Only 11% of AI Citations Overlap Between ChatGPT and Perplexity

What happened

In a 2026 benchmark study, Averi analyzed 680 million AI citations and found that only 11% of the domains cited by ChatGPT also appeared in Perplexity's answers. The rest of the source graph is essentially disjoint. ChatGPT leans heavily on Wikipedia, which accounts for roughly 48% of its cited sources, and gives strong weight to structured reference pages and older, authority-verified content. Perplexity leans on Reddit, which accounts for 47% of its sources, and rewards content published in the last 30 days at more than three times the citation rate of older material. Google AI Overviews sit closer to the classic Google index enriched with authority signals and cite roughly 85% of the time. ChatGPT cites 87% of the time and Perplexity cites 3 to 4 of the 10 pages it visits per query. A single 'be authoritative and helpful' content strategy no longer maps cleanly to any of them.

Why the ChatGPT and Perplexity citation gap matters now

B2B buyers are now splitting research across at least two AI engines during vendor evaluation. If your content is only visible to one, you are being systematically eliminated from the other's answer set before a rep ever hears from that account. Because the two source pools barely overlap, teams that publish once to their own blog and hope for the best are effectively invisible on Perplexity, where Reddit threads and last-30-days content dominate, while teams that rely on Reddit AMAs and news mentions are invisible on ChatGPT, where evergreen reference material wins.

1. 01
   Split your content pipeline by engine target

   Tag each piece of content by which engine it is designed to be cited by. Reference pages, glossaries, and pillar articles ship to your owned domain for ChatGPT. Reddit AMAs, LinkedIn posts, and product changelog notes ship for Perplexity. Google AI Overviews reward both, weighted toward domain authority.
2. 02
   Publish something Perplexity-eligible every week

   Perplexity cites content under 30 days old at roughly 3.2 times the rate of older material. Ship one dated, timestamped piece per week: a data note, a product release, a benchmark update. Old evergreen articles do not compete for Perplexity's freshness weight, no matter how well ranked they are elsewhere.
3. 03
   Track citation share per engine, not aggregate visibility

   Use Profound, Peec, or Otterly to break your citations down by engine rather than reporting a single 'AI visibility score.' A 40% share of voice on ChatGPT paired with a 3% share on Perplexity is a very different problem from an even split, and both need different fixes.

By the numbers: ChatGPT cites Wikipedia in ~48% of sources; Perplexity cites Reddit in ~47%. Perplexity's freshness bonus makes sub-30-day content ~3.2x more likely to be cited. Only 11% of domain citations overlap between the two engines.

What to do this week

Pick your top three commercial queries — the ones a buyer would type before a demo. Run each in ChatGPT and Perplexity and record every domain in each answer. If your brand appears in neither, you have a citation gap on both engines. If it appears in one, you have a channel-specific gap. Spend this week seeding a single well-sourced Reddit thread and updating one evergreen pillar page with a 2026-dated stat — those two moves target the two engines' opposite biases.