Skip to main content
Sources are the foundation of AI optimization. While you can’t directly influence AI responses or training data, you can influence the sources AI models reference. Navigate to Sources in your sidebar to analyze which websites and URLs drive your visibility.

Why sources matter

Optimizing sources gives you the most control over your AI visibility. Sources are all URLs AI models access during response generation, while citations are sources explicitly referenced in the response text. Both contribute to AI responses. You have full control over your own website content and can influence external sources through PR, partnerships, and community engagement. When AI models reference sources where you have a stronger presence, your visibility increases.
Important limitation: AI models only see HTML content. They can’t read behind paywalls or load JavaScript-dependent content.

Source Domains

In the Sources page, you’ll see:
  • Source Retrieval by Domain graph: Shows trends over time for the top 5 domains. Hover over the chart to see which domain each line represents and track how source retrieval changes over your selected timeframe.
  • Sources Type chart: Displays the number of citations by domain category (Editorial, Corporate, UGC, etc.). This helps you understand which types of sources dominate your industry and where to focus your optimization efforts.
  • Domain Table: It displays all domains that have contributed to generating answers for your prompts.

Domain Table

The domain table shows:
  • Source: The domain name of the website.
  • Domain Type: Shows “You” for your domains, “Competitor” for tracked competitor domains, or automatic classification (Editorial, Corporate, UGC, Reference, Institutional, Other).
    • Click on All Domain Types to narrow your analysis to any of the domain types. You can select just one or multiple.
  • Retrieved: Percentage of chats where a domain appeared as a source.
  • Retrieval Rate: Average of how many times a domain’s URLs were retrieved per chat.
  • Citation Rate: Average number of times the domain was explicitly cited when used (in your selected time period).
  • Gap Analysis toggle: Content gaps and opportunities based on the Gap Score column.
Gap Analysis: Higher scores indicate bigger opportunities — sources that appear frequently and mention lots of competitors represent your best targets.

Source Domain Types

Source Type provides a high-level categorization of the entire domain or website. This helps you understand the general nature of the source at a glance.
ClassDescription
CORPORATEOfficial company websites and corporate pages
EDITORIALNews sites, blogs, online magazines, and other publications
INSTITUTIONALGovernment, educational, and non-profit organization websites
UGCUser-Generated Content from social media, forums, and communities
REFERENCEEncyclopedias, documentation, and other reference materials
COMPETITORWebsites and content from direct competitors
OTHERMiscellaneous or uncategorized sources

Domain metrics

In the domain table, you will see key metrics such as Retrieved, Retrieval Rate, and Citation Rate metrics. They all give you an idea of how relevant a source is for LLMs, but differ in their influence.
Image

Retrieved

Retrieved measures the percentage of chats in which one or more URLs from a specific domain appeared in the AI’s answer as a source. This metric indicates whether your or your competitors’ content is being picked up by AI models, demonstrating your domain’s reach even when not explicitly cited. Retrieved is calculated over your selected timespan as:
Retrieved (Domains) = (Chats with at least one domain URL retrieved / Total chats) × 100
For example:
  • If AI models used any URL from your domain as a source in 40 of 100 chats, then the retrieved percentage would be 40% (40 ÷ 100 × 100 = 40%).
This means that your domain appeared as a source in 40% of all chats, regardless of how many times it was cited in the AI responses.

Retrieval Rate

The Retrieval rate measures, on average, how many times a URL from a specific domain is retrieved per chat. Values above 1.0 mean the AI is pulling multiple pages from that domain in a single conversation. This metric reveals depth of reliance; not just whether your domain shows up, but how heavily the AI leans on it. A brand with a high retrieval rate is deeply embedded in AI responses, not just occasionally present.
Note: A single chat can retrieve multiple URLs from the same domain.
Retrieval rate is calculated over your selected timespan as:
Retrieval Rate = Number of unique URLs retrieved from a domain / total chats
For example:
  • If across the same 100 chats, your domain URLs were retrieved 80 times as a source, this would yield a retrieval rate of 0.8 (80 ÷ 100).
This means that in chats where your domain was used as a source, the AI pulled an average of 0.8 different pages (URLs) from it. Indicating reliance, not just a one-off mention.

Citation Rate

Citation Rate measures the average number of times a specific source is explicitly cited within AI responses. This metric indicates how often AI models find your content relevant enough to reference multiple times. Multiple citations in a single response indicate that AI models find your content highly relevant and trustworthy for the topic being discussed. The citation rate is calculated over your selected timespan as:
Citation Rate (Domain) = Total citations of that Domain / Total responses where Domain was used as a source
For example:
  • If your domain was used as a source in 40 AI responses, and from those responses, the content from your domain was cited a total of 60 times, the citation rate would be 1.5 (60 ÷ 40)
Meaning that every time your domain is used as a source, the AI cites its content an average of 1.5 times per response, suggesting high content authority and consistency.

Source URLs

In the URLs view, you can identify specific pages that consistently influence AI responses and discover content formats that perform best in your industry:
  • Source Usage by URLs graph: Shows trends over time for the top 5 URLs, helping you track which specific pages gain or lose influence with AI models.
  • Sources Type chart: A category breakdown covering website page types, such as Article, Comparison, Listicle, Product Page, and others.
In the URL table below, you’ll find:
  • URL: The exact webpage used as a source.
  • URL Type: Automatic classification of the specific page type (Homepage, Article, Listicle, Comparison, Profile, etc.).
  • Mentions: Which brands were mentioned on this specific URL.
  • Retrievals: Total number of times this URL appeared as a source across all chats.
  • Citation Rate: Average number of times the URL was explicitly cited when used (in your selected time period).
  • Updated: The latest time we fetched content from this URL.
  • Details: Click to see what AI models can actually read from this page.
As with the Domains tab, toggle Gap Analysis to see content gaps and opportunities at the URL level. Higher scores indicate bigger opportunities — URLs that appear frequently and mention lots of competitors represent your best targets.

Source URL Types

URL Type offers a more granular classification by identifying the specific type of content on a given page. This allows for a deeper analysis of the context in which your brand is mentioned.
TypeDescription
HOMEPAGEThe main entry page of a website
CATEGORY PAGEA page that lists products, articles, or subcategories
PRODUCT PAGEA page detailing a single product or service
LISTICLEAn article structured as a list (e.g., “Top 10 Laptops of 2024”)
COMPARISONAn article or page that directly compares two or more products or services
PROFILEA directory-style entry for a company, person, or product (e.g. G2, Yelp, Crunchbase)
ALTERNATIVEAn article focused on alternatives to a specific product or service (e.g., “Best HubSpot Alternatives”)
DISCUSSIONContent from discussion forums, comment sections, or community threads
HOW TO GUIDEInstructional content with step-by-step guidance on completing a specific task
ARTICLEGeneral articles, news pieces, features, and other editorial content
OTHERAny page type that does not fit into the categories above

URL Metrics

On the URL page, you will see key metrics such as Retrievals and Citation Rate.

Retrievals

Retrievals is the total number of how many times a URL was used as a source. This metric indicates which URLs are most popular and most credible to AI models. Retrievals over your selected timespan are calculated as:
Retrievals (URLs) = Total responses where URL was used as a source

Citation Rate

Citation Rate measures the average number of times a specific source is explicitly cited within AI responses over the last 7 days. This metric indicates how often AI models find your content relevant enough to reference multiple times. Multiple citations in a single response indicate that AI models find your content highly relevant and trustworthy for the topic being discussed. The citation rate is calculated over your selected timespan as:
Citation Rate (URLs) = Total citations of that URL / Total responses where URL was used as a source

Gap Analysis

shows sources and URLs where your competitors are mentioned, but your brand is not, helping you discover content gaps and partnership opportunities you might otherwise miss. How to prioritize opportunities:
  • Start with high Gap Scores: Focus on sources that are both frequently used and mention multiple competitors.
  • Check URL details: Click Details to see what content AI models can actually read from each page.
  • Action by domain type: Editorial sites may require PR outreach, whereas UGC platforms may allow direct community engagement.

How to use Source insights

Different source types require different optimization approaches, and gap analysis helps you prioritize where to focus first. Quick strategies by source type:
  • Editorial: Focus on digital PR and journalist outreach.
  • Corporate: Explore partnerships and directory listings.
  • UGC: Engage authentically in communities or work with influencers.
  • Reference: Update incomplete information through proper channels.
  • Your own website: Optimize content structure and ensure AI readability.
For URLs, you’ll see what types of web pages get cited (Articles, Listicles, How-to Guides, Comparisons, etc.). Same as source types, you can use this information to adjust your content strategy and create formats that perform well in your industry.