When a buyer asks ChatGPT, Claude or Perplexity to recommend a company in your category, the assistant doesn't read your homepage. It reads a handful of third-party pages — comparison articles, directories, review sites — and answers from those. We measure which ones, and whether they name you.
We ran real buying-intent questions and collected every page that surfaced. For "best transactional email API for developers", every single result was a third-party listicle. Not one vendor's own homepage appeared.
So your standing in AI answers is largely decided on pages you don't own and didn't write. Most companies have never looked at that surface, because until recently there was no reason to.
We publish the underlying measurement for whole categories, free and without signup, so a company can look up its own number without paying anyone: citation research by category — 107 vendors measured across 8 software categories.
If your category is there, read your number, take the source list, and fix it yourself. That's a completely reasonable outcome.
We choose the questions a buyer in your category would actually ask an assistant, collect the third-party pages that surface for them, and fetch each one as a well-behaved crawler. On every page we record each vendor named — by outbound link and by name in the prose — applying the identical test to you and to every competitor.
Every finding links to the page it came from. You can check any line of the report yourself in about a minute, and we'd rather you did.
Nobody controls whether an assistant names a given business on a given day. The systems are opaque, they change without notice, and they vary by phrasing and by user. Anyone selling you a guaranteed placement is selling something they cannot deliver.
What we give you instead is the set of conditions that are within your control, and which of them you currently fail. Being present on the pages assistants read is necessary. It is not sufficient, and we won't pretend otherwise.
An SEO audit examines your own website. This examines everyone else's. The gap it finds isn't "your meta descriptions are weak" — it's "these eleven pages decide your category, your competitors are on nine of them, and you're on two." Different surface, different work, different fix.
No, and nobody can. Assistants are opaque, change without notice, and answer differently depending on phrasing and user. What we can show you is which sources those answers are built from and exactly where you're absent. Anyone promising you a placement is promising something they don't control.
A document, by email, within 3 business days. It contains your citation share against every competitor found, each source with the quote where you appear, the ranked list of pages naming competitors but not you, a technical read on your own site, and an ordered action plan with effort estimates.
Yes, and we'd prefer you did. Every finding links to the public page it came from. Nothing in the report rests on data you can't open yourself.
Ask for a refund within 14 days and you get one, in full, without having to explain yourself.
Full refund on request within 14 days of delivery. No questions, no conditions. Email [email protected] and it's returned to your original payment method, typically within 5–10 business days.
If we haven't delivered your report within 5 business days of payment, you get an automatic full refund without having to ask.
[email protected] — a real person reads this, and replies within one business day.
Citation Footprint · research reports on AI answer sources · [email protected]