AI search field guide
AI citation gap analysis: a repeatable evidence worksheet
A citation gap is a question where a relevant source appears in a saved AI answer while your own useful page does not. Investigate the page and the question before choosing a fix.
Build the worksheet
Create columns for prompt ID, exact prompt, market, language, platform, collection method, capture date, answer ID, competitor source URL, your target URL and proposed change. Keep one row per question-source pair. A domain appearing often does not establish that every page on it is an opportunity. Save the specific URL and the passage relevant to the answer. Use the source-analysis method to distinguish domain usage from individual page evidence.
Work through a concrete example
Illustrative example: an answer to “Which scheduling tool supports round-robin assignment?” cites a rival’s implementation guide. Your pricing page names scheduling but never explains assignment rules. The gap is not simply a missing keyword. Check whether your product supports round robin, then document configuration, plan restrictions and an example if it does. If it does not, another topic is a better investment.
Separate access problems from explanation gaps
Check response status, indexability, canonical URL and the visible explanation on your target page. A blocked page needs a technical owner; an unclear capability needs a product reviewer. A citation to an independent review may call for honest outreach or better supporting evidence, not a duplicate review published on your own domain. Prioritize fixes you can verify.
Define a follow-up that can fail
Save the edit date and a frozen prompt cohort. Repeat the collection under the same method and compare eligible answer counts and citation completeness. A new citation is evidence of a changed response, not proof the edit caused it. If the page still fails its reader-facing acceptance test, improve that problem before increasing the sample.
Sources and further reading
- Google Search Central: AI features and your website
- Google: creating helpful, reliable, people-first content
Sources explain the underlying standards. They do not endorse InfiniteRanks or guarantee a ranking outcome.