AI Citation Index | Trakkr Research

Domain concentration, source categories, and prompt intent across aggregate AI citation appearances. Answer pages, reference facts, and live trackers drawn from this study.

Methodology: Derived from AI Citation Index and updated August 18, 2026.

What this hub contains

Domain concentration, source categories, and prompt intent across aggregate AI citation appearances. Answer pages, reference facts, and live trackers drawn from this study.

Answer Pages

Narrow questions answered directly from the study.

• Which sources dominate AI citations? - YouTube leads the current Trakkr index with 3.47% of citation appearances, followed by English Wikipedia at 1.83% and Reddit at 1.74%. Together, the top three account for 7.04%.
• How concentrated is the AI citation landscape? - The top three domains account for 7.04% of citation appearances, while the index spans 529,165 unique domains and 2.60 million cited URLs. That is a visible head followed by a very broad tail.
• Does AI rely more on reference sites than the open web? - The index cannot support that claim. Reference sources account for 1.98%, while 86.75% remains unclassified in “other”. The taxonomy is too partial to compare reference sites with the open web as a whole.
• Does social content matter for AI citations? - Yes. Social sources account for 7.20% of citation appearances in the current index, led by YouTube, Reddit, and LinkedIn.
• Do review sites still matter to AI models? - Review sources account for 1.15% of citation appearances in the current index.
• How often do docs get cited by AI? - Domains in the current docs map account for 0.17% of citation appearances. That is a partial category result, not a complete measure of documentation pages across the web.
• Why does Wikipedia still win so often? - English Wikipedia is no longer first in the current index. It ranks second with 886,607 citation appearances, or 1.83%, behind YouTube at 3.47%. The aggregate records that pattern but does not establish why it occurs.
• How big is the competitive set for AI citations? - The current index contains 529,165 unique cited domains across 1,643 tracked brands. That is the observed citation surface, although not every cited domain is a commercial competitor.
• What source categories show up most in AI answers? - Social is the largest named category at 7.20%, followed by reference at 1.98% and reviews at 1.15%. However, 86.75% remains in “other”, so the source-type taxonomy is incomplete.
• Do brands need third-party sources to win AI citations? - This index alone cannot answer that. It aggregates cited domains but does not label ownership for every URL or provide an experimental comparison between owned-only and third-party strategies.
• What should a brand prioritize after reading the source data? - Start by mapping which domains recur for the brand’s tracked prompts, then separate observed source gaps from assumptions. The public index is a benchmark, not proof that one source strategy causes visibility.
• Does raw domain count matter more than repeat-cited authority? - Raw domain count and repeat citation count answer different questions. The index shows 529,165 domains and a 7.04% share for the top three, but it does not measure “authority” as a causal variable.

Reference Facts

Short, quotable claims with metrics and methodology context.

• YouTube leads the current citation index - YouTube ranks first in the August 18, 2026 snapshot, ahead of English Wikipedia and Reddit.
• The citation graph spans more than 500,000 domains - The long tail is enormous even when the head is concentrated.
• Reference sites are the clearest named source family - Reference is the strongest named source class even though the long tail is larger in aggregate.
• Social sources are material, not noise - YouTube, Reddit, and LinkedIn keep appearing in the citation layer.
• Review sites are small in share but high in intent - Review domains are a smaller slice of raw volume than many teams expect.
• Docs are a niche but important citation layer - Docs are not broad recommendation winners, but they show up when the prompt becomes technical.
• Most citations remain unclassified by source type - Most citation volume sits outside the hand-maintained named source map.
• The current index tracks more than 2.6 million cited URLs - AI systems do not just cite domains. They return to a large but finite set of pages.

Trackers

Live benchmark views built from the study’s most reusable dimensions.

• AI citation share by source category - Current category mix in the live Trakkr citation index.
• Top cited domains in the live index - Current head domains in Trakkr’s live public citation graph.