Claims we reject, and why
AI search has produced a fast-growing body of advice, and much of it has never been tested. Some of it conflicts with what the platforms themselves publish. This page is a public ledger of popular claims about ranking and AI visibility, each with a verdict and the evidence behind it.
Our measurement study forms the backbone of this ledger. Today's verdicts rest on what the platforms have published. As the study runs, entries gain a further section recording what we measured: the same real small-business questions, put to the major AI systems again and again, with cited and uncited pages measured side by side. Where a claim is testable, we say so. Where our data contradicts our own verdict, we change the verdict. Anchoring the ledger to an ongoing study is what separates evidence from criticism.
Three rules govern every entry. Every verdict cites its evidence. Verdicts change when evidence changes. We publish reversals, because a ledger that records only the times we were right would amount to marketing.
How to read a verdict
Contradicted conflicts with authoritative platform guidance or reliable evidence. Unsupported is commonly repeated without adequate evidence. Context-dependent holds only under conditions the advice usually omits. Experimental is worth testing and not yet established.
1 · "Create an llms.txt file to rank in Google's AI results"
Contradicted (for Google) · last reviewed 2026-08-04
Google states that it does not use llms.txt for Search or for its generative features. Creating the file cannot improve visibility in AI Overviews or AI Mode, and advice that sells it as a Google ranking intervention conflicts with the operator's own statement.
What is defensible: the file costs little to publish and other systems may read it. Treated as an optional experiment for non-Google assistants, it does no harm. Advice that sells it as an AI-ranking fix belongs in this ledger.
Source: Google Search Central, Guide to optimizing for generative AI features on Google Search.
2 · "Break your articles into small, AI-friendly chunks"
Contradicted (as a universal rule) · last reviewed 2026-08-04
Google states that no special chunking is required for its AI features and that there is no ideal page length. Sections and headings should follow the needs of the reader, which good writing required long before AI search existed.
What is defensible: clear structure helps readers and may help systems isolate a relevant passage. That argues for well-organised documents. It does not argue for slicing content to a rumoured machine-preferred size.
Source: Google Search Central, Guide to optimizing for generative AI features on Google Search.
3 · "Add special AI schema to your pages"
Contradicted · last reviewed 2026-08-04
No recognised structured-data vocabulary grants access to Google's AI Overviews or AI Mode. Google states that structured data is not a special lever for generative-search inclusion. Any product selling "AI schema" is selling a thing the platforms say does not exist.
What is defensible: ordinary structured data, describing what is visibly true on the page, supports explicit interpretation and rich-result eligibility. It is a useful foundation and not a citation mechanism.
Source: Google Search Central, Guide to optimizing for generative AI features on Google Search.
4 · "Write a separate page for every long-tail prompt"
Contradicted (as a strategy) · last reviewed 2026-08-04
Google confirms that its AI features issue multiple related searches for one question. The same guidance warns that mass-producing pages to chase query variations can constitute scaled-content abuse. The strategy aims at the mechanism and walks into the penalty.
What is defensible: covering the legitimate sub-questions of a topic, where each section has genuine value to a reader. One useful resource beats a hundred near-duplicates, by Google's stated policy as well as by taste.
Source: Google Search Central, Guide to optimizing for generative AI features on Google Search.
5 · "AI-generated content is automatically penalised"
Contradicted · last reviewed 2026-08-04
Google's published position since 2023: it does not ban content for having been produced with AI assistance. It polices the result, meaning whether the content meets quality standards and whether someone produced it at scale to manipulate rankings.
What is defensible: the caution underneath the myth. AI-assisted commodity content, the generic summary any system could produce, is what Google's current guidance says it does not want to surface. The problem is commodity, whoever the author. Our position on our own process is on the About page: drafted with AI, refined and signed by a named human, evidence required throughout.
Sources: Google Search Central blog, Google Search's guidance about AI-generated content (2023); Guide to optimizing for generative AI features.
6 · "Structured data makes AI systems cite you"
Unsupported (as a citation lever) · last reviewed 2026-08-04
Structured data can make pages eligible for enhanced search displays, and it gives machines an explicit statement of what a page describes. No platform states, and no adequate evidence shows, that it causes AI citation. Google cautions against overestimating it as a generative-search tactic.
What is defensible: structured data as foundation: accurate, matching the visible page, maintained. Our study can inform this claim, because we measure cited and uncited pages alike for schema presence and type. We will update this entry with what we find.
Source: Google Search Central, Guide to optimizing for generative AI features on Google Search.
7 · "More backlinks means more AI citations"
Unsupported (as stated) · last reviewed 2026-08-04
Links and external reputation support discovery and authority, and no platform disputes that. Citation in a generated answer also depends on relevance to the specific question, on what the page contains, and on selection behaviour that varies by system and by run. A general link-building campaign is a different thing from a citation strategy, and manufactured or reciprocal links can register as spam.
What is defensible: genuine external corroboration, meaning independent sources your customers already trust describing you accurately. Which sources matter, and how much, is a question our study is designed to measure.
Sources: Google Search Central, Guide to optimizing for generative AI features; spam policies on link schemes.
8 · "This tool's AI visibility score predicts your citations"
Unsupported · last reviewed 2026-08-04
No third party has access to the internal systems of Google, OpenAI or Perplexity. Citation output varies between engines, between phrasings of the same question, and between repeated runs of the identical question. We measure that variation directly. A single score claiming to predict citation probability compresses away the uncertainty that defines the territory, and Google warns against tools claiming knowledge of internal metrics.
What is defensible: measuring observable readiness. Can systems fetch your pages, is the content extractable, is the entity clear? Report those as separate dimensions with their limits stated. Measurement can offer that honestly. It cannot offer a prediction.
Source: Google Search Central, Guide to optimizing for generative AI features on Google Search.
Claims under test
This is where the ledger grows. The following claims sit in neither the rejected nor the endorsed column. They are registered hypotheses in our study, published before data collection, and each will receive its evidence-backed verdict as results come in:
- AI systems cite third-party editorial sources more than business-owned pages for "best provider" questions, and the reverse holds for specific factual and transactional questions.
- Pages containing original evidence, such as figures, first-hand examples and stated methodology, are cited more often than comparable pages without it.
- Citation sets are less stable than conventional top-ten rankings.
- The sources cited most often are not always the sources that most shape the answer's content.
The full list of ten registered hypotheses, and the method for testing them, is on the methodology page.
Submit a claim
Seen advice that belongs on this ledger? Send it to us, with a link to where the claim is being made. We add entries when a claim is widespread enough to matter and specific enough to test.
Publication of an entry means the claim as commonly stated lacks adequate support. It does not mean the person repeating it acts in bad faith. Most folklore spreads in good faith, which is why a ledger is needed.