Evidence
What AI citations actually tell you about your content gaps
Citations are not vanity metrics. They are the most direct read on which of your sources an agent trusts - and which it ignores.
Every time an AI agent recommends a brand, it leans on a small set of sources. Some are owned (your docs, your blog), some are earned (G2, Reddit, news), and some are nothing to do with you. The mix tells you exactly where your evidence is thin.
Three citation patterns to look for
1. The agent cites only your homepage
This is a yellow flag. It usually means the agent could not find a more specific page that answers the buyer's question. Add a comparison page, a use-case page, or a deeper guide on the topic.
2. The agent cites only third parties
Your content is missing entirely from the synthesis set. The buyer reads about you through someone else's framing. Write the canonical version yourself, or your competitors will.
3. The agent cites a competitor's comparison page
Painful but common. If a competitor wrote "X vs you" and you didn't write "you vs X," they own the narrative. Publish your version with honest tradeoffs and stronger proof.
What a useful citation report contains
- Source URL and domain
- Whether it is owned, earned, or unrelated
- Which prompt and which agent surfaced it
- Whether the citation supports or contradicts the recommendation
From citation gap to content brief
A citation gap turns into a one-line brief: "Agent X recommends competitor Y for prompt Z because the only source it has on us is our homepage." That is enough to write a real comparison page or a focused doc.