Services/AI Search (AEO/GEO)
Search returns a list. An assistant returns a name.
Asked to recommend a business like yours, ChatGPT names two or three. There is no page two, no also-rans, and no notification when you were left out. The shortlist is built before your customer sees anything.
Most businesses have never checked whether they appear. That is the first thing we fix.
Why This Is Different
Being on page one is no longer the same as being considered
A results page gave everyone a chance. Being seventh still meant being seen, and a determined buyer scrolled. An answer engine does not work that way. It returns a short, confident response naming a handful of options, and everything else is simply absent.
That changes what visibility work has to achieve. Ranking is still necessary, because assistants retrieve from pages that rank. But it is no longer sufficient, because the assistant then has to decide you are worth naming, which depends on whether it can describe you accurately and confidently.
The businesses being named are not always the biggest. They are the ones whose details agree across every source, whose pages answer questions completely, and who have enough third-party corroboration that a machine will risk recommending them. That is winnable, and most of your competitors are not yet trying.
What The Work Is
Measure, then make yourself quotable
Find out where you actually stand
Almost nobody has checked. We ask the assistants the questions your customers ask, record who gets named, and establish which competitors are being recommended in your place. That baseline is the whole starting point, and it takes minutes rather than months.
- Prompt testing across ChatGPT, Gemini, Perplexity and AI Overviews
- Which competitors appear when you do not, and what they have that you lack
- What the assistants currently say about you, including anything wrong
- Which of your pages are being retrieved, and which are invisible
Make the machine confident about you
An assistant names businesses it can describe without risk. That confidence is built from structured data, consistent details across every source, and pages that state things plainly enough to be quoted. Ambiguity is the enemy: where a machine is unsure, it picks someone else.
- Schema and structured data, so facts are stated rather than inferred
- Entity consistency across Google Business Profile, Bing Places, Apple Business Connect and the directories in your category
- Pages that answer a specific question completely, in language customers use
- The specifics competitors leave vague: prices or ranges, service areas by name, what you do not do
Earn the citation
Assistants retrieve from the web, and they overwhelmingly retrieve pages that already rank. That is why this is not a separate discipline bolted onto SEO. It is the same underlying work, aimed at a surface that summarises rather than lists.
- Content built to be cited, containing something not already in the model
- Third-party corroboration: coverage, reviews, and profiles that agree with each other
- Monitoring across assistants, since answers move and nobody notifies you
- Reporting tied to enquiries and branded search, not to impression counts
Related
SEO
The foundation assistants retrieve from.
Local SEO & Google Business Profile
Where local AI recommendations get their data.
We write about this regularly in AI Search, and the method sits in Visibility Engineering.
FAQ
Common questions
Less than the acronyms suggest. SEO gets you ranked in a list of results. AEO and GEO are about being named inside an answer that an assistant generates. The work overlaps heavily, because assistants retrieve from the pages that already rank. Treat anyone selling them as entirely separate disciplines with suspicion.
No. Nobody controls what a model outputs, and any agency claiming otherwise is overselling. What can be done is make you the most retrievable, most consistent and most quotable option in your category, which is what determines who gets named. We report on whether that is working.
This is the honest hard part. Assistants rarely pass a referrer, and Google's own AI reporting in Search Console currently shows impressions but not clicks. We combine manual citation testing, branded search volume, direct traffic and a source question on your enquiry form. It is a triangulation rather than clean attribution, and we say so.
Almost certainly not, unless your product is the content itself. A publisher losing page views to a summary has a real grievance. A service business does not sell page views; it wants to be the name that comes up. Blocking is choosing not to appear in the answer. One caution: nosnippet removes you from AI Overviews and also kills your ordinary search snippet, all or nothing.
It is early, which is the argument for doing it. The work compounds and most of your competitors have not started. It is also not speculative: the same foundations that make you visible to assistants make you visible in ordinary search, so the downside case is that you improved your SEO.
Does ChatGPT name you, or a competitor?
Find out before your next customer does. We run your details through the assistants your buyers use and send you what they say.