Skip to main content Scroll Top

AEO Myths: Twelve Claims, Assessed Against Evidence

Updated September 2026 · Written and maintained by the Progression Agency strategy team

Ten of them are false and the two true ones are the cheapest things on the list — plus, for each, how to check it yourself and why the myth survives.

On this page · 14 sections
  1. Twelve claims, assessed
  2. Myth: answer engine optimization replaces SEO
  3. Myth: SEO is dead
  4. Myth: llms.txt is the key to AI visibility
  5. Myth: only big brands get cited
  6. Myth: more content means more citations
  7. Myth: rank tracking measures AI visibility
  8. The claims that are genuinely half true
  9. How to check any AEO claim yourself
  10. Everything else we have written on search, AI and getting found
  11. The twelve claims, in one table
  12. What each myth costs you if you act on it
  13. Where each verdict comes from
  14. The anatomy of an AEO myth

The short answerTen of the twelve claims circulating about answer engine optimization are false. AEO does not replace SEO, SEO is not dead, llms.txt is not the key, placement cannot be bought, volume does not drive citation, keyword density does nothing, rank tracking cannot measure it, and ‘only big brands get cited’ is true on definition queries and false on commercial ones. The two true claims are the cheapest actions available: check crawler access first, and make each section answer its own heading.

Each verdict is labelled by evidence class — provider documentation, our own September 2026 measurement, direct testing, or our observation from client work. No claim is made about internal ranking weights, which no provider publishes.

Twelve claims, assessedTwelve claims, assessed
Ten false, two true. The two true ones are the cheapest things on the list, which is part of why the false ones stay popular — they are more interesting.

Twelve claims, assessed

Ten of the twelve claims circulating about answer engine optimization are false, and the two true ones are the cheapest things on the list. That asymmetry is not an accident — the false claims are simply more interesting than ‘check your robots.txt’.

For each one below we have separated four things that usually get mixed together: what is claimed, what is actually true, how you can check it yourself today, and why the myth persists. The last of those is the one nobody writes about, and it is usually the most informative.

Where a claim is half true we say so rather than scoring it as false, because the half-true ones are where most of the real confusion lives.

Myth 1 — AEO replaces SEO. Answer engines retrieve from indexes SEO builds.
Myth 2 — SEO is dead. Structurally impossible in the direction claimed.
Myth 3 — llms.txt is the key. Low adoption, not a ranking factor.
Myth 4 — You can buy placement. No ad inventory exists at the attribution step.
Myth 5 — More content wins. Retrieval selects passages, not volume.
Myth 6 — Keyword density matters. Similarity is computed on meaning.
Myth 7 — Rank tracking measures it. No ranked list exists inside an answer.
Myth 8 — Only big brands get cited. DA 17 holds a page-one commercial position.
Myth 9 — Schema alone does it. Markup contradicting the page is worse than none.
Myth 10 — AEO is a pure rebrand. Two workstreams are genuinely new.
True 1 — Check crawler access first. Free, minutes, binary outcome.
True 2 — Passage structure matters. Directly serves the extraction step.

Myth: answer engine optimization replaces SEO

It cannot, in the direction people claim. Answer engines retrieve from web indexes, so a page that is not indexed cannot be pulled into an answer. Remove the search work and the answer work has nothing to stand on.

What is true is that emphasis shifts: more hours on technical work and measurement, fewer on content volume. Two workstreams are genuinely new — a second set of crawlers, and a rendering standard stricter than Google’s. That is a real change, and it is not a replacement.

Myth: AEO replaces SEOMyth: AEO replaces SEO
The first four bars are near-identical, which is the answer. Answer engines retrieve from the indexes search optimization builds, so one cannot replace the other in the direction people claim.

What is claimed

That AEO is the successor to SEO and budget should move wholesale.

What is true

AEO depends on SEO. The dependency runs one way and only one way.

How to check it

Take any page that is not indexed and ask an index-led surface a question it answers. It will not be cited, because it was never a candidate.

Why the myth persists

It is a better pitch. ‘Everything changed’ sells a new retainer more easily than ‘most of your existing programme still applies’.

Myth: SEO is dead

The same claim wearing a shorter name, and it fails for the same structural reason. It is also the most expensive myth on this list to act on, because it removes the foundation retrieval depends on.

The honest version: organic click behaviour on informational queries has changed, AI referrals are around one per cent of web traffic in the largest study with a stated denominator, and the index engines still carry the overwhelming majority of demand.

What is claimed

That search traffic is collapsing and search optimization no longer pays.

What is true

Informational click behaviour has shifted. Commercial and navigational demand largely has not. The index engines remain the retrieval substrate for the answer engines.

How to check it

Look at your own Search Console data by query type. The pattern is usually visible and specific rather than universal.

Why the myth persists

It is dramatic, it is partly observable, and it has been repeated annually for well over a decade — which gives it an air of eventual inevitability.

It is dramatic, it is partly observable, and it has been repeated annually for well over a decade — which gives it an air of eventual inevitability.

Myth: llms.txt is the key to AI visibility

Adoption is low, no major provider treats it as a ranking factor, and we have seen no evidence that adding one changes citation frequency. It is hygiene. Add it if you like; do not expect it to do anything.

This one is worth dwelling on because its life cycle is the clearest example of how an AEO myth forms: a real proposal, a confident write-up with the caveats stripped, widespread quotation, and now a small industry of generators that exist to produce one.

How an AEO myth is bornHow an AEO myth is born
The llms.txt story follows this arc precisely: a real proposal, a confident write-up, widespread quotation, and now a set of tools that exist to generate one.

What is claimed

That publishing an llms.txt file makes your site legible to language models and materially improves citation.

What is true

It is a proposed convention with limited adoption. Nobody has published evidence that it affects retrieval or selection.

How to check it

Ask any provider’s documentation whether it is consumed. The absence of an answer is the answer.

Why the myth persists

It is cheap, it feels like doing something, and it is one of very few AEO actions that produces a visible artefact. Tools now exist to generate it, which gives the idea commercial momentum independent of whether it works.

Myth: only big brands get cited

Half true, and the half that is wrong is costing small sites real opportunity. Our own September 2026 capture found a domain with an authority score of 17 holding a page-one position on a commercial services term, and a DA-18 domain at position one on a tool query worth $45.41 a click.

What is true is that definition queries really are publisher-locked. The lowest page-one domain authority we measured on ‘what is answer engine optimization’ was 56, with Forbes, HubSpot, PwC and Coursera above it. Stating this as one universal rule is what makes it wrong.

Myth: only high-authority sites get citedMyth: only high-authority sites get cited
From our own published dataset. The myth is half true: definition queries really are publisher-locked, and commercial and tool queries really are not. Stating it as one rule is what makes it wrong.

What is claimed

That AI answers favour established brands and a small site cannot compete.

What is true

It depends entirely on query type, with roughly forty points of domain authority separating the open end from the closed end of the same topic family.

How to check it

Pick a commercial query in your category, look at who actually ranks, and check their domain authority. It takes ten minutes.

Why the myth persists

Because the queries people test first are definition queries, which are the locked ones. The test that produces the myth is the wrong test.

Myth: more content means more citations

This is the most expensive false belief on the list, because acting on it misallocates budget every single month. Retrieval selects a passage, not a page, and it does not reward volume.

One page whose section answers the question in its first sentence beats twenty that circle the topic. We have watched sites publish steadily for months with no change in citation frequency, then move by restructuring pages they already had.

What believing each myth costs youWhat believing each myth costs you
The two eights are the expensive ones. Publishing volume misallocates budget every month, and rank tracking hides the misallocation from you.

What is claimed

That publishing cadence drives AI visibility the way it once drove indexation.

What is true

Structure drives extraction. Volume drives nothing in particular, and thin volume can actively hurt through templating signals.

How to check it

Take your best-performing existing page, move the answer to the first sentence of the section, and watch extraction rather than publishing something new.

Why the myth persists

Content volume is measurable, schedulable and sellable. Structural rewriting of existing pages is none of those things, so it loses the argument internally.

Myth: rank tracking measures AI visibility

It cannot, and this one is uniquely damaging because it does not merely fail — it reports success while you are absent from every answer a buyer receives.

There is no ranked list inside a generated answer. Three or four sources are named and there is no second page. A rank tracker is measuring a different surface entirely.

What is claimed

That search position is a reasonable proxy for AI citation.

What is true

They are different surfaces with different retrieval. A site can hold position one and be absent from every AI answer in its category.

How to check it

Compare your top-ranking page against the same question asked in two assistants. The disconnect is usually immediate.

Why the myth persists

Because the alternative — a fixed prompt set, run repeatedly, with answers stored verbatim — is more work, and because rank data is already in everyone’s reporting stack.

Because the alternative — a fixed prompt set, run repeatedly, with answers stored verbatim — is more work, and because rank data is already in everyone’s reporting stack.

The claims that are genuinely half true

Six of them, and they cause more confusion than the outright false ones because arguing against them feels like denying something obvious. In each case the claim is true in a narrow condition and false as a general rule.

Schema helps — when it agrees with the visible page, and not otherwise. Backlinks matter less directly — and still drive the indexation retrieval depends on. It is too early — for local trades, yes; for question-led B2B, no.

Only big brands get cited — Half true. True on definition queries, false on commercial.
Schema helps — Half true. When it agrees with the visible page.
Backlinks are less important — Half true. Less direct, still drive indexation.
AI traffic is tiny — Half true. ~1% today, and growing.
Content still matters — Half true. Structure matters; volume does not.
It is too early — Half true. True for local trades, false for B2B.
Why each myth survivesWhy each myth survives
Top-right is the dangerous quadrant: easy to disprove, and someone is selling it anyway. Bottom-left is ordinary confusion rather than anything cynical.

Want this done for your site?We build and maintain the search, content and paid programmes described on this page.

Get a free proposal

How to check any AEO claim yourself

Four questions settle most of them, and the whole process takes under an hour. Is it documented by a provider? Is it observable by testing? Does anyone outside the labs actually know it? And who profits if you believe it?

That fourth question is not disqualifying on its own — plenty of true things are also profitable to say. It is worth noticing specifically on claims that are hard to check, because that combination is where bad advice survives longest.

How to check any AEO claim yourselfHow to check any AEO claim yourself
This is the method behind every assessment on this page, and it is more useful than the assessments themselves.
Provider docs — Check. Crawler behaviour is published.
Test it yourself — Check. Most claims collapse within an hour.
Is it knowable? — Check. Ranking weights are published by nobody.
Who profits? — Check. Worth noticing on hard-to-check claims.
Record the result — Check. Dated notes beat recalled impressions.
Repeat later — Check. Platform behaviour genuinely changes.
How each verdict on this page is supportedHow each verdict on this page is supported
The two ‘part’ rows are labelled wherever they appear in the text. The red row is why this page makes no claim about why a specific source was chosen.
Evidence — Provider documentation. Crawlers, robots.txt, bot policies.
Evidence — Our own SERP capture. 104 positions, September 2026.
Evidence — Direct testing. Scriptless fetch, user-agent requests.
Evidence — Retrieval research. Published RAG literature.
Labelled — Industry consensus. Flagged in the text where used.
Not claimed — Internal ranking weights. Published by nobody.
The myths, in numbersThe myths, in numbers
The first two zeros settle four of the twelve claims on their own. The gap between 17 and 56 settles a fifth.
Guarantees — Costs you. Money for something undeliverable.
Content volume — Costs you. The largest ongoing misallocation.
Rank tracking — Costs you. Hides the misallocation from you.
'SEO is dead' — Costs you. Removes the foundation retrieval needs.
llms.txt focus — Costs you. Cheap, but displaces the real work.
Density habits — Costs you. Wastes effort and degrades the writing.

Want the claims about your own site checked?

An AI visibility audit tests the checkable ones directly — crawler access, rendering, passage structure — and records a dated baseline, so the next set of claims can be measured rather than argued about.

/ai-visibility-audit

Everything else we have written on search, AI and getting found

AI, AEO and what is changing

Websites and design

Choosing and working with an agency

Social, content and brand

By industry and by situation

Not sure which of these applies to you?Tell us the situation and we will say plainly what we would do first, and what we would not.

Talk it through

The twelve claims, in one table

Twelve widely repeated claims, each with a verdict and the basis for it.

Claim, verdict, and how you can check it today
ClaimVerdictWhat is actually trueHow to check it yourself
AEO replaces SEOFalseAnswer engines retrieve from indexes SEO buildsAsk an index-led surface about an unindexed page
SEO is deadFalseInformational clicks shifted; demand did not collapseSegment your Search Console data by query type
llms.txt is the keyFalseLow adoption, not a ranking factorCheck whether any provider documents consuming it
You can buy placementFalseNo ad inventory at the attribution stepLook for the purchase mechanism. There is none
More content winsFalseRetrieval selects passages, not volumeRestructure one existing page instead of publishing
Keyword density mattersFalseSimilarity is computed on meaningCompare a dense page against a clearly-written one
Rank tracking measures itFalseNo ranked list exists inside an answerCompare a page at position 1 against the same question asked
Only big brands get citedHalf trueTrue on definition queries, false on commercialCheck the DA of who ranks for your commercial terms
Schema alone does itHalf trueHelps when it agrees with the visible pageValidate markup, then read the page against it
Backlinks no longer matterHalf trueLess direct, still drive indexationCheck indexation on a page with no links
Check crawler access firstTRUEFree, minutes, binary outcomeRead robots.txt for the five AI agents
Passage structure mattersTRUEDirectly serves the extraction stepRead a section standalone and see if it holds

What each myth costs you if you act on it

The cost of believing each claim, expressed as what you would waste or lose by acting on it.

Ranked by damage rather than by how often it is repeated
MythCost if believedWho typically pays
Buy guaranteed citationsMoney spent on something undeliverableWhoever signed the contract
More content winsThe largest recurring budget misallocation in the fieldContent budget, monthly
Rank tracking measures itReports success while you are absentEveryone, invisibly
SEO is deadRemoves the foundation retrieval depends onLong-term organic performance
Only big brands get citedOpportunity cost — you never enter winnable queriesSmall and mid-size sites
llms.txt is the keyCheap in itself, but displaces the real workA quarter of wasted focus
Keyword density mattersWasted effort and worse writingReaders, then extraction

Where each verdict comes from

The basis for every verdict above, separated into vendor documentation and our own observation.

Evidence class for every assessment on this page
VerdictEvidence classSource
Crawler and robots.txt behaviourDocumentedProvider documentation
Passage-level retrievalDocumentedPublished RAG research
JavaScript execution by crawlersObservableTest with curl on your own site
Sources named per answerObservableCount them
Domain authority by query typeOur own measurementSeptember 2026 SERP capture, published
AI referral share (~1%)Cited, not measured by usThird-party study with a stated denominator
Content volume does not drive citationOur observationClient work, labelled as such
Internal ranking weightsNot claimedPublished by nobody

The anatomy of an AEO myth

Most of these follow the same pattern, and recognising it is more useful than memorising the list.

The same six stages every time
StageWhat happensllms.txt example
Real observationSomething true in a narrow caseA convention is proposed for describing sites to models
Confident write-upCaveats removed for readabilityPosts describe it as the AI visibility file
QuotationCited without original contextOther posts cite those posts
ConsensusAppears widely enough to look settledIt becomes a standard checklist item
Commercial attachmentProducts get built on itGenerators and audits that check for it
PersistenceDisproving it now costs someone moneyThe claim outlives the evidence for it

Getting found in search

AI, AEO and what is changing

Paid media and lead generation

Websites and design

Choosing and working with an agency

Social, content and brand

By industry and by situation

Frequently asked questions

Is SEO dead because of AI?
No. Answer engines retrieve from the web indexes that search optimization builds and maintains, so removing that work removes the foundation. What has changed is click behaviour on informational queries, not the existence of search demand.
Does AEO replace SEO?
No, and it cannot in the direction people claim. AEO depends on SEO. Only two workstreams are genuinely new: a second set of crawlers, and a rendering standard stricter than Google’s.
Does llms.txt improve AI visibility?
There is no published evidence that it does. Adoption is low and no major provider documents it as a ranking factor. Treat it as hygiene rather than strategy.
Can you pay to appear in AI answers?
No. There is no ad inventory, no submission process and no paid tier at the attribution step. Any offer of guaranteed citation is describing something that does not exist.
Does publishing more content improve AI citation?
No. Retrieval selects passages rather than rewarding volume. Restructuring one existing page so its section answers its own heading typically returns more than publishing twenty new ones.
Does keyword density matter for AI search?
No. Similarity is computed on meaning, so repeating a phrase does not increase your odds of retrieval, and dense writing tends to extract worse rather than better.
Can a rank tracker measure AI visibility?
No. There is no ranked list inside a generated answer. Worse, a programme measured this way will report steady success while the brand is absent from every answer a buyer receives.
Do only big brands get cited by AI?
Half true. Definition queries are publisher-locked — the lowest page-one domain authority we measured on ‘what is answer engine optimization’ was 56. Commercial and tool queries are not: we measured a DA-17 domain holding a page-one commercial position.
Is schema markup enough on its own?
No. Schema helps when it agrees with the visible page, and markup that contradicts the copy is worse than no markup at all.
Do backlinks still matter?
Yes, less directly. They drive the indexation that retrieval depends on, and third-party mentions function as corroboration at the reliability step.
Is AEO just SEO rebranded?
No. Two things are genuinely new: a second set of crawlers that must be separately allowed and verified, and a rendering standard considerably stricter than Google’s. The rest is emphasis.
What is actually true about AEO?
Two cheap things: check crawler access first, because it is free, takes minutes and is binary; and structure each section so it answers its own heading in the first sentence, because that directly serves the extraction step.
How can I check an AEO claim myself?
Four questions: is it documented by a provider, is it observable by testing, does anyone outside the labs actually know it, and who profits if you believe it. Most claims collapse or confirm within an hour of testing.
Why do these myths persist?
Usually because they are more interesting than the true answers, easier to sell, or attached to a product. The llms.txt claim follows the clearest arc: real proposal, confident write-up, widespread quotation, then tools built on it.
Which myth is most expensive to believe?
Publishing volume. It misallocates budget every month, and it is usually paired with rank tracking, which hides the misallocation from you.
Is AI search traffic too small to matter?
It is around one per cent of web traffic in the largest published study with a stated denominator. Small, real and growing — and the highest-impact fixes cost hours rather than months, which is what makes it worth doing now.
Is it too early to invest in AEO?
Half true and it depends on your category. For local trades and walk-in retail, listings and reviews still return more. For question-led B2B, professional services and healthcare, the buyers are already there.
Do AI assistants prefer recent content?
Freshness appears to help on time-sensitive queries and matters little on stable ones. We have not measured this ourselves and would not state it as a rule.
Does site speed affect AI citation?
Indirectly. It affects crawling and indexation, which affect eligibility. It is not a selection factor in the way structure and corroboration are.
Can anyone tell me why a model cited my competitor?
No. No provider exposes model reasoning, so every explanation offered is inference. What can be established is whether you were eligible to be chosen at all.
Should I believe AEO case studies?
Ask whether a dated baseline was recorded before the work started and whether verbatim answers were stored. Without both, the result cannot be distinguished from ordinary run-to-run variance.
What should I do first, given all this?
Read your robots.txt for the five AI crawlers and fetch a commercial page with JavaScript disabled. An hour of work, it settles more than any article will, and it is the one thing on this page nobody disputes.

Want this done for your site?We build and maintain the search, content and paid programmes described on this page.

Get a free proposal

Get a free marketing proposal

Tell us what you are trying to grow and we will come back with a plan, not a pitch deck. Same-day reply on weekdays.

Our website uses cookies from third party services to improve your browsing experience. Read more about this and how you can control cookies by clicking "Privacy Preferences".
Privacy Preferences
When you visit our website, it may store information through your browser from specific services, usually in form of cookies. Here you can change your privacy preferences. Please note that blocking some types of cookies may impact your experience on our website and the services we offer.
Privacy Policy
You have read and agreed to our privacy policy
Required
Contact Us