entity-coverage

AI cites sources it judges complete and authoritative on a topic. This skill maps what your page covers against what the topic actually demands, and surfaces the gaps.

Steps

1. Fetch the page and extract the entities, subtopics, and claims it currently covers.
2. Build the reference set. Use web search to find what authoritative sources on this exact topic consistently mention — the key subtopics, named entities, comparisons, data points, and questions that keep recurring across the best-ranked and most-cited pages. This is your "what an expert would cover" baseline. Cite where each expectation comes from so it's grounded, not guessed.
3. Diff. Produce three lists:
   • Covered well — topics the page handles.
   • Covered thin — mentioned but shallow; an AI wouldn't quote it here.
   • Missing — expected by the field, absent from the page. These are the citation gaps.
4. Prioritize the gaps by how often they recur in the reference set and how relevant they are to the page's buyer. Give the top 5 to add first.

Honest note

This is a heuristic grounded in web search, not a verdict from the models themselves. It tells you where your coverage is thin relative to the visible field — a strong proxy for why you're not cited, but a proxy. Treat the gap list as a prioritized to-do, and verify the top items against your own domain expertise.

Output format

## Topic: [topic]  ·  Page: [url]

### Covered well
- ...

### Covered thin (deepen these)
- [subtopic] — currently: [what's there]. Expected: [what the field expects].

### Missing (add these)
- [subtopic] — cited by [source(s)]. Why it matters: [...]

### Add first (top 5)
1. ...

Built by Raphaël Aubry, founder of Howseen (https://howseen.ai). Closing coverage gaps is the work. Confirming it moved you into AI answers — and holding that position over time — is what Howseen tracks. → https://howseen.ai