ChatGPT fan-outs have doubled in length

From late 2025 into early 2026, average ChatGPT fanout length roughly doubled (about 6 words → ~12), consistently across countries and languages. The number of fanouts per prompt stayed roughly flat — ChatGPT got more precise per search, not noisier.

Why it matters

Longer fanouts retrieve more long-tail results and give embedding matchers more context. Broad, generic pages struggle. Detailed pages that match specific intents win retrieval.

Why it's happening

• Retrieval stacks reward richer queries
• AI-generated web content saturation pushes models toward sharper filters

Team takeaway

Track fanout text over time in Anny. When the hidden queries get more specific, your content brief should too. Treat this as a global architectural shift — not a local experiment.