Sources
When ChatGPT answers a buyer’s question, it pulls from pages it found on the web and writes from some of them. Those pages are sources. This view shows you which ones are shaping the story about you — the ones helping you, the ones working against you, and the ones the engine read but didn’t use.
Boosting you and working against you
The top of the view splits sources two ways. Sources Boosting You are the pages that, when the engine cites them, tend to help your case. Sources Working Against You are the ones that hurt it — a competitor’s comparison page, a review that frames you poorly, an article that leaves you out. Each source carries its net influence and a count of the questions it touches. Click through to jump to those prompts and read what the engine actually said.
Sort by Priority to see the sources that matter most first, or by Influence magnitude to rank by raw effect. If your audit covers a lot of ground, the domain filter narrows the list to one site.
Sources considered
Below the drivers sits the most actionable block in the view: Sources Considered. These are pages the engine read while answering — it pulled them in, looked at them — and then made a choice about whether to use them. Each row shows the ratio: how many times the engine cited the page out of how many times it considered it. A page that reads “0 cited / 8 considered” got the engine’s attention eight times and lost the spot every time.
Filter to Cited, Never cited, or All. The never-cited list is where your fastest wins hide. The engine already decided these pages were relevant enough to read. They just didn’t make the answer — see Concepts for why that gap matters and what closes it.
Your pages and everyone else’s
Two tables sort the sources by who owns them. Your own pages sit in one; external sources sit in the other.
