A finance brand ranks on page one for its category. Its content is current, its backlinks are strong, its domain authority is respectable. Ask ChatGPT the same question a prospect would type, though, and the brand never comes up. A competitor half its size does, by name, twice. Nothing about the SEO changed. What changed is which system is answering, and that system does not read the web the way Google’s index does.
Key Takeaways
- AI models cite brands based on trust signals and structured content, not the ranking factors that win Google search results.
- Share of Voice and Share of Citation are different metrics: a brand can dominate mentions across the web and still get zero AI citations.
- Citations compound, so brands with early AI presence pull further ahead the longer competitors stay invisible.
Your SEO Is Working. Your AEO Is Not.
Search engines and AI models solve different problems. Google ranks pages that match intent and rewards freshness, backlinks, and page experience. An AI model is not ranking pages. It is assembling an answer, and it needs to decide which sources are trustworthy enough to pull from and repeat as fact.
That distinction matters because a page can be perfectly optimized for search intent and still be useless to a model. If the content is not structured in a way the model can parse cleanly, if it is not corroborated elsewhere, or if it reads as marketing copy rather than a direct answer, the model has no reason to cite it. Rank one on Google and invisible on Perplexity are not contradictory outcomes. They are two separate scoring systems producing two separate results.
Share of Voice Is Not Share of Citation
Marketing teams often assume that if a brand is mentioned everywhere, it will show up in AI answers too. It doesn’t work that way. Share of Voice measures how often a brand is mentioned across the web. Share of Citation measures how often an AI model actually pulls that brand into an answer and names it as a source. The gap between those two numbers is where most brands lose visibility without noticing.
A brand can run a heavy PR calendar, get quoted in trade press, and still have near-zero Share of Citation if none of that content is structured for retrieval or corroborated across independent sources. Volume of mentions does not automatically earn a model’s trust. Consistency and structure do.
A brand can have high Share of Voice and near-zero Share of Citation.
Why the Engine You Ask Changes the Answer
Part of the confusion comes from treating “AI search” as one thing. It is not. Each model draws from different sources and weighs them differently, so a brand can be well covered on one engine and missing on another for structural reasons, not content quality.
- ChatGPT pulls from across the open web, which means broad coverage and independent corroboration matter more than any single high-authority mention.
- Gemini cross-references Search, YouTube, and Scholar, so video and academic or institutional presence carry weight that plain articles do not.
- Claude leans on high-authority publications and documentation, rewarding brands with technical depth and credible third-party coverage over marketing content.
- Perplexity cites its sources inline, which makes it the most transparent engine for auditing exactly why a brand did or did not get named.
- Grok reads the real-time social web, so a brand absent from current conversation on that platform is functionally invisible to it, regardless of its site content.
A brand missing from Gemini but present in Claude is not an anomaly. It’s a sign that the brand has documentation and press coverage but no video or Scholar footprint. Diagnosing invisibility means checking engine by engine, not asking “are we visible in AI” as a single yes or no question.
Two Kinds of Presence, and Most Brands Only Have Half
AEO splits into two layers that get earned differently. Training-data presence is what a model absorbed during its build, and it comes from wide, repeated coverage over time across independent sources. It is slow to earn and slow to lose. Retrieval presence is what a model pulls live when it answers a question, and it depends on structured, current, technically accessible content that the model can fetch and parse at the moment of the query.
A brand with strong training-data presence but weak retrieval setup will get mentioned in general terms but lose out on specific, current questions where a competitor’s structured page gets pulled instead. A brand with strong retrieval but no training-data history will show up for narrow technical queries and nowhere else. Most invisible brands are missing one layer entirely, and fixing the wrong one wastes months.
Trust Does Not Work Like a Score
Teams used to domain authority look for an equivalent number to chase in AEO. There isn’t one. Trust, as a model evaluates it, is a graph of agreement across independent sources rather than a single score attached to a domain. A model is effectively asking whether enough unrelated, credible sources say the same thing about a brand for that claim to be treated as fact.
This is why one glowing feature in a major outlet rarely moves the needle alone. It is also why citations compound: once a model cites a brand for one query, the next citation on a related query becomes more likely, because the model has already treated that brand as a validated node in the graph. Brands that start early build a lead that gets harder to close over time. Not because of some algorithmic advantage, but because trust, once established, keeps reinforcing itself.
growth in monthly AI mentions for Moburst’s own brand, with 42,435 total AI citations and the number 1 position in its category.See the case study
What To Check This Week
Start by asking the actual questions your prospects would ask, across ChatGPT, Gemini, Claude, Perplexity, and Grok. Record whether your brand is named, and if so, which page the model cites. This alone will tell you whether the problem is training-data absence, retrieval failure, or both.
Then look at the pages a competitor’s citations point to. If they are structured with clear headings, direct answers, and current data, and yours are written as long-form marketing narrative, that is your retrieval gap. If your brand rarely appears in independent third-party coverage outside owned channels, that is your training-data gap. It takes longer to close because it depends on earned coverage rather than a content rewrite.
Neither fix replaces existing SEO or PR work. Structured data, entity optimization, and citation building extend that work into a layer search never had to account for before.
FAQs
Why does my brand rank well on Google but not appear in ChatGPT answers?
Google ranks pages based on relevance and authority signals built for search intent. ChatGPT decides which sources to cite based on trust and how well the content can be parsed and corroborated across independent sources. Strong search rankings do not automatically transfer into AI citations.
Is Share of Voice the same as AI visibility?
No. Share of Voice measures overall mentions of a brand across the web. Share of Citation measures how often AI models actually cite that brand as a source in answers. A brand can score high on one and near zero on the other.
Do I need to check every AI engine separately?
Yes. ChatGPT, Gemini, Claude, Perplexity, and Grok draw from different sources and weigh them differently, so a brand can be well covered on one engine and absent on another for structural reasons specific to that engine.
How long does it take to start appearing in AI answers?
Retrieval presence, driven by structured and current content, can shift within weeks of technical changes. Training-data presence, built through wide corroborated coverage, takes longer because models absorb it slowly and it depends on sustained third-party mentions rather than a single content update.
