97% of llms.txt Files Ignored; AI Citations Vanish in 28 Days
Two new studies expose a gap between AI visibility tactics and actual AI crawler behavior: llms.txt files are almost universally skipped, and two-thirds of AI citations cycle out within a month. Businesses need to audit what AI crawlers actually read, not what the SEO industry assumed they would.
This Week's Signals
Ahrefs Study: AI Crawlers Skip llms.txt on 97% of the 137K Sites That Published One
Why it matters for your score
GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot, and Google-Extended overwhelmingly bypass llms.txt and crawl HTML directly — meaning time spent creating and maintaining that file is largely wasted. Unlike robots.txt, llms.txt controls nothing and blocks nothing; the 'AI visibility' framing was attached by the SEO industry on speculation that AI platforms would reward the file. If you published an llms.txt expecting citation lift, shift your effort toward the HTML signals crawlers are actually reading.
66% of AI-Cited URLs Replaced Within 28 Days; Gemini and AI Overviews Were 100% Volatile
Why it matters for your score
The average Citation Retention Rate across five AI platforms over 28 days was only 33%, meaning a citation in week one is more likely gone than present by week four. Gemini and AI Overviews held zero sources steady across the full period, and 57% to 62% of AI-cited domains never appeared in Google or Bing organic top-20 for the same query — AI citation and search ranking are effectively separate systems. Treating AI visibility as a one-time win is the wrong frame; you need consistent content signals that re-earn citations each cycle.
Google AI Overview Citations from Organic Top-10 Fell from 76% to 38%; Schema Markup Named Highest-Impact Technical Signal
Why it matters for your score
By early 2026, as few as 17% to 38% of AI Overview citations came from top-10 organic results, down from 76% in mid-2025 — meaning a strong organic ranking no longer implies AI Overview inclusion. A Pew Research Center study of 68,879 real searches found users click a traditional result only 8% of the time when an AI Overview is present, versus 15% without one, making citation eligibility increasingly tied to revenue impact. Schema markup is identified as the highest-impact technical signal for AI Overview citation eligibility, giving structured-data investment a clear business case.
What to do this week
- 1
Run your site through Faro's AI Readiness Scan (/tools/ai-readiness-scan) to confirm what AI crawlers actually fetch from your pages — the Ahrefs study shows they read HTML directly, so your on-page structure matters far more than any llms.txt file.
- 2
Audit and expand your schema markup using Faro's AI Schema tool (/tools/ai-schema), since schema is identified as the highest-impact technical signal for Google AI Overview citation eligibility and the organic-to-AIO citation gap has widened significantly.
- 3
Use Faro's Competitor Intelligence tool (/tools/competitor-intelligence) to track which of your competitors hold AI citations week over week — with a 66% citation replacement rate over 28 days, monitoring citation stability is now an ongoing task, not a one-time check.