A four-dimension framework for building GEO prompt libraries, backed by B2B survey data
yangyi · x · 2026-11-02
A practical write-up on the hardest part of GEO (generative engine optimization): constructing intent questions. Core framework: build a question library along four dimensions — decision intent × use scenario × constraints × conversation stage — instead of endlessly expanding brand keywords. Grounded in Semrush's 2026 B2B survey (61% describe scenarios, 56% ask for vendor comparisons, 45% add budget/feature constraints), G2's 1,076-respondent study (comparing vendors is the top AI-search use), consumer research (52% state all constraints upfront, 33% refine over multiple turns), and a Trip.com AI assistant study of 280K requests. It stresses survey ratios can't be extrapolated into real query volume: build coverage matrices first, then use third-party AI search-volume tools to prioritize content.
More from Venture
- Grok Bot Quotas Jump 20-40% Overnight as Musk Admits Subsidy Is Unsustainable — yangyi · 2026-11-02
- What 150k-follower Bilibili gaming creators charge: 8,500 yuan per video — ezshine · 2026-11-02
- Microsoft issues $25B bonds in $92.28B revenue week as OpenAI trims round by $10B — YvesMulkers · 2026-11-02
- Amazon's $215B capex exceeds $168B operating cash flow, FCF turns -$42B — thealexbanks · 2026-11-02
- Saudi Aramco takes 49% stake in AI firm Humain in $2.6 billion deal — The Decoder · 2026-11-02
- Stratechery: Google's earnings miss mirrors Facebook's—but its agent play is bolder than credited — Stratechery · 2026-11-02