The list of sources an engine cited is the most actionable artefact you have, because it names exactly which pages you would need to be on.
Citations are a map of what to influence
When an engine composes an answer about your category, it draws on specific URLs. That list is not trivia. It is a directory of the pages that currently shape what buyers are told, and therefore the concrete targets for any off-domain work.
Read them as a set, not individually. Look for the pages that recur across many prompts — those are the load-bearing sources in your category, and presence on one of them is worth more than presence on a dozen one-off mentions.
Cited does not mean the page mentions you
A subtle and important distinction: an engine can cite a page that has nothing to say about your brand. The citation tells you the engine used that page to answer the question, not that the page endorsed you.
So a page cited on a prompt you lose is a target — being covered there could change the answer. A page cited on a prompt you win, which does mention you, is an asset — worth protecting and worth understanding, because whatever it says about you is currently working.
Where this lives in SolCrys
The citations surface lists source URLs from your responses, filterable by whether the response mentioned your brand.
The mentions-you filter is the fast way to separate assets from targets.
Sources are classified by type, which the next lesson uses.
Lab: Separate assets from targets
Needs a free SolCrys account (no credit card). No quota consumed.
List the ten most frequently cited URLs across your responses.
Mark each as an asset (mentions you favourably), a target (does not mention you but shapes a prompt you lose), or neutral.
Note which single target appears across the most prompts. That is your highest-leverage off-domain opportunity.
AI engines like ChatGPT, Perplexity, Google AI Overviews, and Claude choose sources using overlapping but distinct signals. This guide maps the 7 signals that drive citation eligibility and the engine-specific weighting differences.
How ChatGPT search picks sources in 2026: crawlers, search providers, 7 citation-eligibility patterns, and what August 2026 changed about the sources it names.
A 252,000-trial study across six LLMs ranks what makes AI answer engines cite a source. Then it anonymizes away the brand trust that decides citations in production. How to read it without over-rotating.
Turn AI answer gaps into governed marketing execution.
Start free with a ChatGPT visibility read, then add multi-engine tracking, Corporate Context governance, and the action-to-result loop when you are ready.