Why 90% of brands are still invisible in AI search

AI search is still wide open. A study sponsored by Victorious found that only 18 out of 177 brands earned any AI mention rate above zero across 107,011 responses from eight major AI platforms in the first quarter of 2026. That means 89.8% of the brands tested were effectively invisible in AI-generated answers.

For SEO and GEO teams, that is the real headline. The gap is not just about ranking in traditional search. It is about whether large language models mention your brand at all, whether they cite your site, and whether those two things happen together. This study suggests they often do not.

What did the study actually measure?

It measured the distance between classic organic performance and AI answer visibility by looking at the same brands across both environments.

The dataset covered 177 brands in five verticals: healthcare, SaaS, financial services, ecommerce and retail, and legal services. Researchers tested those brands across eight platforms, including ChatGPT, Gemini, Perplexity, Claude, Copilot, Meta AI, Google AI Overview, and Google AI Mode.

Mention rate is how often a brand was named in an AI answer. Citation rate is how often the model linked to the brand's domain as a source. That distinction matters because a brand can be visible in one way and absent in the other.

• The study first selected a cross-section of brands by vertical.
• It then collected 107,011 AI responses using vertical-specific prompts.
• For each response, it logged whether the brand was mentioned and whether the brand's website was cited.
• It matched that data with Q1 2026 Semrush domain metrics, including traffic trends and Authority Score.

A simple example makes the difference clear. An ecommerce brand might be named in an answer because the model recognizes it, while the sources behind that answer come from Amazon, Reddit, or review sites instead of the brand's own domain. In that case, the brand has mentions without owning the evidence layer.

Why are most brands still absent from AI answers?

Because AI systems do not reward brand size alone. They appear to reward brand clarity, outside validation, and sourceable content.

The strongest performers in the study had signals that models could attach to a brand identity. In healthcare, that often meant clear entity markers such as names, locations, specialties, and network affiliations. In SaaS, it meant discussion on platforms like G2, Reddit, and LinkedIn. In financial services, it meant strong editorial coverage on sites such as MarketWatch, Bankrate, and NerdWallet.

That helps explain why so many brands are missing. Many companies still treat AI visibility as a downstream effect of SEO. The study points in a different direction: visibility in answer engines depends on whether models can confidently connect your brand to trusted pages, trusted entities, and trusted third-party references.

There is also an important nuance here. This was a sponsored study with a sample of 177 brands across five sectors, so it should be read as directional evidence, not as a universal rulebook. Even so, the pattern is hard to ignore. Most brands are not competing aggressively in AI search yet because most brands are barely present there at all.

What do the vertical patterns tell us?

They tell us there is no single AI visibility playbook. The way brands appear in AI answers changes by sector, which means GEO strategy has to start with the shape of the market, not with generic advice.

• Healthcare, SaaS, and financial services: these categories were both mentioned and cited. The common thread was stronger entity definition, stronger editorial presence, or stronger third-party validation.
• Ecommerce and retail: these brands were mentioned more than they were cited. Models recognized them, but often relied on marketplaces, aggregators, and review sites for supporting evidence.
• Legal services: this category showed the opposite pattern. Legal sites were cited regularly, but the firms behind the content were rarely named, which suggests the content earned trust while the brand entity stayed weak.

That last point is especially useful. If you are a law firm, your content may already be doing part of the job, but your brand may not be receiving the credit. If you are an ecommerce brand, the problem may be the reverse: people know the name, but your own site is not the source an AI system wants to rely on.

The study also found that different AI platforms prefer different source sets. That means a brand can look stronger in one model and nearly absent in another, even when the prompt is similar. For marketing teams, cross-model tracking is no longer optional. It is the work.

BotRank's Take

The most useful part of this study is not the 89.8% figure. It is the fact that mention rate and citation rate were tracked separately. Too many teams still talk about AI visibility as if it were one number. In practice, a brand can be named often and cited rarely, or cited often while staying invisible as a brand. Those are different problems, and they need different fixes.

This is why BotRank's AI Visibility feature matters in this context. It lets teams run reusable prompts across multiple LLMs, compare how brands and competitors appear over time, and inspect the actual sources behind those answers. It also shows whether cited pages truly mention the brand, which is a small detail with big consequences. If a model keeps pulling supporting pages that never reinforce your name, your visibility problem is not solved just because a URL appeared. That kind of separation turns AI search from a vague concept into something measurable and actionable.

What should brands do next?

Start by separating the problem into three layers: mentions, citations, and model coverage. If you lump them together, you will miss the reason visibility is weak.

• Audit entity clarity on your site. Make sure your brand, products, locations, specialties, and authorship signals are easy to understand and consistently connected.
• Build pages worth citing. Ecommerce brands, in particular, need content on their own domains that can compete with marketplaces, forums, and review sites as a source.
• Earn third-party reinforcement. SaaS and financial services brands benefited from discussion and editorial coverage outside their own properties.
• Track by model, not just by keyword. A brand can perform well in Perplexity and poorly in Gemini, or vice versa, because source preferences differ.
• Treat brand mention and source citation as separate KPIs. Legal brands are a good example of why. Content can earn trust before the brand earns recognition.

The opportunity is obvious. If only 18 brands in the sample earned any non-zero mention rate, then most categories still have open space. The first teams that build strong entity signals, publish citation-worthy pages, and measure results across models will have an easier path than the teams waiting for AI visibility to happen on its own.

AI search visibility is not a future problem anymore, but it is still an early one. If you want to know where your brand is actually showing up, which competitors get named, and which pages models trust enough to cite, BotRank gives you a way to measure that before the window gets crowded.

FAQ

What is the difference between an AI mention and an AI citation?

An AI mention means the model names your brand in its answer. An AI citation means the model links to your domain as a supporting source. You can have one without the other.

Does strong SEO automatically lead to AI search visibility?

No. The study compared organic performance with AI visibility and showed that traditional strength does not guarantee mentions or citations in AI answers. Strong SEO helps, but it is not the whole system.

Which sectors looked strongest in the study?

Healthcare, SaaS, and financial services showed the strongest pattern of being both mentioned and cited. Ecommerce and retail were mentioned more than cited, while legal services were cited more than mentioned.

Why should brands track multiple AI platforms instead of just one?

Because different platforms appear to rely on different source sets and visibility patterns. A brand can look established in one model and nearly absent in another. That is why cross-model measurement matters.