AI Search Visibility (GEO / AEO)
Getting your brand named and cited in AI-generated answers: entity consistency, structured data, and the content formats assistants actually quote.
A growing share of commercial research now happens inside ChatGPT, Perplexity and Google’s AI Overviews, where there is no blue link to rank for — only sources the model chooses to cite. Getting cited is a different discipline from ranking, and it has measurable drivers: entity consistency across the web, structured data, content formats built for extraction, third-party presence, and freshness. This service builds all five.
## How the three systems actually choose sources
They do not work the same way, and a strategy that treats them as one thing under-performs on all three.
What each system favours
| System | How it retrieves | What it favours |
|---|---|---|
| ChatGPT | Training data plus selective Bing retrieval — web search triggers far more often on commercial queries than informational ones | Encyclopedic, comprehensive pages; FAQ schema and inline citations; presence on third-party review platforms |
| Google AI Overviews | The organic index, but increasingly decoupled from top-10 position | Semantically complete, self-contained answers; multi-modal pages; verified data; regular updates |
| Perplexity | Real-time search on every query, no knowledge cutoff | Freshness above almost everything; question-shaped headings; visible statistics and named sources; community platforms |
Behaviour summarised from Leapd, 2026 and Omnibound’s 2026 citation dataset.
## The five levers we pull
- Weeks 1–4
Entity resolution
Make the web agree on who you are. Identical name, address and description across your site, LinkedIn, Crunchbase, review platforms and industry directories, bound together with
sameAsin Organization schema. An assistant that cannot resolve you as a single confident entity will not name you. - Weeks 2–6
Structured data as a graph
Not scattered blocks — one connected
@graphper page with stable IDs, so Organization, Service, WebPage, Article and FAQ nodes cross-reference each other. Pages carrying FAQ schema and inline citations attract materially higher citation weighting. - Ongoing
Citable content formats
Self-contained answers in the opening section, question-shaped headings, comparison tables, original statistics with named sources. Comparison content alone captures around a third of all AI citations.
- Ongoing
Third-party presence
The overwhelming majority of AI citations come from earned media rather than a brand’s own site. Review platforms, industry directories, journalism and community discussion are where that presence gets built.
- Quarterly
Freshness discipline
Recently updated pages are cited several times more often than stale ones, and the effect is strongest in Perplexity. Every page we manage carries a visible review date and sits in a refresh queue.
## What we measure
Monthly prompt testing across ChatGPT, Perplexity, Google AI Overviews and Claude: a fixed set of buyer-intent prompts for your category, run repeatedly, tracking whether you are named, how you are described, and who is named instead of you. That last one is the useful part — it tells you which competitor’s content the models trust.
Nobody can guarantee an AI citation, and any agency that does is selling you something. These systems are opaque, they change without notice, and the same prompt can return different sources on different days. What is measurable is the direction: whether you are named more often this quarter than last, across a consistent prompt set. That is the metric we report against.
Questions buyers ask us
These are the questions that come up most often on first calls. If yours is not here, ask — we answer in writing within one working day.
Is AI search optimisation different from SEO?
It overlaps heavily but is not identical. Ranking is about a position in a list; citation is about being a source a model trusts enough to quote. Strong SEO helps — AI Overviews draw from the organic index — but the specific levers differ: entity consistency, extractable structure, third-party presence and freshness carry unusual weight.
Can you guarantee we will be cited by ChatGPT?
No, and neither can anyone else. What we can do is build the measurable drivers and report on a consistent prompt set month over month so you can see whether it is working. Anyone offering a guarantee here is either misunderstanding the systems or misrepresenting them.
Does schema markup actually help with AI visibility?
The evidence points that way, particularly FAQ markup and a well-connected entity graph. The stronger mechanism is probably indirect: schema makes your content unambiguous, and unambiguous content is easier for a retrieval system to use confidently.
Should we block AI crawlers to protect our content?
That is a real strategic decision with a genuine trade-off, and it depends on your business model. For a services business trying to be found by buyers, blocking is almost always the wrong call — you are opting out of a growing discovery channel. For a publisher whose product is the content itself, the calculation is different.
How quickly does this show results?
Perplexity can pick up new content within hours of indexing. ChatGPT and AI Overviews move more slowly, usually over 8–16 weeks, because entity signals and third-party presence take time to accumulate.