LearnCornerstone
August 6, 2026 · 3 min read
By Marcus Bransbury · Founder, Robot Visible
AEO evidence: what can it prove?
A practical guide to AEO evidence: what crawler checks, captured answers, citations, and before-and-after comparisons can prove—and where their limits are.
Quick answers
What does AEO stand for?
Answer engine optimisation. It is the practice of making a business the answer an AI assistant gives when a buyer asks for a recommendation, as distinct from AI visibility, which is the measured outcome of that practice.
Can you guarantee an AI citation?
No, and any tool or agency that offers one is describing something it cannot control. There is no submission mechanism, and answers vary by provider, surface, time, location, and personalisation. What can be done is measure what answers say today, change what the evidence points at, and check comparably afterwards.
Start with the category definition
The AEO category definition belongs on the category page: answer engine optimisation is the practice of making a business the answer an AI assistant gives when somebody asks for a recommendation. This guide takes the narrower next question—what evidence can show whether that work is succeeding?
It is a practice, not a metric. The thing you measure is AI visibility: whether an answer mentioned you, recommended you, or cited one of your pages as a source. Those are three different outcomes and collapsing them into one number is the most common mistake in this field.
The four words people use interchangeably, and should not
| Term | What it actually refers to | Use it when |
|---|---|---|
| AEO | The practice: the work of becoming the answer. | Describing what a team does. |
| AI visibility | The outcome: whether and how you appear in answers. | Describing what you measured. |
| AI search | The channel: assistants that answer with live retrieval. | Describing where it happened. |
| GEO / LLMO / AI SEO | Common synonyms for the practice. | Meeting readers who use those words. |
AEO, GEO and AI visibility describe different layers›
How AEO differs from SEO in practice
Most SEO fundamentals still apply. A page an engine cannot fetch, parse, or trust will not be cited by anything. Google's own guidance is explicit that established search fundamentals remain the foundation rather than being replaced by a new discipline.1
Three things do change. First, the unit of competition is an answer rather than a results page, so being tenth is usually the same as being absent. Second, the winning source is often not the page you expected: a review directory, a community thread, or a competitor's comparison page can supply the sentence the assistant repeats. Third, the result is not stable in the way a ranking is. The same question can produce different answers by provider, API surface, time, location, and personalisation.
What can honestly be measured
- Whether a named crawler is permitted to fetch a specific URL, and what the initial HTML contains before JavaScript runs.
- What a given answer engine returned for a specific question, at a specific time, on a specific API surface.
- Which brands that answer named, and which role each held: mentioned, recommended, or cited.
- Which pages the answer used as sources.
- Whether a published change is publicly visible.
- Whether a later, comparable check of the same question produced a different result.
A readable site can still be absent from every answer›
What cannot
Nobody can guarantee a citation, and no tool can submit a page to an assistant for inclusion. A readiness score is not evidence that an assistant recommends you: a site can pass every technical check and still be absent from every answer, which is the single most useful thing a first measurement usually reveals.
A provider that was not checked is not a negative result, and one answer at one moment is a sample rather than a trend. Publication proves a change is live; only a comparable later check can support any claim about what moved, and even then it is a correlation across a stated interval rather than an attribution.
What an AEO practice actually looks like
The work is a loop rather than a checklist, because each stage produces the input to the next: measure what an answer engine says today, understand which brands and sources won it, decide the one change most likely to close that gap, ship it, then ask the same question again and record what happened.
Where to start
Start with a measurement rather than a plan. Ask one question your buyers actually ask, read the answer in full, and note who is in it. That single observation tells you more about your position than any audit, because it is the thing your buyer sees.
Sources and further reading
- A new resource for optimizing for AI features — Google Search Central. Acknowledges AEO and GEO as terms in common use while stating that established search fundamentals still apply rather than being replaced.
- AI Visibility Checker — Ahrefs. Uses AI visibility for the measured outcome and identifies answer engine optimisation as the underlying practice, which is the split this guide describes.
- ChatGPT Search — OpenAI. Describes query rewriting and live retrieval from search providers, which is why a measured answer is a sample at a moment rather than a stable ranking.
Continue learning
See where your website stands
Run a free scan and get your AI readiness score across all six categories, with the gaps to fix first.