Content analyzer
What is an evidence gap (and how does Seeklyn find them)?
Seeklyn detects up to 6 evidence gaps per passage across 5 gap types — vague, comparative, freshness, hedge, and source — and emits a typed, actionable prompt for each. Detection only: Seeklyn never invents a figure, but it flags the exact phrase and explains what kind of concrete data is missing.
Evidence gaps appear beneath each passage in a Seeklyn content report, grouped by type with a suggested fix — run your page free at /dashboard/content.
What is an evidence gap?
An evidence gap is a specific claim in your content where Seeklyn can see that a concrete figure, source, or year is missing — and its absence makes the passage harder for an answer engine to quote as authoritative. Princeton's GEO study (KDD 2024) found that adding statistics or quoted sources raises AI citation rates by +30–41% per intervention. An evidence gap prompt tells you exactly which phrase to strengthen and what kind of evidence to add.
Evidence gaps are detection-only: Seeklyn identifies the structural pattern (a vague quantifier, a bare comparative, a freshness word without a year, a hedged claim, a stat without attribution) but never generates a replacement figure. Fabricating a number would be worse than the gap it fills — so Seeklyn names the problem and the type of evidence needed, leaving you to provide the verifiable fact.
What are the 5 evidence gap types?
Seeklyn classifies every gap into 1 of 5 types based on the linguistic pattern that triggered it. Each type has a different fix: vague quantifiers need a specific figure; bare comparatives need a magnitude and baseline; freshness words need an anchor year; hedges need either a cited figure or a plain declarative rewrite; source gaps appear when a fact is present but attribution is missing.
| Type | Trigger pattern | Example phrase | Fix |
|---|---|---|---|
| vague | Non-specific quantifier (significant, many, several, numerous…) | "significant improvements" | Replace with a specific % or count |
| comparative | Bare comparative with no magnitude or baseline (faster, cheaper, better…) | "loads faster" | Add magnitude and comparison ("2× faster than X") |
| freshness | Recency word without a year in the passage (latest, currently, recently…) | "currently the best" | Anchor with a year ("as of 2026") |
| hedge | Hedge word in a sentence with no fact and no attribution (may, might, could, possibly…) | "may improve rankings" | State plainly or cite a figure |
| source | Figures present but no attribution phrase in the passage | "conversion rates rose 18%" | Add source ("according to Ahrefs 2026") |
How many gaps does Seeklyn surface per passage?
Seeklyn caps evidence gaps at 6 per passage to keep the report a focused to-do list rather than an overwhelming wall of brackets. When more than 6 patterns fire in a single passage, Seeklyn surfaces the first 6 in detection order (vague and comparative issues appear before freshness and hedge issues in the scan). The 6-cap means the highest-priority gaps always appear; fix them first, then re-run to surface any remaining ones.
Evidence gaps are reported at the passage level alongside the 8 extractability flags, but they are a separate signal. A passage can have 0 extractability flags (clean structure) yet still carry 3 evidence gaps (vague claims, bare comparatives, missing attribution). Both matter for citation quality, but they fix different things: flags fix structure, gaps fix evidence depth.
Frequently asked questions
Why does Seeklyn detect gaps instead of filling them automatically?
Generating a figure that does not exist in your research would be a fabrication — and a fabricated stat cited by an AI engine is worse than a vague one. Seeklyn can identify the structural gap reliably; only you can supply the verified fact. Detection is honest, generation would not be.
What is the difference between an evidence gap and an extractability flag?
Extractability flags (like HEDGE or NO_ATTRIBUTION) affect the passage's citability score (0–100) and are the primary score drivers. Evidence gaps are advisory prompts that do not change the score directly — they tell you what evidence to add to raise fact density and citation lift in the next revision.
Does fixing all evidence gaps guarantee a citation?
No — citations also depend on the Retrieval stage (query-term overlap) and the Citation stage (all 8 extractability flags). Evidence gaps are 1 of several quality dimensions. Fixing them strengthens the evidence base, which Princeton GEO (KDD 2024) found raises citation rates by +30–41% per improvement — but Retrieval and structural flags must also be addressed.
How does the 'source' gap differ from the NO_ATTRIBUTION flag?
The NO_ATTRIBUTION extractability flag fires when numeric claims lack an attributable source and reduces the passage's 0–100 citability score. The 'source' evidence gap is a companion advisory prompt that provides the specific fix text: it appears when figures are present but attribution is missing, nudging you toward the exact change ('according to …' or a named study). The flag scores the defect; the gap tells you how to fix it.