Is this AI SEO? Can you get us recommended by ChatGPT?
No, and we will not claim otherwise. Citerva measures what AI assistants say about you: how often you are named, whether you are cited, and who is named instead. Nobody can guarantee that a content change will make a model recommend a brand, because model updates, retraining and even prompt phrasing move these numbers on their own. What we give you is a measurement you can act on and then re-check.
AI answers change every time. How is that a measurement?
Because it is a rate, not a ranking. We ask a defined set of prompts, count the share of answers that named you, and always show the counts behind the percentage. One scan is a single sample; the trend across repeated scans is the signal. Small samples are flagged, and a one-run difference is labelled as sampling noise rather than a trend.
Which AI assistants do you check?
OpenAI, Anthropic, Google Gemini and Perplexity. Only providers with credentials configured are ever scanned. If a provider was not asked, or was asked and failed, that run is excluded from every rate and shown as a gap. We never report a provider as checked when it was not.
What is the difference between a mention and a citation?
A mention is an assistant naming your brand in its answer. A citation is it linking to a page you own as a source. They move independently: being described from third-party reviews without ever being linked to is common, and it is a different problem from not being named at all.
Why is citation rate blank for some providers?
Perplexity and Gemini return the sources they used. OpenAI and Anthropic answer from the model unless web search is enabled, so they return none. Counting them as having cited you zero times would understate your attribution for reasons that have nothing to do with your content, so they sit outside that denominator and the exclusion is shown on screen.
Why are questions that name my brand treated differently?
Asking whether a named product is any good will nearly always produce an answer mentioning it. Including those would push mention rate towards 100% while measuring nothing about discovery, so they are excluded from the headline rates and reported separately. They are still worth asking, because they show what a model says and cites once it already knows who you are.
Do you write the prompts, or do I?
We generate a set from your product description, competitors and topics, and you review, edit, approve or reject every one before anything is sent. Generation is free; asking an AI provider costs money, so nothing runs until you approve it.
Can I see the actual answers?
Yes. Every answer is stored exactly as the provider returned it and shown against the prompt that produced it, with each mention and citation traceable to the text it was read from. If you doubt a number, you can read its source.
Does it work for agencies with several clients?
Yes. A workspace holds many projects, so you can run one project per client, or give each client its own workspace when billing and access need to be separate.
Can I change or cancel my plan?
At any time, from your billing settings. Plans are sized by provider calls per month rather than by feature checkboxes, because provider calls are what a scan actually spends.