ComScore Q2 2026 AI Intelligence Report separates two layers of an AI answer: the source links a response draws on, and the cited links it shows to the user. The report found that brands appear as source links far more often than citations. For example, Tripadvisor appears as a source in 61% of AI responses related to lodging, but is only cited in 21% of lodging-related responses. 

Frequency is just part of the equation  

The report shows how often a specific domain is used as a source and cited by AI, but doesn’t expose how much of an AI answer is shaped by that domain. This is a problem because two pages that are cited an equal number of times in AI answers can have very different levels of influence over those answers.  

Bluefish measures that difference in impact directly. For every source page an AI response cites, Bluefish calculates an Impact Score between 0 and 100. The Impact Score measures how strongly a page's content shaped the answer, based on the semantic similarity between the page and the generated response. Bluefish combines Impact Score with Source and Citation Frequency to calculate Influence Rank, which ranks pages and domains by their overall sway on AI answers. 

Read together, Impact Score and Frequency show marketing teams which pages to change and what kind of change they need. A page with high Frequency and low Impact is a prime candidate for content changes because AI systems can easily find the page, but its content is not shaping the answer. On the other hand, a page with high Impact and low Frequency has content that sways AI, but may need technical updates to be considered by agents. 

In our Black Friday 2025 report, for example, we found that while tech publications like CNET and TechRadar weren’t cited very frequently, they were hugely impactful in shaping how AI systems described brands over the Black Friday period. If a brand only considered the number of times they appeared in AI answers, they would likely focus their efforts on high-visibility sites like Reddit or Best Buy, unaware that the tech publications actually have a greater impact on answers.  

What this means for AI measurement 

ComScore’s report concludes that influence matters more than clicks. We agree, but counting how often a site appears as a source or a citation doesn’t answer the bigger question for brands: What kind of content does AI reward in a given product category or topic when making recommendations for a specific audience?  

Frequency tells a brand where they show up in AI. Impact tells a brand where showing up changes the AI’s answer. Marketers that only track Frequency end up spreading effort across every site that happens to be cited, while teams that track both can focus their budget on the pages and publishers that shape what AI says and drive a higher return. 

Request a demo to see how much influence your brand's content has over AI answers.