to the minute is usually the whole tell.

When this is not your problem

If your pipeline runs on referrals, or you sell into a market where four buyers know every supplier by name, none of this changes your quarter. Machine readability matters most when a buyer does not already know who you are, which is another way of saying it is a top of funnel problem. If you do not have one, spend the time elsewhere.

The risk on the other side is treating this as a content project. Publishing twelve more blog posts does not put you on a shortlist if your pricing is still a contact form. The order matters more than the volume.

Editor’s note: This area changes quickly, so check the latest platform policy before making compliance decisions.

Write down the five questions a buyer would ask an assistant before shortlisting a company like yours, then run them. If you are not named, compare what the named companies publish in plain text against what you publish in a PDF. If you want a data source rather than a manual log, Microsoft now reports which topics AI links to your brand. That gap is usually the entire answer.

The post AI Builds the Shortlist Before Anyone Calls You. Here Is What It Can Read. appeared first on Mark8ng.com.

What buyers are actually saying

Sam Huston, svp of media at Dept, said there is not a client who would say no to testing AI visibility ads, given how disruptive GEO has already been to some businesses. But Stephan Kopp, managing director of Mediaplus Performance, argued the opposite: you can try to influence a large language model with paid content, he said, but it will not happen, and it is only a matter of time before AI developers tune their systems to ignore the workaround, the same way Google has repeatedly rewritten its search algorithm to close loopholes. A WARC study cited in the same reporting estimated that 63% of a brand’s visibility in AI answers comes from long-term brand equity, and only 26% from current marketing activity. If that number holds up, most of what a GEO vendor is selling you is a way to move a small slice of a much bigger picture.

The tools are shipping faster than the proof

None of that skepticism is slowing the vendor side down. Onclusive launched GEO Analytics in the US this week, tracking how brands appear across five AI engines, Gemini, ChatGPT, Grok, Claude, and Perplexity, and scoring which sources most influence what those engines say. It joins a growing list that includes Scrunch and Profound, all promising a dashboard where a founder or marketing lead used to have a spreadsheet and a lot of manual searching. The pitch is reasonable. The open question, the one none of these tools can fully answer yet, is whether what they are measuring actually predicts anything about revenue.

What you can check before paying anyone

A two-person marketing team at a regional B2B services company does not need a subscription to start. Pick the ten questions a prospect would plausibly type into ChatGPT, Claude, or Perplexity before hiring a company like yours, run them once a week by hand, and log which brands get named. Do the same for your closest three competitors. After four weeks you will know, without paying for a platform, whether your brand shows up at all and whether that changes week to week. That log is also the only way to sanity check a vendor’s dashboard later, because you will already know what the honest numbers look like.

When the paid version is worth it, and when it is not

A GEO tool earns its cost once you are tracking more brand and competitor combinations than a person can reasonably check by hand, or once someone above you wants a number to report every month regardless of what it means. This probably is not worth paying for if your marketing budget is still in four figures a month and you have not run the manual version for at least a month first. The risk is buying a dashboard before you know what question you are trying to answer with it, and then mistaking a rising citation count for a rising number of customers. Those are not the same thing, and nobody selling the dashboard is incentivized to remind you of that.

Editor’s note: This area changes quickly, so check the latest platform policy before making decisions based on AI visibility data.

If you already run Microsoft’s free topic-level AI visibility data, start there before you evaluate a paid tool. It will not give you the polish of a GEO Analytics dashboard, but it is free, and free is the right price for a category this unproven.

The post The GEO Tool Market Is Booming. Media Buyers Still Are Not Sure It Works. appeared first on Mark8ng.com.

That calculation is meaningless. It is worth understanding why before anyone sets a quarterly target on it.

What the report actually is

Google announced dedicated generative AI performance reports for Search and Discover on 3 June 2026 and has been rolling them out to a subset of site owners since. The Search version shows impressions from AI Overviews and AI Mode, broken down by page, country, device and date. Experimental Search Labs features are excluded.

What it leaves out is the longer list: queries, clicks, click-through rate, average position, where in the answer your link sat, which passage Google used to build the answer, and anything resembling revenue. In the Search Central announcement Google said it is still working with site owners to decide which metrics to add.

So the report answers one narrow question. Which of your pages appear inside Google generative features, and how often.

Why the share calculation fails

AI impressions divided by total impressions is your own data measured against your own data. There is no market in the denominator. It cannot tell you whether a competitor appears more often, whether query volume grew, or whether AI Overviews simply triggered more this month than last.

There is a second problem. The two impression types are not the same unit. In classic search an impression usually means a listing was put in front of someone for consideration. Inside an AI Overview, a link only counts once it has been scrolled or expanded into view, and the product on the page is the synthesised answer, not your listing. Adding the two together gives you a number, not a meaning.

One more quirk, before anyone accuses Search Console of failing at arithmetic. The chart aggregates at property level, so two of your URLs appearing in a single AI response can count as one impression. The page table may credit each URL separately. Page-level totals will not always add up to the chart total. The dimensions are different, that is all.

The comparison that is actually useful

Here is the version that changes a decision. Take a kitchen fitting company with roughly forty pages. Export the AI-visible URLs for a full month, then pull classic impressions, clicks and average position for the same URLs across the same dates. Put them side by side and look for the mismatches.

Two groups usually fall out. The first is pages that rank well and pick up almost no AI impressions. These are often service pages written as persuasion, where the actual answer to the underlying question sits on line nine behind three paragraphs of positioning. The second group is the interesting one: pages with modest rankings taking a disproportionate share of AI impressions. For that kitchen fitter it was a single page explaining how long a fitted kitchen takes to install, with the answer in the opening sentence and a table of timings underneath. It ranked fifth for a low-volume query and was the most used page on the site.

That is the finding worth acting on. Not “our AI share is 12 percent”, but “the pages Google can lift a clean answer from all look like this, and most of ours do not”.

What can go wrong

The most common error is treating one revision as proof. Change a heading, watch impressions rise three days later, announce that you have reverse-engineered AI Mode. Demand moves, seasonality moves, competing sources move, and Google changes its own systems constantly. A sustained rise over several weeks following a substantial rewrite is evidence. A three-day bump is weather.

The second error is a blended click-through rate across both impression types. You can produce the number. It does not describe anything a human being did.

The third is reaching for the off switch. Google is testing a control that lets a site exclude itself from AI Overviews and AI Mode without leaving classic search. Opting out forfeits both the impressions and whatever traffic those features were sending. Get a baseline first. We looked at how far AI Overview coverage now reaches and the honest answer is that most sites do not yet know what they would be giving up.

When this is not worth your time

If your site collects a few hundred organic impressions a month, this report will not tell you much yet. The sample is too small, and because the rollout is partial an empty report may mean you are not included rather than that you are invisible. Spend the hour on the pages instead.

It also depends on what you sell. If your buyers research through AI assistants before they ever open a browser tab, appearing in a synthesised answer matters even when nobody clicks. If you are a local trade found through a map listing, it matters considerably less, and the time is better spent on reviews.

Editor’s note: This area changes quickly, so check the latest platform policy before making compliance decisions.

The practical next step is small. Export one month of AI-visible URLs, put classic performance next to them, and find the three pages that overperform. Then open those pages and work out what they have in common. That is a morning of work, and it will tell you more than any visibility score ever will.

Update, 7 August 2026: AI Mode queries are in Search Console, just not where you would look

A reporting detail has surfaced that sharpens the point above. AI Mode queries do appear in Search Console, but they sit in the main search performance report, not in the generative AI reports. Google’s John Mueller confirmed the split. The generative AI reports carry impressions, pages, countries and devices, with no query data attached, while the query-level view of AI Mode activity is folded into the standard performance report where it is hard to separate from everything else.

In practice that means the AI impressions number is even less self-contained than it first appeared. You cannot open the generative AI report and ask which questions triggered you, and you cannot open the main report and cleanly isolate AI Mode. The workaround people are using is to look for conversational, question-shaped queries in the main report, which is a heuristic rather than a filter. Treat anything you conclude from it as a hint, not a measurement.

The advice in the original piece stands unchanged. Export the AI-visible URLs, put classic performance beside them, and look at pages rather than numbers. If anything, the reporting split is another reason not to build a target around the impressions figure this year.

The post Google Now Shows Your AI Search Impressions. Here Is What That Number Does Not Mean. appeared first on Mark8ng.com.

Here is what the control actually does, what it quietly takes with it, and why most small businesses arguing about this are arguing about the wrong decision.

What the setting covers

The control lives under Settings in Search Console, labelled Search generative AI. Google is rolling it out to a subset of verified properties, so it may not be in your account yet. Google’s own Search generative AI control documentation is the only source worth trusting on the specifics, because the behaviour has already changed once.

Leave it on the default and your content can appear as links inside AI Overviews, AI Mode, and the generative features in Discover, and it can be used to ground those answers. Switch it to exclude and your site drops out of all of that, links included. Google says it began respecting these changes on 17 June. Processing takes a few days, then content usually falls out within a day or two, though caching can stretch that.

What it does not do matters just as much:

• It is not a ranking signal. Your position in regular search results should not move.
• It does not override Merchant Center or Google Ads. Shopping participation stays a separate decision.
• It does not stop your content being used for AI training. That runs through Google-Extended in robots.txt, which is a different lever entirely.

Those three distinctions catch people out constantly. If your goal is to keep your writing out of model training, this toggle does nothing for you.

The part nobody mentions in the setup guide

Google treats AI features and regular results as separate things. Google’s actual results page does not always agree.

Top Stories carousels have started rendering inside AI Overviews. John Shehata of NewzDash put the rate at roughly 1 in 6 US trending news queries, or 15.5% of tracked results where Top Stories appeared, with the UK figure at 17.46%. His reading is that opting out of AI features will likely pull publishers out of Top Stories when the carousel sits inside an AI Overview.

Two caveats worth holding onto. NewzDash sells news visibility tracking, so it has commercial interest in the finding, and it has not published sample sizes or collection dates. Google has not confirmed how the control handles a traditional feature displayed inside an AI feature. Search Engine Journal covered the full detail of the opt-out trade-offs if you want the longer version.

Still, the principle holds even if the number moves: a property-level switch can remove placements that were never described as AI placements.

You cannot currently measure what you would lose

This is the real problem with the decision.

The generative AI performance report, also rolling out to a subset of accounts, shows impressions from AI features broken down by page, country, device and date. Impressions only. No clicks. No queries. Your analytics can show Google organic referrals, time on site and conversions, but it cannot separate a visit that came from an AI Overview from one that came from a blue link.

So the maths is: unknown traffic value, unknown conversion value, against a change you can reverse but only after several days of processing and caching. The UK’s Competition and Markets Authority has required Google to supply click data and click-through rates, with most of those measures starting in December, and page-level controls by March 2027. Until then you are guessing.

A worked example

Take a six-person accountancy firm in Manchester running a blog on tax deadlines and allowances. Their content is exactly the kind that gets summarised: short factual answers with numbers in them. The partner reads about AI Overviews eating clicks and wants to opt out.

Before touching anything, the useful questions are:

1. Does the property even have the control yet? If not, the decision is theoretical.
2. What does the generative AI report show in impressions, and how does that compare to total search impressions? If AI features are 3% of impressions, this is not where the business is won or lost.
3. What is a blog visit actually worth here? For that firm, a tax-deadline article is not monetised by the pageview. It is monetised when someone reads it, trusts the firm, and calls. Being cited inside an AI answer with a link still supports that. Being absent supports nothing.

That third point is where most of these conversations go wrong.

The uncommon takeaway

The opt-out is a publisher’s tool wearing a webmaster’s clothes.

It makes sense if your revenue comes from impressions on your own pages, because in that model an AI answer that satisfies the user is a direct substitute for your business. That is the situation news publishers are in, and it is why the CMA forced the control into existence.

Most small businesses are not in that model. They sell services, software or products, and their content exists to get them in front of someone who might buy. For those sites, opting out removes a discovery surface and protects nothing measurable. The honest recommendation for the large majority is to leave it alone and spend the same hour on getting cited inside AI answers instead.

What can go wrong

Three things, in rough order of likelihood.

You lose placements you did not know were connected. The Top Stories pattern is the known case. There may be others that nobody has tracked yet, because the boundary between AI surfaces and traditional ones keeps moving.

You cannot prove the decision either way. With impressions-only reporting, you will not be able to show anyone, including yourself, whether opting out helped. That is an uncomfortable position to be in six months later when traffic has drifted for unrelated reasons.

It is all or nothing. The control works at property level. If you have one section of the site where being summarised genuinely costs you and another where visibility helps, you cannot split them until page-level controls arrive in 2027.

This may not be worth thinking about at all if AI features are a small share of your impressions, if you sell something with a considered buying process, or if your property has not received the setting yet.

Editor’s note: This area changes quickly, so check the latest platform policy before making compliance decisions.

What to do this week

Open Search Console and check whether the Search generative AI control and the generative AI performance report exist on your property. If they do, record your current AI impressions figure and today’s date somewhere you will find it again. Then wait. The click data Google has been required to supply starts arriving in December, and that is the first point at which this becomes a decision you can actually make with evidence rather than instinct.

If you find yourself impatient with that answer, ask what you would do differently tomorrow if the numbers came back either way. For most sites, honestly, nothing.

The post Google’s AI Search Opt-Out Is Live. Here Is What You Give Up. appeared first on Mark8ng.com.