Entity Confidence Registered Trade Mark

Author: Tony Lord

  • The GEO Advice You Followed Was Written for a World That No Longer Exists

    The GEO Advice You Followed Was Written for a World That No Longer Exists

    In January 2025, ChatGPT held 86.7% of AI chatbot web session share. By January 2026, that figure had fallen to 64.5%. US mobile share had dropped below 40% for the first time. In the same twelve months, Google Gemini grew from 5.7% to 21.5% — nearly a fourfold increase. Perplexity grew 370% year-on-year.

    Most published GEO advice was written when ChatGPT had 87% of the market. At that concentration, treating “optimise for AI” and “optimise for ChatGPT” as synonymous was reasonable. It is no longer reasonable. The market has fragmented faster than the advice has updated.

    What was built for 87%

    The practical playbooks for AI visibility — which outlets to target, which content formats to prioritise, which technical signals matter — were largely derived from studying ChatGPT’s citation behaviour. ChatGPT was the obvious choice: it was available for testing, it had public documentation, and at 87% it was functionally the entire market. When practitioners said “AI search,” they meant ChatGPT.

    That produces a body of advice that is not wrong so much as platform-specific. Targeting Reuters, the Financial Times, and Axios makes excellent sense for ChatGPT — Muck Rack’s analysis of 1 million+ citations (July 2025) confirms these are among ChatGPT’s most-cited journalism outlets. Prioritising fresh coverage from the last twelve months makes sense for ChatGPT, which draws 56% of its journalism citations from that window.

    This advice is well-evidenced and worth following. It is just not advice for “AI.” It is advice for ChatGPT — written at a moment when that distinction did not seem to matter.

    Why Gemini’s rise is not the threat it looks like

    The obvious interpretation of the market share shift is that Gemini has eaten into ChatGPT’s dominance. That is true. The less obvious interpretation is that Gemini’s growth is, paradoxically, an argument for traditional search quality — not against it.

    The reason is architecture. University of Toronto researchers tested domain overlap between each major AI engine and Google’s top-10 search results across 1,000+ queries (arXiv:2601.16858, January 2026). The results by engine:

    • GPT-4o: 4.0% overlap with Google
    • Gemini: 11.1% overlap with Google
    • Claude: 12.6% overlap with Google
    • Perplexity: 15.2% overlap with Google

    Gemini uses Google Search as its retrieval grounding. Its domain overlap with Google is nearly three times ChatGPT’s. The platform taking market share from ChatGPT is the one most tightly coupled to the search index that SEO builds. As Gemini’s share grows, the fraction of AI interactions that run through Google’s retrieval infrastructure grows with it.

    The claim that AI search requires a strategy separate from traditional search is structurally weaker today than it was a year ago — not stronger.

    One partial reprieve

    There is a second piece of data that reduces the complexity somewhat. Muck Rack’s outlet analysis finds that ChatGPT and Gemini share an almost identical journalism citation profile: Reuters, Financial Times, Time, Forbes, Axios. The two biggest non-Google AI engines — one holding 64% of the market, the other growing fastest — are drawing from the same outlet pool.

    For brands that built their earned media strategy around ChatGPT’s citation preferences, Gemini’s rise may not require wholesale reprioritisation. The outlet list that serves ChatGPT is likely to serve Gemini at similar rates. The outlet divergence problem is currently concentrated in Claude, not Gemini.

    This does not mean the advice translates frictionlessly. Gemini’s higher Google-grounding means that its citation behaviour is also more dependent on search ranking than ChatGPT’s — the 4% vs 11.1% domain overlap gap is not just trivia, it is a statement about what prerequisite work you need to have done. But the outlet-level strategy is more transferable than the platform shift headline implies.

    Where the fragmentation actually bites

    The sharper problem is not ChatGPT-to-Gemini substitution. It is the growth of everything else.

    Claude is structurally different. It cites Reuters approximately 50 times less than ChatGPT (Muck Rack, July 2025). Its top journalism sources — Good Housekeeping, TechRadar, Harvard Business Review — have almost nothing in common with the wire-service profile that serves ChatGPT and Gemini. Claude also operates on a longer temporal window: only 36% of its journalism citations come from the last twelve months, compared to 56% for ChatGPT.

    Perplexity grew 370% year-on-year from a smaller base but 1.2 billion monthly AI chatbot sessions (Similarweb / Vertu, 2026) means even minority platforms carry volume. Perplexity’s source mix draws heavily on domain overlap with organic search (15.2%, the highest of the four) but blends in social and video content in ways the other engines do not.

    Semrush’s longitudinal tracking (October 2025) documents the platform-specific volatility: Reddit dropped 82% in ChatGPT’s citation share while rising 74% in Google AI Mode in a single quarter. The same source, moving in opposite directions simultaneously, on the two biggest platforms. That is not noise — it is a structural signal that platform-specific dynamics are already operating at a level that single-channel strategy cannot capture.

    What a Gemini-first strategy looks like

    For most brands, the market share data points to the same practical conclusion from two directions.

    First: Gemini’s growth strengthens the case for search quality. Its retrieval architecture means that ranking in Google’s index is not just a prerequisite for ChatGPT citation — it is a more direct input for Gemini. The fraction of AI sessions where search ranking materially influences citation outcomes is growing, not shrinking.

    Second: the outlet strategy for the ChatGPT/Gemini bloc (now roughly 85% of the market combined) still centres on wire services, financial press, and authority publications — and it requires consistent recent coverage, not historical presence.

    What is not covered by that strategy is Claude — and to a lesser extent, Perplexity. For brands whose customers index toward research-oriented, knowledge-worker, or specialist professional contexts, Claude’s growing share matters in ways a wire-service PR strategy will not address.

    The advice most brands received about AI visibility is not wrong. It is increasingly incomplete — and the incompleteness has a specific shape. The market that the advice was written for is gone. The question now is which slice of the new market your customers actually live in.

  • Mass Adoption Is Already Here

    Mass Adoption Is Already Here

    The conversation about AI search has been framed, for two years, as a transition story. Brands were preparing for a future state. Watching adoption curves. Waiting for the numbers to become undeniable.

    The numbers are now undeniable. The transition is over.

    The Threshold That Just Passed

    ChatGPT had 400 million weekly active users in February 2025. By July 2025 it had 700 million — growth of roughly 11–12% month over month across five months, according to OpenAI’s own published usage research.

    The threshold that matters — the point Rogers’ diffusion model identifies as the mainstream tipping point — is 20% of global internet users. At 5.65 billion internet users, that is approximately 1.13 billion weekly active users.

    At 5% monthly growth — the most conservative plausible rate — ChatGPT crossed that threshold around April 2026.

    The window for establishing AI search presence before mass adoption normalised was narrow. It has closed. Brands that have not yet invested in AI visibility are entering the space as latecomers, not early movers.

    Why This Is Different From “AI Is Growing”

    Every month brings a new statistic showing AI tools gaining ground. That story is familiar and easy to discount — adoption curves always look steep in early innings.

    This is not that story.

    What changed in the OpenAI usage data published in September 2025 was not the scale number. It was the behaviour number.

    In June 2025, asking questions — information queries of the kind that were previously the exclusive domain of search engines — accounted for 51.6% of all ChatGPT interactions. Generating content (writing, summarising, creating) had fallen to 34.6%.

    ChatGPT started as a content generation tool. Within a year of that framing being established, search had become its primary use case. Not a secondary one. The primary one.

    The implication: the 700 million weekly active users are not mostly doing work tasks. They are mostly asking questions that used to go to Google.

    The Non-Work Signal

    One data point anchors this as a consumer behaviour shift, not a productivity tool story.

    In June 2024, non-work messages accounted for 53% of ChatGPT interactions. By June 2025 that figure had risen to 73%.

    A tool that is 73% non-work has crossed from enterprise utility to everyday utility. It is being used to make decisions — about purchases, products, services, brands — by people who are not thinking of it as a workplace tool. They are thinking of it the way they once thought of Google: as the place you start when you want to know something.

    For brands, this matters because it relocates the audience. The users who are forming opinions about your product category, your competitors, your company, are doing so through an interface that does not show them your website. It shows them whatever the model has assembled about you from the sources it trusts.

    The Traffic Picture

    AI-referred traffic to retail sites grew 393% year over year in Q1 2026, according to Adobe’s analysis of over a trillion visits to US retail sites.

    At current penetration, that traffic is still a small fraction of total visits. But the quality signal has already flipped. In March 2025, AI-referred visitors converted 38% worse than organic search traffic. By March 2026 — one year later — they converted 42% better. They also spent 48% longer on site and browsed 13% more pages.

    The implication of that reversal: being recommended by AI is already a filtering mechanism. Users who have formed a positive view of a brand through AI interaction and then choose to visit the site are arriving with higher intent than a typical search click. Volume is low now. Quality is high now. Scale is arriving fast.

    The brands AI already recommends are capturing that quality audience. The brands it doesn’t are invisible to it.

    What Mass Adoption Changes

    Before mass adoption, a brand’s absence from AI responses was an edge-case problem. Only a small segment of users would encounter the gap.

    After mass adoption, absence is mainstream exposure. A category buyer with a question — about your market, your product type, your company — is now more likely to encounter AI-mediated information than a traditional search result page.

    Two shifts follow from this:

    First: the strategic priority is no longer building presence ahead of a future state. It is closing a visibility deficit that already exists.

    Second: the measurement framework has to change. Traffic analytics capture what a fraction of users do after they decide to visit. They do not capture what the majority of users encounter, conclude, and act on before they ever reach your site. Recommendation measurement — how often AI mentions your brand, in what context, with what confidence — is the only view into the larger surface.

    The Early Mover Window

    Early mover advantage in AI search follows the same logic as early mover advantage in traditional SEO. The brands that built consistent earned media presence, clear entity signals, and cross-platform consistency in 2024 and early 2025 are compounding that investment now. The knowledge base about them is richer. The recommendation rate is higher. The confidence language around them is stronger.

    That window is not closed forever — AI models are continuously retrained and updated. But it does not get easier to enter a market after mass adoption than before it. The compounding that early movers have accumulated does not reset.

    The question for every brand that has not yet acted is not whether to invest in AI visibility. The threshold passed in April. The question is how much gap to close, and how quickly.

    *Sources: OpenAI, “How People Use ChatGPT” (September 2025), via Search Engine Journal; Adobe.

  • The Six Factors That Actually Decide Whether AI Mentions Your Business

    The Six Factors That Actually Decide Whether AI Mentions Your Business

    AI search engines don’t rank pages the way Google does. They decide, page by page, whether a piece of content is worth quoting in an answer — and the data now shows that decision comes down to roughly the same six things, again and again, across independent studies.

    The Short Version

    Multiple studies — practitioner datasets covering thousands of campaigns, and academic research — converge on the same handful of factors that predict whether AI systems cite a business. None of them require a large budget. All of them require consistency.

    The Six Factors

    1. Can AI actually read the page? This comes before everything else. Adobe’s analysis of over a trillion US retail site visits found that product pages are, on average, only 66% readable to AI systems — homepages fare a little better at 75%. The most common cause is JavaScript-rendered content that never reaches the raw HTML an AI crawler sees. A page that fails here doesn’t get a lower score on the other five factors. It gets ignored entirely.

    2. Does the content answer the question immediately? Content that states its main point in the first sentence of each section — rather than building up to it — is cited roughly 2.3 times more often than content with a traditional narrative structure, according to campaign data spanning hundreds of clients. AI systems extract answers; they don’t reward you for making them scroll to find one.

    3. Is it written plainly? Content that’s needlessly hedged, padded, or complex costs citations even when the underlying facts are accurate. An analysis of how ChatGPT selects and quotes source material found that direct, definitional sentences — “X is Y” — are cited nearly twice as often as explanatory or discursive prose, and that headings framed as direct questions are twice as likely to be quoted as narrative ones. AI systems extract whichever available source states a fact most clearly — not necessarily the most authoritative-sounding one.

    4. Is it current? AI-cited content skews newer than traditionally-ranked content. One large multi-platform study — 12,500+ queries across 8,000 domains — found a 3.2x citation multiplier for content refreshed within the last 30 days. Perplexity is the most aggressive platform on this factor, deprioritising unrefreshed content after just 2-3 days.

    5. Are you mentioned by name, not just linked? Across several independent datasets, being mentioned by name in other people’s content correlates about 3 times more strongly with AI citation than the number of backlinks pointing at your site. A link is a reference. A named mention is a recommendation.

    6. Is your business unambiguous? Name, address, and category consistency, plus basic structured data markup, form a disambiguation layer underneath everything else. Without it, the other five factors get diluted across fragmented, unrecognised versions of your brand — the AI can’t credit consistent evidence to an entity it can’t reliably identify as one thing.

    What This Means in Practice

    None of these six factors require a large marketing budget. They require structural discipline: checking whether your pages are actually machine-readable, writing answers before context, keeping content current, and making sure your brand is described the same way everywhere it appears. Most businesses have never audited even the first one.

  • The SME Advantage in AI Search

    The SME Advantage in AI Search

    Almost every piece of advice written about AI search optimisation treats the challenge as universal. Get cited by AI. Build authority. Earn coverage. It applies to all brands equally.

    The data says otherwise. A peer-reviewed study from the University of Toronto, published in January 2026, found that large, well-known brands and mid-sized niche brands are playing structurally different games in AI search — and that for one of those groups, the investment required to improve AI visibility is dramatically lower than most people assume.

    Summary: The brands best positioned to move the AI visibility needle quickly are not the category giants. They are the well-regarded specialists that AI models have not yet fully formed a view about — where the next piece of coverage in the right outlet can change what AI says.

    Small and Mid-sized SMEs are that group.

    Read on…

    Why Big Brands Are Stuck

    Start with the uncomfortable finding for large brands.
    Chen et al. (arXiv:2601.16858, January 2026) ran a series of perturbation experiments on GPT-4o: they manipulated the evidence the model received about a set of brands — shuffling it, restricting it, even swapping brand names into irrelevant snippets — and measured how much those manipulations changed the model’s rankings.
    For well-known, popular brands, the results were clear. Changing the retrieved evidence barely moved the output.

    PerturbationPopular brands (Δavg)Niche brands (Δavg)
    Snippet shuffle — randomise evidence order2.304.15
    Strict grounding — restrict to provided snippets only1.520.46
    Entity swap — substitute brand names into irrelevant snippets2.604.63

    For popular brands, even aggressive manipulation of the retrieved evidence produces low rank deviation. The model’s internal hierarchy — built during training on years of web content — dominates. The model already knows who the major players are. Retrieval is used to confirm and support what it already believes, not to discover who deserves to be recommended.

    The citation miss rate data makes the same point from a different angle. For well-known automotive brands, the researchers logged how often a brand appeared in AI rankings without any supporting snippet from retrieved content:

    BrandAI ranking without citation support
    Toyota6%
    Honda3%
    Kia10%
    Chevrolet26%
    Cadillac58%
    Infiniti73%

    Cadillac and Infiniti appear in AI answers without any retrieved evidence more than half the time. The model is drawing on training priors, not retrieved content. Toyota appears without evidence only 6% of the time — because mainstream outlets have covered Toyota extensively, and that coverage dominates the retrieval pool.

    For a large brand trying to improve its AI positioning through content publication, this is a structural problem. The model has already formed its view. Publishing fresh content does not change a pre-trained ranking. Getting into future training runs — through Wikipedia, persistent authoritative web mentions, and community discussion at scale — is what moves the needle for large brands. That is a long-cycle activity measured in years, not campaigns.

    Why SMEs Are Not Stuck

    Niche brands — the mid-sized, specialist businesses that serve specific categories or audiences — face a different situation entirely.
    In the same experiments, niche entities showed high rank sensitivity to the retrieved evidence. When snippets were shuffled, rankings shifted significantly. When brand names were swapped into unrelated content, the model was again highly sensitive. The critical result: when retrieval was restricted to only the provided snippets, niche entity rankings stabilised dramatically (Δavg dropped from 4.15 to 0.46).
    The interpretation: the model has no stable internal hierarchy for niche brands, so it follows wherever the retrieved evidence leads. In the absence of strong training priors, retrieval is driving the answer — not confirming one.
    The alignment data reinforces this. Kendall τ (alignment between holistic and pairwise rankings) for popular brands is 0.911 under normal conditions and reaches a near-perfect 1.000 under strict grounding. For niche brands, it is 0.556 under normal conditions and only 0.689 under strict grounding — reflecting genuine model uncertainty rather than stable, training-formed views.

    For a mid-sized SME, this uncertainty is an opportunity. Because the model lacks pre-trained confidence about niche entities, fresh earned coverage in a trusted source can shift AI rankings in the next retrieval cycle. The investment required to change AI outcomes is proportionally smaller — and the feedback loop is measured in weeks rather than years.

    The Niche Convergence Bonus

    There is a further structural advantage for niche brands that the research documents.
    When a query is narrow enough that both AI and Google converge on a small pool of specialist sources, the gap between SEO and AI search optimisation largely disappears. The paper found that niche queries produce 3–4 percentage points more overlap between AI citations and Google’s top-10 results than popular queries do.
    For a popular brand query — “best smartphones” — GPT-4o and Google are drawing from completely different domain ecosystems (the study measured only 4% domain overlap for GPT-4o overall). For a niche query — “top ultramarathon GPS watches” — both systems converge on the same small cluster of specialist review outlets, because that is where the authoritative content lives.

    This means that for mid-sized brands in specialist categories, AI search optimisation and traditional SEO are largely the same work. Securing coverage in the specialist publications that cover your category — the review sites, trade outlets, and editorial destinations your potential buyers use — builds Google ranking and AI citation simultaneously. There is no separate AI strategy required.

    What This Means in Practice

    The distinction between popular and niche entities implies a different investment logic for different types of brand.
    For large, well-known brands, AI ranking is governed by training-time knowledge. The work that moves the needle — Wikipedia presence and Wikidata records, repeated mentions in high-authority publications over years, community discussion at scale — is the kind of brand-building that has always mattered for prominence. Short-cycle content publication produces minimal effect on AI rankings for popular queries. The horizon for this work is long.
    For mid-sized SMEs, retrieval dominates. The practical priorities are:
    Fresh earned coverage in trusted specialist outlets. AI systems cite content that is, on average, 62–90 days old (Chen et al.). A consistent cadence of editorial coverage in the publications that cover your category keeps your brand in the active citation pool. Historical coverage from eighteen months ago is likely no longer being cited.
    Structured, extractable content. For niche entities, the model is building its answer from what retrieval surfaces. Content that is clearly attributed, well-structured, and names the brand in the evaluative sentence — not buried in surrounding context — is more likely to enter the context window and generate a recommendation.

    Coverage in category-trusted outlets, not just any press. AI citation authority is concentrated in relatively small clusters of specialist publications per category. For consumer electronics, TechRadar, Tom’s Guide, RTINGS, and CNET dominate AI citations. For automotive, Consumer Reports and Car and Driver. The equivalent tier exists in virtually every category. Knowing which publications AI systems trust in your space is the highest-value strategic input for an earned media brief.

    The Inversion That Most AI SEO Advice Misses

    Most writing on AI search positions large enterprises as the natural leaders in this space — the brands with the biggest content budgets, the most established authority, the longest track records. In traditional SEO, that advantage compounds.
    In AI search, the data suggests the opposite logic applies at the margins. Large brands are locked in by training data they cannot rapidly change. Mid-sized SMEs are operating in the retrieval layer, where fresh coverage, consistent presence, and specialist-outlet relationships translate directly into AI visibility — faster, and with a smaller investment floor.

    The brands best positioned to move the AI visibility needle quickly are not the category giants. They are the well-regarded specialists that AI models have not yet fully formed a view about — where the next piece of coverage in the right outlet can change what AI says.

    Sources: Chen et al. arXiv:2601.16858 (University of Toronto, January 2026); ConvertMate GEO Benchmark Study 2026.

  • Google Told You to Optimise for Google

    Google Told You to Optimise for Google

    In May 2026, Google published a guide to generative AI optimisation. The headline message: “Optimising for generative AI is still SEO.” Nick Fox, Google’s VP of Search, has said the same thing in public. A senior Google executive, using the authority of one of the most trusted technical brands on earth, telling marketers clearly: the skills you have are the skills you need.

    The statement is correct. It is also written by Google, for Google. And Google is no longer the only AI your customers are using.

    What the data actually shows

    The arXiv study that tested this most rigorously (Chen et al., University of Toronto, January 2026) measured domain-level overlap between each major AI system and Google’s top-10 search results across 1,000 queries:

    • GPT-4o: 4.0% overlap with Google
    • Gemini: 11.1%
    • Claude: 12.6%
    • Perplexity: 15.2%

    GPT-4o — the engine behind the most-used AI chatbot on the planet — shares four percent of its cited domains with Google’s top results. Not forty percent. Four.

    BrandFeatured’s analysis adds a further dimension: less than 1% of citations overlap between ChatGPT and Perplexity. The two most-used AI answer engines are drawing from almost entirely different source pools.

    Google’s own AI products sit at the high end of the overlap range (Gemini at 11%). This makes sense: Gemini is built on Google’s infrastructure and uses Google Search grounding. Of course it draws from Google’s index. That’s also why Google’s guide is correct — for Google’s AI.

    The guide answers the right question about the wrong platform

    Nick Fox’s statement is accurate. For Google AI Overviews, 76% of citations come from pages already in Google’s top-10 results (Ahrefs data). Rank in Google, appear in Google’s AI. The logic holds. The advice works.

    The problem is the implicit premise: that optimising for Google is optimising for AI search. That premise was once defensible. A strong Google presence got you most of the way there across platforms.

    By mid-2026, that premise has collapsed. A page that ranks well in Google has roughly a one-in-twenty chance of appearing in GPT-4o’s cited sources for the same query.

    Google’s guide doesn’t mention this. It doesn’t need to. It is a guide to Google’s products.

    The platforms diverge in ways that change the strategy, not just the tactics

    The divergence is not just about which pages each AI retrieves. It runs through every lever brands can actually pull.

    Earned media vs owned content. The arXiv study found GPT-4o cites 57% earned media; Gemini cites 46% earned and 46% brand-owned content. Google’s own AI gives substantially more weight to brand-owned pages than ChatGPT does. Optimising your website for Google’s AI is rational. That same owned content is a marginal signal for ChatGPT, which is dominated by independent editorial coverage.

    Freshness windows. ChatGPT’s median citation age in consumer electronics is 80 days. Google’s is 130 days. A content programme calibrated to Google’s freshness cadence — one that refreshes every quarter — may be running too slow to reach ChatGPT’s retrieval window for the same queries.

    Outlet profile. Claude cites Reuters roughly 50 times less than ChatGPT. Claude’s top-cited journalism outlets include Good Housekeeping, TechRadar, and Harvard Business Review. ChatGPT and Gemini share an almost identical outlet profile anchored by Reuters, the Financial Times, Time, Forbes, and Axios. A PR programme built for Google’s citation preferences will serve Gemini well and largely miss Claude.

    These are not marginal differences. They are structural divergences across the platforms that now collectively account for the majority of AI-mediated information discovery.

    The market share shift makes this urgent

    ChatGPT’s share of AI chatbot traffic was 87% in early 2025. By May 2026 it had fallen to 64%. US mobile share had dropped below 40% for the first time. Gemini went from 5.7% to 21.5% in the same period.

    Most AI visibility advice was written for an 87%-ChatGPT world. That world is gone. The distribution across AI platforms is fragmenting rapidly — which means the gap between “optimise for Google” and “optimise for AI” is widening every quarter.

    Google’s guide was accurate when it was published. The market it describes is changing underneath it.

    What Google’s guide gets right (and why that still matters)

    The guide’s underlying recommendations are not wrong. Demonstrate expertise, publish original research, earn independent editorial coverage, maintain consistent entity signals. These are correct — and they are the inputs that help with ChatGPT, Perplexity, and Claude as well as Google’s AI.

    The fundamentals converge. The execution diverges.

    Which outlets you target for earned coverage. How fast you refresh content. Which platforms you prioritise when budgets are constrained. Those decisions are platform-specific. Google’s guide does not help you make them, because it was not written to.

    The question the guide doesn’t ask

    Google’s guide starts from the premise that you are optimising for Google. That is a reasonable premise for a guide published by Google.

    It is not a sufficient premise for a brand strategy in 2026.

    Your customers are not on one platform. A B2B buyer researching software options may be using Claude or Perplexity. A consumer comparing products may be in ChatGPT on iOS. The AI-mediated moments that shape purchase decisions are distributed across platforms with fundamentally different citation logic.

    The question the guide doesn’t ask is the one that matters most: which AI is your customer using when they form the opinion that drives the decision?

    Google’s guide is the right answer. It’s just the answer to a narrower question than the one you need to be asking.

    Sources: Google AI Optimization Guide (https://developers.google.com/search/docs/fundamentals/ai-optimization-guide, 2026-05-15); Chen et al., arXiv 2601.16858 (University of Toronto, January 2026); BrandFeatured AI ranking factors analysis (2026); Ahrefs AI Overview citation study; Muck Rack Generative Pulse (via Nieman Lab, July 2025).

  • Why AI Search Engines Default to Famous Brands

    Why AI Search Engines Default to Famous Brands (and Why That Doesn’t Last)

    AI search engines recommend famous brands more often not because those brands are better, but because they were already well documented before the AI was ever built. That’s a fixable gap, not a permanent one — and for most of the questions real customers actually ask, it doesn’t apply at all.

    The Short Version

    Every AI system — ChatGPT, Gemini, Google’s AI Overviews — was trained on a huge snapshot of the internet before it ever answered a single question. Big, famous brands had spent decades building up Wikipedia pages, news coverage, and mentions across the web. That presence got baked into the AI during training. Small and mid-sized businesses, however good they are, usually hadn’t built up the same paper trail — so the AI never learned much about them.

    That’s the actual mechanism behind “AI only recommends famous companies.” It isn’t a judgement of quality. It’s a gap in what the AI happened to read before it was finished being built.

    The Evidence

    A 2026 University of Toronto study tested this directly. Researchers fed an AI model manipulated evidence about different brands and measured how much its answers changed. For famous brands, changing the evidence barely moved the AI’s recommendations at all — its mind was already made up from training. For lesser-known brands, the same manipulation changed the results dramatically. The AI had no fixed opinion to defend, so whatever it found when it looked things up is what actually decided the answer.

    In other words: fame makes an AI’s opinion sticky. Obscurity makes it responsive.

    Where Small Businesses Actually Compete

    That second finding is the opportunity. Most real customer questions — “best plumber near me,” “who’s reliable for X,” “is this company any good for Y” — aren’t the kind of query where an AI has a fixed opinion already. For questions like these, the AI has to go and look at current information to answer. That’s exactly the territory where a small business’s recent reviews, press mentions, and clear web presence can shape the answer directly.

    The famous-brand advantage is real, but it’s concentrated in the broad, generic searches that were already decided years ago. It doesn’t extend to the specific, current, local questions that make up most of what customers actually type.

    What This Means in Practice

    Beating a famous competitor’s name recognition isn’t the goal, and isn’t realistic. The goal is making sure that when an AI system goes looking for current evidence — which it does constantly, for exactly the questions where small businesses compete — it finds consistent, credible information about your business. That’s a matter of current evidence, not decades of accumulated fame.

    Go Deeper

    This piece summarises the underlying research and its philosophical framing — including the specific study data and the academic argument for why this amounts to more than a marketing disadvantage — in Art #0049: The Structural Exclusion Problem.

    Article: #0072
    Date: 2026-07-16

  • Why Mid-Sized Brands Are Locked Out of AI Knowledge

    The Structural Exclusion Problem: Why Mid-Sized Brands Are Locked Out of AI Knowledge

    Most GEO and EC advice frames the SME visibility problem as a competitive disadvantage. You are behind the large players. Here is how to close the gap.

    That framing is wrong — or at least, it is not wrong enough. The problem is not that mid-sized brands are losing a race. It is that the race was designed without them.

    The Training Data Problem Is Not Random

    Before AI systems answer questions about your brand, your industry, or your category, they have already formed a view. That view was assembled during pre-training — the process by which a model learns the world’s knowledge from an enormous corpus of text, before it ever answers a query.

    That corpus was not a neutral sample of what exists. It was a weighted sample of what had been digitally published, cited, covered in large-circulation media, and documented in the reference sources that trained models weight most heavily: Wikipedia, academic databases, major editorial outlets, trade press with decades of archive depth.

    Large brands — the multinationals, the household names, the category incumbents — accumulated exactly these kinds of presence over decades. They had Wikipedia pages. They had Reuters coverage. They had academic case studies, analyst reports, Financial Times profiles. That presence existed before any training data was assembled. When the models trained on the web, they trained on a web that had already organised itself around the visible and the established.

    Mid-sized SMEs, by design, had none of this. A regional services firm with thirty years of operational excellence but no analyst coverage and no Wikipedia entry had produced no signal that pre-training data collection would recognise as authoritative. The AI system did not decide that firm was unimportant. The training data never recorded its importance in the first place.

    This is not market inefficiency. It is structural exclusion embedded in how knowledge was assembled.

    What the Evidence Shows

    The academic basis for this is not theoretical. Chen et al. (arXiv:2601.16858, January 2026) ran perturbation experiments on GPT-4o: they manipulated the evidence the model received about brands — shuffling retrieved snippets, restricting retrieval to only provided content, injecting brand names into unrelated material — and measured how much those manipulations moved the model’s output rankings.

    For well-known, popular brands, the results were stark. Average rank deviation under snippet manipulation: 2.30–2.60. For niche entities: 4.15–4.63. Popular brand rankings barely shifted regardless of what the retrieved evidence said. The model already knew the answer. The citation miss rate for Cadillac was 58%. For Infiniti, 73%. Those brands appeared in AI answers without any supporting retrieved content more than half the time — drawn entirely from training priors. (Chen et al., arXiv:2601.16858)

    The mechanism is clear. Popular entity rankings are governed by pre-trained knowledge. Retrieved evidence is used to confirm what the model already believes, not to discover who deserves to be recommended. For niche brands — where the model holds no stable prior — retrieval actually drives the answer. The two populations are not playing the same game.

    The training-data corpus itself reflects this asymmetry. Analysis of earned media citation patterns shows that 82% of AI citations come from earned media sources — but those sources are heavily concentrated in a small cluster of high-authority outlets with long publication histories. (Muck Rack, What Is AI Reading?, 2025) The outlets that trained AI systems to recognise credibility are the same outlets that were historically accessible only to companies with significant PR infrastructure. Small and mid-sized businesses have always been systemically underrepresented in major national media. That underrepresentation was baked into training data.

    The Philosophical Reframe

    Mark Coeckelbergh (2025) draws on Dotson’s (2014) concept of epistemic oppression — “a persistent and unwarranted infringement on the ability to utilize persuasively shared epistemic resources that hinder one’s contribution to knowledge production” — and extends it to AI-mediated knowledge environments. The argument is that AI does not merely repeat existing power asymmetries. It embeds them structurally.

    The relevant extension for brands is not just about producing knowledge. It is about the knowledge environments where customers form beliefs. If an AI system has no training-data basis to surface a brand in the answers your potential customers receive, that brand is excluded from the epistemic environment where purchase decisions begin. A customer who asks an AI assistant “what are the best firms for X?” receives an answer shaped entirely by what training data recognised as authoritative before any query was submitted. If your brand was not visible to the training data, you are absent from the answer — not because you lack capability, but because the epistemic infrastructure never recorded it. (Coeckelbergh, Social Epistemology 39(1), 2025)

    This is a consumption-side exclusion, not just a production-side one. The SME is not only excluded from contributing knowledge; it is excluded from the environments where knowledge shapes customer belief.

    Framing the SME AI visibility problem as a competitive gap misses this point. The gap did not emerge because large competitors worked harder or invested more in the last two years. It emerged because AI training data weighted types of presence — Wikipedia coverage, academic citations, major-media editorial — that mid-sized businesses have never had the infrastructure to accumulate. That is not a level playing field with a laggard on one side. That is a structural condition. And structural conditions require structural responses.

    What EC Work Actually Is

    The standard commercial framing for AI visibility tools — GEO vendors, entity optimisation platforms, AI citation monitoring services — presents the work as a competitive instrument. Get cited before your competitors do.

    That framing is not false. But it understates what the work is.

    Earned media in credible outlets, consistently maintained across a publication cadence, does two things simultaneously. In the retrieval layer, it provides fresh evidence for AI systems to draw on. In the training layer — for future model updates — it begins to build the kind of cross-source corroboration that pre-training data collection recognises as authoritative. Structured entity data (schema markup, sameAs identifiers, Wikidata records) creates the machine-readable signals that allow AI systems to resolve which firm you are and connect disparate mentions into a coherent entity record. Original research and defined analytical frameworks give AI systems something causally grounded to cite — not just statistical pattern-matching from general web content.

    Each of these is not just a tactic for climbing citation rankings. Each is a mechanism for entering the epistemic infrastructure from which training data bias has structurally excluded mid-sized brands.

    The EC toolset, properly understood, is not a route to competitive advantage. It is a route to epistemic access — the ability to participate in the knowledge environments where your potential customers form beliefs. Large brands already have that access, because their historical presence built it for them. Mid-sized brands are building it now, with tools that did not exist when the training data was assembled.

    That distinction matters for how you explain the work, how you measure its value, and how you evaluate whether a GEO vendor is offering you a tactical campaign or a structural solution.

    The Limit of the Argument

    This framing should not be stretched beyond its evidence base. Coeckelbergh’s epistemic justice concept is a normative philosophical framework, not an empirical claim about AI systems. The pre-training bias data from Chen et al. is robust, but it covers a single model (GPT-4o) and a narrow category (automotive brands). The training data composition claims are grounded in observed citation patterns, not disclosed training corpus analyses.

    What can be said with confidence: AI training data demonstrably weighted large-media presence, reference database coverage, and academic documentation in ways that structurally disadvantaged brands without those resources. That weighting was not deliberate exclusion — it was an artefact of using the web as training material. But the effect is structural regardless of intent. And the practical response — earned media, entity data, original research — addresses it at the level it operates: the evidence base from which AI systems draw their knowledge.

    The Close

    Generic GEO vendors offer tactics. Citation audits. Source gap analysis. Content briefs for the current citation map.

    The problem with that framing is that the citation map moves — 120% average source volatility in three months in late 2025 (Semrush AI Visibility Index, Geaney, Oct 2025) — while the underlying epistemic exclusion does not. Chasing this quarter’s top-cited sources does not change the structural fact that a brand with no pre-training signal is operating from a deficit that tactical content alone cannot close.

    The bias is not in the algorithm. It is in the data that trained it. And data — accumulated evidence, earned recognition, consistent cross-source corroboration — can be changed. It just requires understanding what it is you are actually trying to change.

    Sources: Coeckelbergh, Social Epistemology 39(1), 2025; Chen et al., arXiv:2601.16858 (University of Toronto, January 2026); Semrush AI visibility trend update, October 2025; earned media AI citation analysis, multiple sources.

    Article: #0049
    Date: 2026-07-16

  • It’s the Same Job

    It’s the Same Job: Why GEO, AEO, and SEO Are Different Names for Identical Work

    **Published**: 2026-07-14

    **Status**: Active

    A new discipline is born every few months. First it was GEO — Generative Engine Optimization. Then AEO — Answer Engine Optimization. Then LLMO — Large Language Model Optimization. Some agencies now offer AI SEO, AIO, or AIGC as distinct service lines. Each comes with its own acronym, its own framework, and an implicit pitch: what you were doing before is no longer enough.

    Most of this is noise. The core claim — that optimising for AI search requires fundamentally different work from optimising for traditional search — is not supported by the evidence. Three independent sources, spanning a Google vice president, an independent practitioner, and Wikipedia, all reach the same conclusion. The empirical data backs them up.

    What a Senior Google Executive Actually Said

    Nick Fox is Vice President of Product at Google. When asked about GEO, he said:

    “Optimizing for AI search is the same as optimizing for traditional search (SEO).”

    That statement appears in Wikipedia’s article on Generative Engine Optimization, published in September 2025. It is not a caveat or a nuanced position — it is a direct equivalence from the person responsible for Google’s search product.

    It is worth pausing on the source here. Fox is not a detached observer. Google has a commercial interest in AI search. If AI-specific optimisation were genuinely necessary, Google would be in a strong position to sell it. The statement is not a disclaimer about future divergence; it is a description of how the system currently works.

    The Empirical Evidence

    The Nick Fox position is not just an executive’s opinion — it has a clear empirical basis.

    Ahrefs analysed 1.4 million ChatGPT prompts and measured where ChatGPT’s cited pages actually come from. The search channel accounts for **88% of all ChatGPT citations** — pages indexed and ranked in traditional web search. The news channel adds another 12%. Reddit, YouTube, and academic preprints each contribute less than 2%.

    A separate Ahrefs study of AI Overviews found that, as of July 2025, **76% of AI Overview citations came from pages already ranking in Google’s top 10** — not just indexed pages, but pages ranking prominently. That figure has since fallen sharply: by January 2026, Ahrefs’ Brand Radar analysis put the top-10 share at 38%, as query fan-out pushes AI Overviews to cite from an increasingly wide and dispersed pool of pages.

    The implication is mechanical, not philosophical. AI engines need a way to find and evaluate web content. The most comprehensive, freshness-maintained, and authority-weighted content database available to them is the search index. A page that does not rank in web search is almost certainly invisible to AI citation pipelines for the same reason: it is not in the database AI engines are drawing from. Being in the index is still the precondition for citation — but, as the AI Overview figures show, ranking prominently within it is a weaker predictor of citation than it was six months earlier.

    The fastest path to AI visibility is the same as the fastest path to search visibility — because AI citation largely *is* search visibility, viewed through a different lens.

    The Practitioner View

    Jon Monk, Head of Performance at ASP, made the same argument in a session at Event Tech Live in London (November 2025), speaking from a practitioner perspective rather than a Google executive’s podium.

    His position was unambiguous: GEO, AEO, and SEO are different labels for the same underlying work. Ranking well on Google is the single most valuable move an organisation can make for AI visibility. The fundamentals that produce strong search rankings — useful content, genuine depth, clear organisation, durable structure — are the fundamentals that produce AI citation.

    Monk’s specific example concerned event websites that tear down their pages after an event ends, resetting years of accumulated authority to zero. The lesson generalises: content permanence, topical depth, and authentic expertise compound over time for both search and AI. There is no separate accumulation curve for “AI authority.”

    Three independent sources — a Google vice president, an independent SEO practitioner, and Wikipedia — had reached the same conclusion by late 2025, without coordinating.

    The Convergence Is Strongest Where It Matters Most

    The arXiv study from the University of Toronto (Chen et al., January 2026) adds a further layer to this. When testing how AI engines handle popular versus niche queries, the researchers measured domain overlap between AI citations and Google’s top results.

    Niche queries — the long-tail, specialist, category-specific searches that characterise most SME visibility problems — produced **3–4 percentage points more overlap** between AI and Google than popular entity queries did. When the AI model lacks strong pre-training knowledge of a topic, it converges more closely with Google’s ranking logic. Both systems are drawing from the same narrow pool of specialist sources.

    This is the convergence point that matters commercially. Large brands fighting for presence in popular queries — “best smartphone,” “top credit card” — are operating in territory where AI has strong pre-trained views that retrieval alone cannot easily shift. The optimisation challenge there is genuinely harder, involving Wikipedia presence, years of authoritative mentions, and community discussion at scale.

    Niche entities — a regional accounting firm, a specialist software vendor, a B2B services brand — are in discovery mode. The AI model has limited prior knowledge and follows retrieved evidence closely. In that territory, the SEO and AEO challenge are nearly identical: get authoritative, structured, fresh content into trusted sources that AI engines retrieve.

    —

    ## What Is Genuinely Different

    Convergence does not mean identity. Two things are genuinely new in AI search, and it is worth being precise about what they are.

    **The measurement layer is different.** Traditional SEO measurement tracks rank position, impressions, and click-through rate. These metrics do not exist in AI search. A brand cited in an AI answer cannot track its “rank” — AI answers are not ranked lists. The commercially relevant metrics are citation presence (is the brand named in relevant AI responses?), confidence language (does the AI recommend or hedge?), and consistency across regenerations (how stable is the citation probability?). The instruments are new even if the underlying work is not.

    **The community channel is new.** Reddit, forums, and user-generated content shape what AI models know about brands during training — independently of whether those sources appear in citation outputs. Reddit is retrieved at massive scale by ChatGPT (over 16 million data points in the Ahrefs study) but is cited at only 1.93%. The model uses community content to build background knowledge, then cites institutional sources when formulating its answer. This training-mode influence has no direct equivalent in traditional SEO’s link-and-rank model. Community presence affects what AI engines know; search presence affects what they cite.

    Everything else — content structure, depth, authority signals, earned media, topical consistency — operates the same way it always has.

    —

    ## The Practical Conclusion

    If GEO, AEO, and SEO are the same job, the practical question is not whether to buy a new strategy — it is whether you are executing the existing one well enough to appear in AI answers.

    For most brands, the answer is no, and the reasons are familiar: insufficient depth on key topics, inconsistent authority signals across platforms, insufficient third-party coverage from the sources AI engines trust. These are search problems that have existed for years. AI search makes them more consequential because the citation threshold is higher — not every page that ranks gets cited, and citation requires a level of authority that ranking alone does not.

    The genuine addition that AI search requires is measurement. Without visibility into whether you are being cited, named, recommended, or ignored across AI platforms, you cannot manage the outcome. That measurement layer did not exist in traditional SEO because the output (a ranked list) was directly observable. AI responses are not. Building the ability to track citation presence across platforms is new work — but it is not a new discipline. It is instrumentation for a discipline that already existed.

    The brands that will win in AI search are not the ones that invest in the right acronym. They are the ones that execute on authority-building fundamentals — earned media, topical depth, content structure, consistent entity signals — and add the measurement layer to track what AI engines are actually saying about them.

    —

    *Sources: Nick Fox (Google VP) via Wikipedia GEO article (September 2025); Ahrefs 1.4M ChatGPT prompt study; Ahrefs AI Overview brand correlation research (July 2025 baseline); Ahrefs Brand Radar, AI Overview citation analysis (February 2026, data collected January 2026); Jon Monk, Event Tech Live London (November 2025); Chen et al. arXiv:2601.16858, University of Toronto (January 2026).*

  • The Shortlist You’re Not On

    The Shortlist You’re Not On

    Last updated: 2026-05-25

    Type: Original article / reactive commentary — Google I/O 2026

    At Google I/O last week, Google announced that AI Mode has crossed one billion monthly users. Queries are more than doubling every quarter. The search bar is being redesigned for the first time in 25 years.

    Those are the headline numbers. The more important announcement was buried beneath them.

    Google is launching information agents: persistent, background AI processes that monitor the web 24/7 and surface synthesised findings without being asked. No query required. The agent watches topics on a user’s behalf and delivers what it finds — relevant brand updates, price changes, competitor activity, vendor options — as push notifications, at whatever moment the information matters.

    And alongside it: Universal Cart, a cross-merchant purchasing layer built into Search, Gemini, YouTube, and Gmail. Agent Payments Protocol (AP2), a framework allowing AI agents to complete purchases on users’ behalf within conditions the user specifies upfront.

    Conditions like: preferred brands.

    This is not zero-click search. Zero-click still starts with a query. The user types something. A result appears. A brand gets cited or it doesn’t. The interception point exists.

    What Google announced last week eliminates the interception point.

    Not Zero-Click. Zero-Query.

    The progression is worth stating plainly.

    Traditional search: user forms a query, gets ten blue links, clicks one or two. Brand visibility = ranking.

    Zero-click search: user forms a query, gets an AI-generated answer at the top of the page, doesn’t click. Brand visibility = citation in the AI answer.

    Zero-search discovery: user delegates a monitoring task to a persistent agent. The agent researches, shortlists, and reports back. The user sees the conclusion. They never formed the query. Brand visibility = being in the agent’s knowledge before the task was assigned.

    Each stage doesn’t merely reduce traffic. It moves the moment of brand inclusion further upstream — further away from any content your marketing team can optimise in real time, and closer to what the AI already believed about your brand before it was asked.

    Google’s information agents are the first mass-market implementation of zero-search discovery. The concept was documented as theoretical in December 2025. It shipped in May 2026. One billion monthly AI Mode users now have access to it.

    The Shortlist Was Already Closed

    Here is the detail that changes the strategic picture.

    When Google launched Agent Payments Protocol, they gave users the ability to specify conditions for autonomous agent purchases: preferred brands, product types, spending limits. The agent acts within those constraints. It doesn’t ask at checkout what brand you want. It was told upfront.

    That preference list was formed before any agent was launched. It reflects what the user already believed about brands in a category — the brands they think of as credible, relevant, and worth considering. That belief was shaped by what they have read, heard, and asked AI assistants over weeks and months.

    Which is exactly what AI training knowledge and entity confidence determines.

    The shortlist a user gives their agent is, in practice, a projection of the AI’s own brand knowledge back through the user. A user who regularly asks ChatGPT about project management tools, CRM platforms, or accounting software has absorbed an AI-mediated view of their category. The brands that appeared reliably in those answers — cited clearly, described accurately, associated with the right attributes — are the brands that make the agent’s preference list.

    The brands that didn’t appear, or appeared inconsistently, or were confused with a competitor, are not on the list. They cannot optimise their way onto it now. The list is closed.

    The Agent Doesn’t Browse. It Recalls.

    There is an important distinction between what retrieval-optimised content does and what the agent draws on when it shortlists.

    Retrieval optimisation — the discipline of structured data, BLUF formatting, semantic HTML, citation-ready content — affects what happens when an AI engine executes a live query. The agent fetches pages, scores them, extracts information, and constructs an answer. Content quality determines whether your brand makes that answer.

    But when an information agent is monitoring a topic on a user’s behalf, it is not starting from scratch on every cycle. It is drawing on its established understanding of the category — who the credible players are, what they offer, which ones are reliably accurate — before it retrieves anything new. The shortlist it builds reflects both what it retrieves and what it already knew.

    This is the distinction the wiki captured as “training mode” versus “retrieval mode.” For popular entities, training-time knowledge dominates. The model’s answer about who the leading vendors in a category are reflects patterns absorbed across billions of documents, not the page it retrieved this morning. Retrieved content updates and corrects at the margins; it does not override.

    For mid-sized brands — the ones not prominent enough to have saturated training data — retrieval matters more. Fresh, well-structured earned coverage in credible sources can shift what the agent finds and reports. But the baseline the agent works from is still its existing category understanding.

    Which means: retrieval optimisation keeps you visible to agents actively searching. Entity confidence determines whether you were on the shortlist before the search began.

    What This Means in Practice

    Google’s information agents launch first for paid subscribers (Google AI Pro and Ultra) in the US this summer. Universal Cart rolls out across Search and Gemini, with YouTube and Gmail to follow. The distribution is initially constrained.

    But AI Mode already has one billion monthly users. Queries are doubling every quarter. The agents are the next layer, not a separate product. The direction of travel is unambiguous: AI-mediated purchase journeys are moving from “the user searches and decides” to “the user delegates and approves.”

    In that model, the purchase decision has already been shaped before the agent receives its instructions. The category shortlist, the trusted vendors, the brands worth considering — these are background knowledge that the user carries into the agent conversation. They were formed over months of AI-assisted research, not in the moment of purchase.

    For brands, the practical implication is uncomfortable. The content sprint, the PR push, the structured data audit — these are retrieval optimisation. They affect what the agent finds when it searches. They do not reach back in time to correct the category knowledge the user already has. They don’t change what the user tells their agent to prefer.

    The gap between “what we optimised for” and “what the agent knew before it started” is the gap where most brands are losing.

    The Entity Confidence Response

    The question is not whether information agents change things. They do. The question is what brands can do about it, and on what timeline.

    Three things are in scope:

    Build category knowledge in AI systems now, not later. The training-mode channel — community presence, third-party editorial coverage, Wikipedia-adjacent authority signals — shapes what AI systems know about a category before any retrieval happens. This is the long cycle. It cannot be compressed. Brands that start it now are building the knowledge base that will be in the next generation of training data; brands that wait are not.

    Make your entity unambiguous. If an AI agent is shortlisting vendors in your category, the first thing it needs to do is correctly identify who you are. Identity drift — the failure mode where a brand appears under multiple name variants without canonical linking — ensures you are either absent from the shortlist or confused with a competitor. Before any agent can recommend you, it must know that the company the user mentioned, the domain it found, and the brand in the editorial coverage are the same entity. That is a structured data and canonical naming problem. It is solvable today.

    Measure what agents can see, not what analytics can count. Zero-search discovery is invisible to standard analytics. No query, no impression. No click, no session. No session, no attribution. The brands that will navigate this era are the ones that begin measuring AI citation frequency and entity representation now — not as a supplement to traffic data, but as the leading indicator for a category of brand influence that traffic cannot capture.

    The Closing Argument

    Google’s information agents are not a feature. They are the endpoint of a direction that has been visible for three years: AI moving from answering questions to completing tasks; from citing brands when asked to recalling brands when needed.

    The brands that win in that model are not the ones that appear when searched. They are the ones that were already part of the AI’s understanding of the category — present in training data, clearly identified as entities, accurately described across multiple independent sources — before any user opened a search box.

    You cannot optimise your way onto a shortlist that closed before the query.

    The work that determines whether your brand is on that list is happening now, in the slow accumulation of editorial coverage, entity signals, and community mentions that shape what AI systems know. It is not a campaign. It is not a sprint. It is the background condition on which every future agent decision will be based.

    Start building it before the agents start working.

  • E-E-A-T Is Entity Confidence

    E-E-A-T Is Entity Confidence — So Why Don’t We Say That?

    Summary: A provocation piece arguing that Google’s E-E-A-T framework and the practitioner concept of “entity confidence” describe identical phenomena — and examining what is genuinely new about the AI measurement layer.

    Last updated: 2026-07-07

    Type: Original article / insight post

    The GEO industry has spent two years building new vocabulary. Entity confidence. Citation frequency rate. AI visibility scores. Semantic authority. Confidence language analysis. These terms fill product decks, vendor websites, and optimisation guides published from 2024 onwards.

    Google invented most of this framework in 2014 and called it something different.

    But there is a catch — and it matters more every month. E-E-A-T was designed for one search engine. Half the search behaviour it was built for is now happening somewhere else entirely.

    What E-E-A-T Measures

    E-E-A-T — Experience, Expertise, Authoritativeness, Trustworthiness — is Google’s formal quality framework. Google’s HJ Kim described it as “a template they use to rate every single site for every single query.” Marie Haynes distilled the core of it more precisely: “E-E-A-T is a measure of the legitimacy of your entity as a destination for the topics you cover.”

    The measurement mechanism was described by Gary Illyes at Pubcon 2018: “E-A-T is largely based on links and mentions on authoritative sites. If the Washington Post mentions you, that’s good.”

    Danny Sullivan made the ranking connection explicit: E-E-A-T is not a directly disclosed score but a cluster of proxy signals — third-party mentions, backlinks, review reputation, entity recognition — that approximate what human quality raters would assess.

    And the loop closes formally through Pandu Nayak’s antitrust testimony: quality rater assessments generate the Information Satisfaction (IS) Score, and IS-scored documents train the deep learning systems that power Google Search. E-E-A-T signals → rater feedback → IS Score → ranking model training. It is not theoretical; it is a documented production mechanism.

    What Entity Confidence Measures

    Entity confidence score — as described by the vendors building AI citation measurement tools — is a composite measure of how certain an AI model is about the accuracy and authority of information associated with a brand. Primary vendors measure it through: citation frequency (how often AI mentions a brand for relevant queries), confidence language (whether AI uses “according to Brand X” versus “some sources suggest”), and response position (whether the brand appears first or sixth in an AI answer).

    The signals that build a high entity confidence score: authoritative off-site mentions, structured entity data (schema, sameAs markup), cross-platform consistency, third-party recognition from credible sources.

    The Mapping

    Hold the two frameworks side by side:

    E-E-A-T ComponentEntity Confidence Equivalent
    Authoritativeness — cited by credible third partiesRelationship Mapping — industry ecosystem recognition
    Expertise — topic depth, credentials, definitional claritySemantic Authority — topic clustering, definitional precision
    Experience — first-hand, original knowledgeInformation gain, non-AI-replicable original content
    Trustworthiness — transparent, consistent, no manipulationContextual Consistency — coherent, verified, cross-platform aligned

    The signal overlap is not approximate — it is complete. E-E-A-T is built from off-site mentions, links, entity recognition, review reputation, schema, and community presence. Entity confidence is built from the same list.

    The data confirms this at the empirical level. Ahrefs analysed 75,000 brands against AI Overview citation outcomes. The strongest correlate with AI citation was branded web mentions (off-site) at a correlation coefficient of 0.664. Not schema. Not content length. Not author bios. Earned media mentions — the Gary Illyes signal, showing up again in 2025 data for an entirely different platform.

    The finding replicates across independent datasets. 82% of AI citations across 1M+ analysed prompts came from earned media (Muck Rack/MacroLingo). And 76% of AI Overview citations came from pages already in the top-10 search results — meaning AI answers are substantially inheriting Google’s E-E-A-T rankings rather than computing something new from scratch.

    Did AI Engines Inherit E-E-A-T — or Reinvent It?

    The most important question the data raises is whether AI citation systems independently arrived at the same signal set as Google, or simply inherited it.

    For Google’s own AI surface, the evidence points clearly to inheritance. In the Ahrefs 1.4M-prompt study, 88% of ChatGPT citations came from the search channel — pages indexed and ranked in traditional web search. For Google AI Overviews and AI Mode, this is even more direct: those systems are grounded against Google’s index by design. AI Overviews are, to a significant degree, Google ranking with a synthesised output layer.

    The chain is documented end to end for this surface:

    E-E-A-T signals → off-site mentions and links → IS Score trains ranking models → pages rank in search → AI retrieval pulls from ranked pages → AI citation emerges

    There is no layer in that chain where AI engines are separately deciding what constitutes a trustworthy source. They are, in large part, delegating that judgment to Google’s ranking systems — which were themselves trained on E-E-A-T rater feedback.

    But this chain only describes part of what is now happening.

    The Search That Google Doesn’t See

    In June 2025, OpenAI published its first major usage research. The finding that matters here: 51.6% of all ChatGPT interactions are now search-like information queries — the primary use case for the platform, having overtaken content generation in a single year. ChatGPT has 700 million weekly active users. At 11–12% monthly growth from early 2025, it passed the mass adoption threshold around April 2026.

    These are not Google searches with a different interface. They are queries going directly to an AI engine that is not Google — and they are not passing through Google’s ranking systems before generating a response.

    The University of Toronto’s peer-reviewed study of 1,000+ queries (Chen et al., arXiv:2601.16858, January 2026) makes the structural difference measurable. Domain-level overlap between GPT-4o’s citations and Google’s top-10 results: 4%. Claude’s overlap: 12.6%. Perplexity’s: 15.2%.

    GPT-4o is making its own authority decisions on 96% of the domains it cites. It is not routing those decisions through Google’s ranking systems. For that 96%, E-E-A-T → search ranking → AI citation is not the chain. The AI is drawing on training data, its own retrieval logic, and entity signals that exist independently of Google’s index.

    The pre-training bias finding from the same study sharpens this further. For well-known brands, AI rankings remain highly stable even when retrieved supporting evidence is removed or shuffled entirely — the model’s pre-trained understanding of brand authority dominates regardless of what is currently ranking on Google. For those queries, Google’s ranking is not the input. Training data is. And training data is shaped by years of accumulated web content, entity signals, and third-party coverage — not by last week’s ranking positions.

    Zero-click search has been driving this shift at the structural level: approaching 65–70% of queries now resolve in AI-generated answers without a click-through. Google AI Mode runs at approximately 93% no-click. The search that is happening is increasingly not search in the traditional sense — it is AI-mediated information retrieval across multiple platforms, only some of which route through Google.

    Then What Is Entity Confidence Actually Adding?

    If E-E-A-T is entity confidence, the concept is not new. But what it covers — and how it is measured — genuinely is.

    E-E-A-T was built to track brand authority in one system: Google Search. The signals are right. The gap is the scope.

    Entity confidence, properly defined, covers two surfaces that E-E-A-T tracks only partially:

    The Google-adjacent surface: AI Overviews, Google AI Mode, and AI systems that retrieve primarily through search indices. Here, E-E-A-T signals are the right inputs and search ranking is a reasonable proxy for AI citation eligibility. This is the surface where the inheritance chain holds.

    The direct AI surface: ChatGPT queries without web search enabled, Claude, Perplexity, and all AI interactions where the model answers from training knowledge or its own retrieval logic rather than Google’s index. Here, E-E-A-T signals still matter — the underlying signals (earned media, entity clarity, cross-platform consistency) are what train these models and what their retrieval systems learn to trust. But Google ranking is not a reliable proxy for AI citation on this surface. A brand can hold Google position 3 and have no meaningful AI presence if its entity signals are thin in the training corpus.

    This is the surface that is growing fastest. The 51.6% of ChatGPT interactions that are search-like queries is not a static number — it is a rising proportion of a user base that grew 75% in five months.

    The second genuine addition is the measurement layer. E-E-A-T has no disclosed score. It is tracked by proxy: rank improvement, traffic, domain authority. These are position-based, deterministic metrics.

    AI citation is stochastic. Only 30% of brands maintain consistent visibility across multiple regenerations of the same query. A brand is not in AI position 3. It appears in approximately 65 out of every 100 relevant AI responses. That is a probability distribution, not a rank — and it can differ significantly across AI platforms using different retrieval architectures and training histories.

    The third addition is the narrative framing layer. Google returns a ranked list. AI platforms synthesise a recommendation and present it as a direct answer. There is a meaningful difference between being described as “the industry standard” and “a budget-friendly alternative” — both are appearances, but at very different confidence levels. That qualifier is invisible in any rank tracking tool.

    What This Means for Strategy

    Brands being sold AI visibility programmes as something new and distinct from their existing SEO and PR investment should ask: what specifically are we being asked to do that we are not already doing?

    For the Google-adjacent AI surface, the honest answer is: at the foundational level, very little. The inputs are shared:

    • Build authoritative off-site mentions → E-E-A-T Authoritativeness = EC Relationship Mapping = earned media programme
    • Demonstrate expertise in original content → E-E-A-T Expertise/Experience = EC Semantic Authority = owned content programme
    • Maintain entity clarity via schema and sameAs markup → E-E-A-T entity signals = EC entity disambiguation = technical SEO
    • Build consistent brand presence across platforms → E-E-A-T Trustworthiness = EC Contextual Consistency = brand management

    For the direct AI surface — the faster-growing, non-Google half — the inputs are still the same signals. But the tracking is not. Google rankings tell you nothing about how Claude or ChatGPT without web search represents your brand. Measuring that requires different instruments: repeated sampling across regenerated prompts, platform-specific citation profiling, confidence language analysis. This is what EC measurement adds that rank tracking cannot.

    The framing shift matters too. Google Search is where you compete for rank position. Direct AI search is where AI decides whether to recommend you based on what it has learned about your brand — and that learning happened before the user typed a word.

    The Vocabulary Question

    Why doesn’t the entity confidence industry say “we’re measuring E-E-A-T for AI engines”?

    The uncharitable read: novelty sells. New vocabulary justifies new products and new budget lines.

    The more generous read — and the more accurate one — is that E-E-A-T is a Google framework, and the problem is now multi-platform in a way it was not when the GEO vocabulary was being built. Saying “build your E-E-A-T” implies Google Search is the destination. Saying “build your entity confidence” points to the full citation surface: Google AI Overviews, ChatGPT, Perplexity, Claude, Gemini — including the surfaces that have no relationship to Google’s ranking systems.

    That vocabulary change reflects a real scope change. The underlying signals — what you need to do to build brand authority — are E-E-A-T. The measurement layer, the platform coverage, and the probabilistic framing of visibility scores are new. And the urgency is new: a search behaviour that took Google a decade to accumulate is now being replicated across multiple AI platforms at mass-adoption pace.

    The question worth asking is not whether your E-E-A-T programme is doing its job. It is whether you know what your brand looks like on the surfaces that E-E-A-T tracking was never designed to see.