Concepts
What is query fan-out?
Query fan-out is the process by which an AI answer engine silently expands 1 user prompt into multiple sub-queries before retrieving passages. Seer Interactive (2026) found AI engines average ~10.7 sub-queries per prompt, with ~95% carrying zero search volume — meaning standard keyword tools cannot see them. Seeklyn models 8 fan-out angles and grades each as covered, weak, or missing.
Fan-out coverage appears in every Seeklyn content report — run one free at /dashboard/content.
What is query fan-out?
Query fan-out is the silent expansion of a single user prompt into multiple sub-queries that an AI engine uses to retrieve passages from across the web before synthesizing an answer. Seer Interactive (2026) measured that AI engines average ~10.7 sub-queries per head prompt, and roughly 95% of those sub-queries have zero recorded search volume — so standard keyword tools that only show search-volume data are structurally blind to most of the retrieval that actually happens.
The implication for content is that optimizing for the 1 visible head query is no longer sufficient. A page that answers "what is AEO?" but has no passage addressing "how does AEO work?", "AEO vs SEO", or "is AEO worth it?" will be absent when the engine fans out to those angles — even if it ranks well in classic search results.
What 8 fan-out angles does Seeklyn model?
Seeklyn models 8 deterministic fan-out angles drawn from common sub-query patterns observed in AI engine behavior. For each angle, Seeklyn checks whether any of your content passages covers it (covered), mentions it tangentially (weak), or misses it entirely (missing). The angles mirror real sub-query composition: definition intent, how-it-works intent, why-it-matters rationale, comparison/alternative framing, cost/pricing intent, best-options framing, worth-it/pros-cons weighting, and temporal/freshness signals.
| Angle | Sub-question template | What makes a passage "covered" |
|---|---|---|
| definition | What is [topic]? | Passage defines the term with "means", "refers to", "is defined as", or similar |
| how | How does [topic] work? | Passage explains a process, steps, or mechanism |
| why | Why does [topic] matter? | Passage states a reason, rationale, or benefit |
| comparison | [topic] vs alternatives | Passage uses "vs", "compared to", "alternative", or names a competitor |
| cost | How much does [topic] cost? | Passage mentions cost, price, fee, free, or budget |
| best | Best [topic] options | Passage frames top, best, recommended, or a ranked shortlist |
| worth | Is [topic] worth it? | Passage addresses pros, cons, drawbacks, or a worth-it verdict |
| temporal | [topic] in 2026 | Passage carries a year, "latest", "current", or a recent date signal |
How is the fan-out coverage score calculated?
The fan-out coverage score is computed as (covered×100 + weak×50) / total, where total is always 8 (the fixed angle count). A covered angle — one where a passage both mentions the topic and carries 2 or more answer-signal terms — contributes 100 points; a weak angle contributes 50; a missing angle contributes 0. The result is a 0–100 score: 8 covered angles = 100, 8 missing = 0.
Seeklyn requires co-location to score an angle as covered: the topic terms and the answer-signal terms must appear in the same passage, not just anywhere in the document. A stray mention of "cost" in an unrelated passage does not cover the cost angle — the passage must also be about the topic.
Why do ~95% of fan-out sub-queries have zero search volume?
About 95% of the sub-queries AI engines generate have zero recorded search volume because they are restatements, reformulations, or implied sub-questions that no human has typed verbatim into a search bar at any scale (Seer Interactive, 2026). These sub-queries are constructed dynamically by the model based on the user's intent, not selected from a keyword database. Standard keyword tools only report on queries that have been searched frequently enough to record volume, so the vast majority of AI sub-queries are invisible to them — which is exactly why a fan-out coverage model is needed.
Frequently asked questions
Does every AI engine fan out queries the same way?
No — the number and composition of sub-queries varies by engine and query type. The ~10.7 average (Seer Interactive, 2026) is a measured average across engine types, not a fixed rule for each one. Seeklyn's 8 angles are a deterministic model of common intent patterns, designed to be conservative and auditable rather than a black-box prediction.
Can I see which of my passages covers which angle?
Yes — Seeklyn's content report shows each of the 8 fan-out angles for your target query, marks each as covered / weak / missing, and links the angle to the specific passage that best addresses it. Missing and weak angles also show a one-line hint explaining what to add.
Is a covered score of 100 achievable?
Yes, for tightly scoped topics where all 8 angles are addressable. Some topic types (a narrow technical procedure, a product with no competitors) may structurally lack a cost or comparison angle, in which case those angles will remain missing without harming the content's citation performance for the angles it does cover.