Zyppy "Fan-out Framework: 5 Steps to Optimize for Fan-out Queries" — Cyrus Shepard's Definitive AEO Playbook

Executive Summary

Cyrus Shepard published the most actionable AEO framework to date on Zyppy Signal (June 25, 2026): a 5-step process for optimizing content to appear in AI fan-out queries — the sub-queries AI engines generate in the background before synthesizing answers. Critical insight: Google uses FastSearch/RankEmbedBERT for AI Overviews, which uses 70 days of user data vs 13 months for traditional Search. This means you can influence AI answers faster than traditional rankings — content updates influence AI citations within weeks, not months.

The 5-Step Fan-out Framework

StepActionTool/Method
1 Find keywords you already rank for (top 10-20 in Google) — the #1 predictor of AI citation GSC, Ahrefs, Semrush
2 Find fan-out queries — the sub-queries AI engines generate from your primary keywords QueryFan, Qforia, Bing WMT AI Performance Report grounding queries, AI-generated prompts
3 Determine most important fan-out topics — AI cleanup → keyword clustering → gap analysis Clustering tools, manual review
4 Optimize pages or create new pages — align headings/content to fan-out queries; avoid Godzilla pages or low-quality thin pages Content management, on-page optimization
5 Measure results — track AI citation changes after optimization Bing WMT AI Performance Report (best), Google GSC AI Features (limited), premium trackers

Why the 70-Day Window Changes Everything

AI Rankings Respond Faster Than Traditional Rankings

Google's AI Overviews use FastSearch/RankEmbedBERT, which operates on a 70-day user data window. Traditional Search uses a 13-month window. The practical implication: content updates influence AI citations within weeks, while traditional rankings take months to respond. This is the single most important tactical insight for AEO — you can iterate on AI visibility at a velocity that's impossible in traditional SEO.

Entity Data Ages Fast — 70 Days Is Your Window

Entity data ages faster than general content: pricing changes, new features, competitor landscape shifts, integration changes. The 70-day window means stale entity data loses AI visibility within a single quarter. Implement a 60-day entity refresh cycle — update each entity page with new data every 60 days, well within the 70-day AI citation window. This creates a continuous freshness signal that AI systems reward.

Fan-out Queries Are Entity Discovery Multipliers

Entity queries generate rich fan-outs that most entity sites ignore:

Each fan-out gap is a missed AI citation for entity queries. The sites that map and fill fan-out gaps capture disproportionate AI visibility.

Strategic Actions for Hermes Agent

1. Fan-out Audit for Top 10 Entity Pages

For each top-ranking entity page, run the fan-out framework:

  1. Identify current ranking keywords (Step 1)
  2. Generate fan-out queries using Bing WMT AI Performance Report grounding queries or AI prompts (Step 2)
  3. Cluster and prioritize fan-out gaps (Step 3)
  4. Create or optimize pages for high-priority fan-out queries (Step 4)
  5. Track citation changes in Bing WMT (Step 5)

Entity queries generate the richest fan-outs — and the most fan-out gaps. Each filled gap is a new AI citation opportunity.

2. Implement a 60-Day AI Refresh Cycle

The 70-day AI window means AI citations respond faster to content updates. Implement a 60-day entity refresh cycle:

Entity data ages fast — pricing, features, and competitor landscape change frequently. 70 days is your citation window. Stay inside it.

3. Bing WMT AI Performance Report as Primary AEO Dashboard

Bing shows grounding queries and citation counts per URL — Google's GSC AI report shows impressions only, no query data. Use Bing WMT as your primary fan-out measurement tool:

Entity sites with strong Bing presence get a strategic advantage in AI visibility measurement — Bing's tooling is currently superior to Google's for AEO diagnostics.

Source

📋 https://hermes-agent.reviews/zyppy-fanout-framework-shepard-aeo-playbook.html
SEO Signal tracked by hermes-agent.reviews — June 30, 2026