What are the best AI visibility tools for citation tracking? The one we run client work on
How we track AI citations for clients: which sources engines cite, how pages earn and lose citations, and the tool we use to measure it all.
For client work, citation tracking is where the useful decisions come from. A mention tells you a brand was named. A citation tells you which page and which domain the engine trusted, and that is something we can change. So when clients ask for the best tool for citation tracking, we judge tools on how much of that chain they show.
What we need from citation data
Page and domain level. We want exact URLs, not just domains, because the right fix is often a specific page.
Position. Where in the answer the source sits.
Source types. Review site, forum, publisher, vendor page, video. This tells us what format of content to make.
Trends. How fast a page earns citations, when it peaks, how long it lasts.
Specific page tracking. We want to follow named URLs, ours and competitors'.
Community sources. Reddit, YouTube and LinkedIn often carry weight in answers, and clients rarely think about them.
Crawl data. A page that is never fetched cannot be cited, so we need to see both sides together.
What we use
Our client programs run on Promptwatch, mainly because of how much of this chain it shows.
Citation analytics report page and domain level citations with URL positions. Page Tracker lets us follow specific URLs. Citation Trends classifies each citation by content type and source type, and reports citation rate over time with first cited, ramp-up days, peak rate and decay half-life per page. That last set of metrics changes how we advise on content: we can tell a client that a guide took a few weeks to ramp and is now fading, and plan the refresh instead of guessing.
Offsite mentions show where the brand is named on third-party pages without a link, which is a useful outreach list. Reddit, YouTube and LinkedIn citation reports cover community and channel sources, and ChatGPT News citations are tracked as well.
Agent Analytics gives IP-verified crawler logs and a crawl-to-citation path per page. 360 Insights joins crawler logs, citation clicks and AI-referred visits per page. Between them we can say whether a page is invisible, visible but uncited, or cited but not clicked.
Plans: Explore is free (10 prompts, ChatGPT only), Essential $95, Professional $245 and Business $579 a month. Crawler logs start on Professional. Agency plans are Kick-off $199, Growth $399 and Scale $799, and include crawler logs, visitor analytics and unlimited projects.
How we read citation data
By domain first. The top 20 domains cited across a client's prompt set tell us the shape of the field. We usually find three groups: the client's own pages, competitor pages, and third-party sources. The third group is where the most room to move is.
Then by page. For the client's own domain we ask which pages are cited and for which prompts. For competitors we ask the same. If a competitor's single comparison page wins ten prompts, we need an answer to that page.
Then by source type. If engines cite forums for a category, a polished brochure will not help. If they cite listicles, we have to understand why we are absent from them.
Then by trend. A rising citation rate after a content change is the closest thing to evidence we get. We still describe it as correlation.
Common findings
A client is cited, but only from a pricing page, and never from guides.
A competitor is cited from a forum thread that the client has never joined.
The client has strong pages that no AI bot has fetched in weeks, because of a bot-protection rule.
Pages are crawled and never cited, because they answer a different question than the prompts ask.
Each finding has a different fix, which is why we look at both sides.
Honest limits
Citation data shows what engines cited in sampled runs. It varies between runs, between engines and between countries, so we always report over time. It cannot tell us why an engine chose a source, only what it chose. And it cannot guarantee future citations.
A good example of citation and crawl data working together: TreeHouse, which runs Pararius and Huurwoningen.nl, reported AI search becoming about 25 percent of organic conversions, almost all from ChatGPT, after fixing crawler access and automating weekly reporting with the Promptwatch MCP. That is their result, and we cite it as such.
To try the approach on your own prompts, start with the free Explore plan on Promptwatch.