AI Visibility
AI Citation Snapshot
Enter your domain and we generate a set of buyer prompts, run them across the AI engines, and hand back a snapshot of where you are cited right now and where you are missing.
What a citation rate is, and what one number is worth
The snapshot builds a buyer prompt set from what you sell, spreads it across intents so it is not four versions of the same question, and asks each connected engine every prompt. An engine's rate is the share of its answers that named your domain. The headline percentage is the average of those per engine rates, so an engine that failed to answer does not drag the number down.
A first reading is a baseline, not a grade. There is no industry average worth comparing yourself against, because the rate depends on how contested your category is and how specific your prompts are. The value of the number arrives on the second run, when the prompt set is unchanged and the difference means something.
The per-prompt list underneath is the part that turns into work. A rate tells you the size of the problem. A prompt where every engine names someone else tells you which page to write.
What your rate is telling you
| Rate | The usual cause | The first move |
|---|---|---|
| 0 percent | Either no engine can retrieve you, or nothing on the site answers a buyer question in a form worth quoting. | Confirm crawler access and indexing before anything else. A blocked crawler makes every content fix pointless. |
| 1 to 25 percent | You win narrow prompts, the ones close to your brand or a niche use case, and lose every shortlist question. | Work the comparison and evaluation prompts. Those are winnable with pages you control, unlike the best X question. |
| 25 to 60 percent | You are inside the consideration set. The average has stopped being the interesting number. | Fix by prompt rather than by site. Read the answer for each miss and match the source it leaned on. |
| Over 60 percent | Usually a narrow category, or a prompt set weighted toward your own brand. | Widen the prompt set until it starts to hurt, then keep the harder set as the baseline you measure against. |
Compare a snapshot only against itself. A rate measured on five prompts today and twenty prompts next month is two unrelated numbers, so freeze the prompt set before you start reading the trend.
Citation snapshot questions
Why did the same domain score differently an hour later?
Engines search live and re-rank sources on each request, so borderline answers flip. On a small prompt set one flip is a large percentage move. Treat any change smaller than a full prompt as noise, and re-run on a schedule if you want a trend you can trust.
Does the snapshot check my whole site or one page?
The whole domain. Any URL on your domain appearing in an answer counts as a citation. Which page earned it is worth knowing, because it is frequently a blog post you forgot about rather than the product page you would have picked.
Which engines are included?
The ones with an API key on this deployment, from ChatGPT, Gemini, Perplexity, Google AI Overviews, and Google AI Mode. Any engine without a key is marked as not connected and left out of the rate entirely, rather than counted as a miss.
Is a low rate a content problem or an authority problem?
The answers tell you which. If the engines cite roundups and review sites while you are absent, the gap is authority: you need to be named in the sources they already trust. If they cite a competitor's own product page, the gap is content, and you can close it by writing a better version of that page.
Previews the Reporting Agent
This is today. Rankfastt turns the snapshot into a tracked baseline and shows the real change in your clicks and citations after every fix.