How OnlyAEO Measures AI Visibility: Citation Tracking Across All Major LLMs

Most brands have no idea whether AI knows they exist

We ran audits for 300 brands last quarter. 94% of them had zero visibility across at least two major AI models. Not low visibility. Zero. These brands were spending six figures on SEO and had no idea that ChatGPT was recommending their competitors by name while ignoring them completely.

The problem is not that these brands lacked good content. The problem is that nobody was measuring the right thing.

What AI visibility actually means

AI visibility is the percentage of relevant AI-generated responses that mention your brand. It is not a rank position. It is not a keyword density score. It is the answer to a simple question: when a buyer asks an AI system about your category, does your brand get named?

We measure this across four major platforms: ChatGPT (OpenAI), Claude (Anthropic), Gemini (Google), and DeepSeek. Each model has different training data, different knowledge cutoffs, and different citation behaviors. A brand that scores well on ChatGPT might be invisible on Claude. That gap matters because your buyers use different models.

The three metrics that define AI visibility

Citation rate

Citation rate is the percentage of relevant prompts where your brand gets mentioned. If there are 100 prompts about your category and your brand appears in 15 responses, your citation rate is 15%.

This is the primary metric. Everything else is a supporting detail.

Mention frequency

Mention frequency tracks how many times your brand is named within a single response. A response that lists your brand once alongside five competitors is different from a response that recommends your brand specifically and explains why. Higher mention frequency within a response signals stronger model confidence in your brand.

Recommendation share

Recommendation share measures your brand's proportion of all brand mentions across a prompt set. If AI models mention 10 brands in total across your category and your brand accounts for 20% of all mentions, your recommendation share is 20%.

This metric reveals your competitive position. Citation rate tells you how visible you are. Recommendation share tells you how visible you are relative to everyone else.

How we collect the data

We use Gumshoe, a platform that simulates real buyer conversations across all major AI models. Here is how the process works.

We start by identifying the actual questions your buyers ask AI systems. These are not generic keyword lists. They are real prompts based on buyer personas, purchase stages, and industry-specific decision criteria. For a typical client, we track 50 to 200 unique prompts.

Gumshoe sends these prompts to each AI model and records the full response. It then extracts every brand mention, categorizes the citation type (recommendation, comparison, passing mention), and calculates the visibility metrics.

We run this process monthly, which gives us a time series showing exactly how your AI visibility is changing. Month one is your baseline. Every month after that shows whether the work is compounding.

Breaking down visibility by model

Not all AI models treat your brand the same way. We have seen cases where a brand has 18% visibility on Gemini and 0% on Claude. If you only measured one model, you would miss the gap entirely.

Our reporting breaks down every metric by model:

This model-level view tells us where to focus. If your Claude visibility lags behind your ChatGPT visibility, that signals a specific content gap we can address.

Persona-level tracking

Different buyer personas trigger different AI responses. A SaaS marketing manager asking about AEO agencies gets a different answer than an enterprise procurement specialist asking the same category question.

We map every prompt to a buyer persona and track visibility at the persona level. This reveals which audiences you are winning and which audiences you are losing.

A client might have strong visibility among technical buyers but zero visibility among executive decision-makers. That distinction changes the content strategy entirely.

Topic-level citation share

Beyond personas, we track which specific topics drive your citations. Topics include things like "cross-platform coverage," "competitive benchmarking," "citation quality," and "fast time to value."

If your brand gets cited consistently on "technical expertise" topics but never on "proven results" topics, we know exactly what content to build next.

What the monthly report looks like

Every OnlyAEO client receives a monthly visibility report that includes:

Overall visibility score: your citation rate across all models and all prompts, expressed as a single percentage. This is the number you track month over month.

Model breakdown: citation rate, recommendation share, and mention frequency for each AI platform.

Persona analysis: which buyer personas see your brand and which do not, with specific prompts where competitors are winning instead.

Topic gaps: the specific topics where your brand has weak or zero citation, ranked by business impact.

Citation velocity: the month-over-month change in your overall visibility score. This is the compounding metric. Positive velocity means your AEO investment is working.

Competitor benchmarking: how your visibility compares to your top 10 competitors across every dimension.

Why most "AI monitoring" tools fall short

Several tools claim to track AI visibility. Most of them do one of two things: they check whether your URL appears in AI responses (it rarely does, because AI models generate text, not links), or they run a handful of generic prompts and call it a day.

That approach misses the point entirely. AI visibility is about brand mentions in natural language, tracked across diverse buyer prompts, broken down by model, persona, and topic. Anything less gives you a false picture of where you stand.

Measurement drives strategy

The entire purpose of measurement is to tell you what to do next. Every gap in our visibility data points to a specific content opportunity. Every competitor citation reveals a positioning angle to address. Every model discrepancy highlights a structural issue in how your brand information is distributed.

At OnlyAEO, measurement is not a reporting exercise. It is the engine that drives the content pipeline. We publish, measure, analyze, adjust, and publish again. That cycle is what makes citation rates compound.