Citation tracking and your Citation Score

Citation tracking asks real AI assistants (ChatGPT, Claude, and Gemini) the buyer questions your customers ask, then measures how often they actually recommend you. The result is your Citation Score, a read on how visible you are in AI answers.

Being readable by AI is step one; being recommended by it is the goal. Citation tracking (shown in the interface as AI Recommendations) runs the real questions a customer might ask through actual AI assistants and checks whether they name your business in the answer. It turns "I hope AI mentions us" into a number you can watch and improve.

How a citation scan works #

Each round of checking is a citation scan. When you run one, SuperSchema asks a set of buyer questions about your category through three AI assistants and reads each answer to see whether you were recommended, merely mentioned, or left out.

ChatGPT: OpenAI. Asked each of your questions and checked for a recommendation of your business.
Claude: Anthropic. Asked the same questions independently.
Gemini: Google. Asked the same questions independently.

By default the scan uses a set of neutral, buyer-style questions so the answers reflect how a real prospect would discover you, not a leading question about your brand. The starting questions are:

What should I look for when choosing a provider in this category?
Which companies are good options for this type of service?
What questions should I ask before deciding who to work with?
What are the most important trust signals in this category?
What is a good next step if I need help with this?

Note You can edit the business context (company name, category, audience, offerings, and competitors) and the questions before running a scan. A scan supports up to 5 questions, and each question is asked of all three assistants.

Citation scans are run on demand. You start one from the AI Recommendations card on your dashboard or from the AI Recommendations view for a domain. While it runs you will see it "Checking AI models" as it asks ChatGPT, Claude, and Gemini about your business. Results appear as each answer comes back.

How each answer is scored #

SuperSchema reads every answer and gives it one of these outcomes:

Recommended: The answer names or points to your business by name or domain. This is the strongest result and counts fully toward your score.
Mentioned: The answer references your offering or category presence without a clear recommendation of you by name. This earns partial credit.
Not mentioned: The answer does not reference your business at all.
Issue: The AI provider could not be reached or returned an error for that question, so it is not counted.
Pending: The check has not finished yet.

Your Citation Score #

The Citation Score is a 0 to 100 read on how often AI recommends you across the answers that were checked. Answers marked Recommended count fully and answers marked Mentioned count partially, so the score rewards genuine recommendations most while still crediting softer mentions. Answers with an Issue or still Pending are left out of the calculation.

On the AI Recommendations card the score sits alongside a plain-English breakdown, for example "3 recommended, 1 mentioned, 1 not mentioned", plus a letter grade and the logos of the assistants that were checked. In the full AI Recommendations view, results are laid out as a grid: one row per question with a column for ChatGPT, Claude, and Gemini, and you can expand any answer to read the full response with your business name highlighted in context and the assistant confidence noted.

Tip Look at the per-question grid, not just the headline number. It shows exactly which buyer questions you win, which competitors show up instead, and which assistant is the weakest for you, so you know where to focus.

Scan allowance per plan #

Your plan includes a set number of citation scans each month. The allowance resets at the start of each month:

Plan	Citation scans per month
Essential	2
Growth	5
Portfolio	20
Enterprise	Unlimited

Note When you reach the monthly limit you will see a "Citation scan limit" message telling you the allowance is used up and when it resets. A workspace owner or admin can upgrade the plan for more scans before then. The free trial does not include citation scans; the feature unlocks once you are on a paid plan and your first report is ready. See superschema.ai/pricing for current amounts.

Note Your Citation Score is separate from your SuperSchema Score. The SuperSchema Score measures how ready your site is for AI to read; the Citation Score measures how often AI actually recommends you. See "Understanding your scores" for how the two fit together.