Skip to main content
When ChatGPT uses web search to support an answer, it doesn’t just generate text — it pulls from specific external pages and lists them as citations. Those cited domains are the information sources the model treats as credible and relevant for that query. The Sources view tracks exactly which domains are cited in responses to your tracked prompts, how often each one appears, and which brands they’re associated with. This gives you a direct view into the third-party information ecosystem that shapes AI answers in your category.

​
What “sources” means in this context

A source is any external domain that ChatGPT cites in a response to one of your tracked prompts when the model performs a web search. Citations appear in the response as a table or inline links, referencing the pages the model consulted before generating its answer. Not all ChatGPT responses include citations. The model only performs web search for certain prompt types — typically those requiring current information, product comparisons, or recent events. When it does, the cited pages represent the live web content that directly influenced the response text. GenRank logs those citations on every prompt run that produces them.

​
Why source tracking matters

The domains that get cited are the domains that have a seat at the table when ChatGPT forms its answer. A review site that consistently appears in citations for comparison prompts in your category is shaping what the model says about those products — including yours. A media publication that covers your industry and frequently gets cited has more influence over your AI visibility than a site that writes about you but never gets cited. Understanding the citation ecosystem lets you move beyond your own domain and understand the full information environment you’re operating in. You can see:
  • Which third-party sites carry the most weight for prompts in your category
  • Whether the sites that cover you positively are actually being cited
  • Which sources your competitors are drawing citation support from
  • Where your own domain stands relative to the rest of the ecosystem

​
How your domain performs as a source

The Sources view shows your own domain’s citation metrics alongside all other cited domains. Key figures include: Your citation count: How many times your domain appears in citations across all captured responses. This is the raw frequency of your domain being treated as a credible source. Your citation rate: The percentage of responses that generated citations where your domain was included. This normalizes against the total number of cite-eligible responses rather than all responses. Prompts you’re cited on: The specific tracked prompts where your domain appears in citations. This tells you which question contexts ChatGPT currently treats your site as a relevant source for. Prompts you’re missing from: Tracked prompts that generate citations but don’t include your domain, even when your brand is mentioned in the response text. This gap — mentioned but not cited — means ChatGPT is drawing on your brand’s reputation from training data but not treating your own site as an authoritative reference.

​
Using source data to guide content and PR strategy

1

Identify the high-influence domains in your category

Sort the sources list by citation frequency across your tracked prompts. The domains at the top are the ones ChatGPT most consistently turns to when forming answers in your space. These are the publications, review sites, directories, and communities that carry disproportionate weight.
2

Evaluate your coverage on those domains

Check whether your brand is mentioned — and mentioned accurately and favorably — on the highest-citation domains. If a site appears in 30% of your tracked prompt citations but doesn’t mention your brand, or covers you only briefly, that’s a concrete PR and content gap.
3

Find the prompts where you're cited and where you're not

Compare the prompts where your domain is cited against those where it isn’t, especially for prompts where you’re already mentioned in the response text. Being mentioned but not cited means you’re present in the model’s general knowledge but not earning source authority on that topic.
4

Prioritize based on citation potential

Focus content investment on the topics and prompt clusters where the citation ecosystem is active — where other domains are regularly getting cited — but your site isn’t among them. Those are the areas where publishing strong, citable content has the clearest path to improving your citation rate.
5

Track citation changes over time

After publishing new content or earning coverage on a high-influence domain, monitor your citation rate on the relevant prompts. Because GenRank tracks citations daily, you can see when a new page starts getting picked up as a source.
Being mentioned and being cited are two different things with different leverage points. Increasing your mention rate is primarily a question of brand presence in the information environment — how well-known you are in your category. Increasing your citation rate is a question of content authority — whether the specific pages on your domain are treated as credible references for specific query types. You can improve citation rate even on prompts where your mention rate is already high by publishing content that directly addresses those prompt topics in a format ChatGPT finds citable.

​
Reading the web search queries

Alongside citations, GenRank also logs the web search queries ChatGPT executes before generating a response. These queries reveal how the model decomposes a user’s prompt into specific information-seeking steps. For example, a user prompt like “what’s the best CRM for a small consulting firm?” might cause ChatGPT to run searches like “CRM software for consulting firms 2025,” “best CRM small business comparison,” and “CRM pricing small team.” Each of those queries represents a distinct information angle the model is trying to cover — and each is an opportunity for your content to appear in those search results and subsequently get cited.
Web search queries are logged only for prompt runs where ChatGPT performed a web search. Not every prompt triggers a search. The proportion of search-enabled responses varies by prompt type and changes over time as ChatGPT’s behavior evolves.

​
Connecting sources to your broader strategy

The Sources view ties together Response Tracking and your upstream content and PR work. Mentions in AI responses come from somewhere — from training data, from live web search, from cited third-party coverage. Source tracking shows you the live-web layer of that equation: which pages are being retrieved, which domains are being trusted, and where your site fits in that ecosystem.
Not necessarily, but citation rate is a strong signal of source authority. A high citation rate means ChatGPT is treating your domain as a credible reference for those prompt topics. Brands with high citation rates tend to also have higher visibility percent scores, because cited content directly shapes response text. Citation rate and mention rate together give a more complete picture than either metric alone.
Yes. The per-prompt response detail shows the full citation table from that response run, including the specific URLs that were cited. This lets you see exactly which pages are earning citations and which topics they’re associated with.
This is one of the most valuable signals the Sources view surfaces. A frequently cited domain that mentions competitors but not your brand is a clear PR target. Getting covered on that site — especially in content that explicitly compares or recommends products in your category — is one of the most direct ways to improve your citation rate on the prompts where that domain appears.