What is answer engine optimisation (AEO)?

Why it emerged

For two decades, search worked one way: you typed a query, you got ten links, you clicked one. Optimisation meant competing for those positions.

That changed when assistants started answering questions directly. A person now asks ChatGPT, Perplexity or Google's AI which product to buy and receives a short recommendation — often without visiting a single website. The answer became the destination.

The consequence is measurable. Ahrefs, studying 300,000 keywords, found click-through to the top organic result fell by around 58% when an AI overview appeared above it. Pew Research found people clicked a link in roughly 8% of visits where an AI summary was shown, compared with 15% where it wasn't.

Sources: Ahrefs (300,000 keyword study); Pew Research Center.

How AEO differs from SEO

They overlap, but they are not the same discipline, and assuming they are is why many brands with good rankings are invisible in AI answers.

• The unit is different. SEO optimises pages. AEO optimises passages, because retrieval systems pull chunks of text rather than reading a page top to bottom.
• The goal is different. SEO wants a click. AEO wants a citation, which may produce no click at all — but does put the brand in front of the buyer at the moment of decision.
• The evidence is different. Search ranking leans heavily on links. AI citation leans heavily on corroboration: whether independent sources agree the brand is a credible option.
• The overlap is smaller than people assume. Only around a third of pages cited in AI answers also rank in the traditional top ten for the same query.

What the work actually involves

Question-shaped content

Content built around the exact questions buyers ask, with the answer stated in the first two or three sentences. A page that builds towards its conclusion gets skipped, because the retrievable chunk doesn't contain the answer.

Structured data

Machine-readable markup identifying the organisation, the people behind it, the services offered and the questions the site answers. This turns prose into facts a system can hold with confidence rather than inferences it has to make.

Entity consistency

The same name, description and category signals everywhere the brand appears, with explicit links between those profiles. Inconsistency splits one strong entity into several weak ones.

Third-party corroboration

Independent mentions across sources the model already trusts. This is the slowest part of the work and the one that moves the needle most, because a model treats outside agreement as far stronger evidence than a brand's own claims.

Crawler access

Explicitly permitting the AI crawlers in robots.txt. Straightforward, but frequently overlooked — and everything else is wasted if the content can't be read.

Does AEO replace SEO?

No. Traditional search still drives enormous volume, and the technical foundations of SEO — crawlability, site structure, page quality — remain the base layer that AEO is built on.

The accurate framing is that AEO is an additional surface to compete on, and currently a far less contested one. Most competitors in most categories are still optimising exclusively for links, which is precisely why the opportunity exists.

How long does AEO take to work?

Technical changes register within days to weeks. Entity consistency settles over a few weeks. Corroboration takes one to three months of genuine activity.

Specific, long-tail buying questions move first, because they're less contested. Broad category terms take considerably longer.