What are AI citations? How PR teams can track and earn them

When a buyer asks ChatGPT to recommend a vendor in your category, your brand either appears in the answer or it doesn't. Most PR teams have little visibility into why some brands are cited while others aren't. Muck Rack's analysis of more than 25 million AI-cited links found that earned media accounts for about 84% of AI citations, while paid and advertorial content accounts for just 0.3%. Earned media plays a major role in shaping AI-generated answers, but measuring that impact requires a different scorecard than press clips or web traffic. This guide explains what AI citations are, where they come from and how PR teams can track and earn more of them.

What is an AI citation?

An AI citation is a source URL or reference an AI engine surfaces when constructing an answer. Unlike a search backlink, its purpose is attribution within a synthesized response. The reader may see a footnote, a source panel or an inline reference depending on the platform. Regardless of the format, the model is signaling its information source.

As Nieman Lab reported in its analysis of the same study, journalistic content serves as a core component of generative AI answers rather than a peripheral input.

For PR teams, a citation guarantees presence in the answer even if it doesn't guarantee a click. For a buyer evaluating vendors, that presence shapes their shortlist before they visit any website.

Why AI citations are the new currency of PR visibility

Buyers no longer start their vendor research with a Google search. Sixty-eight percent of B2B buyers open their first vendor conversation inside an AI tool, according to 2026 Wynter research. That reshapes what "visibility" means for a PR campaign.

Google AI Overviews now appear on roughly 48% of tracked queries, representing a 58% year-over-year increase as of February 2026 (BrightEdge). More than half of all searches may surface an AI-generated answer before the blue links. These answers contain citations that PR teams are not yet measuring.

The budget signal is equally direct. Gartner projects a two-fold increase in PR and earned media budgets by 2027, as AI answer engines replace traditional search for discovery. Buyers who never click through still form vendor impressions from what AI tells them.

Traditional traffic metrics don't capture much of this visibility. Earning strong AI citation frequency builds pipeline through a channel that does not show up in Google Analytics. Ignoring this channel results in coverage that works for SEO but may fail to reach buyers during their research phase.

Where AI citations come from: what the data shows

Earned media accounts for 84% of AI citations, based on an analysis of 25 million+ links across ChatGPT, Claude and Gemini in 17 industries. Journalism comprises 27% of sources. Paid content represents only 0.3%. The harder problem is that the source mix shifts by platform.

Google AI Overviews

Google AI Overviews favor brand-owned domains, which account for 59.8% of citations in that engine (OtterlyAI analysis of one million data points). For brands with established owned content and strong domain authority, Google AIO is the most accessible citation surface. Thought leadership on your own site can feed Google's citation layer without a third-party intermediary, provided it is crawlable and structured correctly.

Perplexity

Perplexity draws 16.9% of its citations from Reddit and community forums, a higher share than any other major engine. A press strategy alone won't capture that. Contributed posts, expert responses in vertical forums and product reviews on specialized platforms all feed the citation pool here.

ChatGPT

ChatGPT cites in 96% of responses. Its citation mix skews heavily toward journalism and third-party sources. Social media accounts for roughly 9% of total AI citations as of early 2026 (Tinuiti and Profound). Reddit holds a disproportionate share of that total.

Earned media plays a much larger role across ChatGPT, Claude and Gemini, driving the 84% aggregate figure. Brand-owned content matters more for Google AI Overviews, while community participation matters for Perplexity. PR teams that focus only on journalists may miss opportunities to build visibility across these other AI citation surfaces.

How to track AI citations for your brand

To track AI citations, answer three questions. Each one points to a different data source.

Are you appearing in AI answers at all?

The "Share of Model" question measures how often your brand surfaces when AI engines respond to queries in your space. You might currently check this by manually prompting ChatGPT or Perplexity with a handful of queries. That's useful for a spot check, but it doesn't scale to the frequency needed for trend measurement.

Who is driving your citations?

Even when a brand appears in AI answers, most teams don't know which outlets or journalists are responsible for it. Identifying the top-cited outlets for your topic queries allows you to prioritize media relationships effectively.

Which URLs are AI engines actually citing?

Source-level visibility identifies whether coverage on named publications is feeding AI retrieval. A placement in a well-trafficked outlet that AI engines do not cite provides less citation value than a placement in a niche trade that they do.

Generative Pulse answers all three questions within a single workflow. The AI Share of Voice Dashboard tracks brand representation across LLMs. The Top Cited Journalists view identifies media relationships driving pickup, while Source-Level Insights show referenced URLs. As Muck Rack's AI citation FAQ puts it, AI rewards credibility signals. The data shows those signals come from earned media, connecting press coverage directly to AI presence.

How to earn more AI citations

Four levers are within a PR team's direct control.

Many assume that AI citations require placement in national publications. Citation data reveals a different reality. Niche trade outlets generate citations for industry queries where broad national publications simply don't publish. Those trades provide the depth AI engines need for fintech compliance or supply chain software queries. Sandy Pell of Invoca describes how targeting the right outlets has been central to their AI-driven comms strategy. For most B2B brands, a steady presence across five relevant vertical trades will generate more citations than a single mention in a general business outlet.

Common reasons brands miss out on AI citations

Good media relations can still miss the citation layer. Three failure modes explain why.

• Technical crawler blocks. OtterlyAI's 2026 analysis found that 73% of sites have technical barriers blocking AI crawler access. These include robots.txt rules, CDN configurations, or JavaScript rendering issues. Coverage that sits behind a crawler block never enters the citation pool.
• Targeting outlets AI engines don't pull from. Data shows that AI news citations tend to come from a small group of outlets, per news citation research published in July 2025. Mid-tier generalist press can generate strong clip counts while missing the publications AI engines actually sample.
• Infrequent coverage. Content ages out of AI retrieval windows quickly. Brands generating one or two placements per quarter are not maintaining a citation-eligible presence between those windows. Consistent coverage often matters as much as, or more than, landing a single marquee placement.

Track AI citations with Generative Pulse

Generative Pulse maps to the three questions this article walks through. The AI Media List Builder identifies journalists whose coverage generates citations. The Share of Voice Dashboard tracks LLM presence over time. Source-Level Insights show which retrieval URLs AI engines are pulling.

If you don't know which outlets are feeding AI citations in your category, you don't know whether your earned media is working. PR teams already have the relationships and instincts this requires. The scoreboard has just been missing.