aaron-he-zhu/aaron-marketing-skills / seo-geo/implement/geo-content-optimizer
geo-content-optimizer
VerifiedUse when the user asks to "optimize for AI citations"; improves citation readiness for ChatGPT, Perplexity, AI Overviews, Gemini, and Claude. Not for structural on-page SEO — use on-page-seo-checker; not for net-new drafting — use content-writer. AI引用优化/GEO优化/AI搜索
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 9.1 KB |
| references/ai-citation-patterns.md | markdown | 10.5 KB |
| references/ai-overview-recovery.md | markdown | 4.8 KB |
| references/geo-optimization-techniques.md | markdown | 3.4 KB |
| references/instructions-detail.md | markdown | 4.1 KB |
| references/medium-github-surfaces.md | markdown | 5.8 KB |
| references/quotable-content-examples.md | markdown | 3.6 KB |
references/ai-overview-recovery.md
View on GitHub# AI Overview Recovery Playbook
Referenced from [SKILL.md](../SKILL.md). Use when AI Overviews are diverting traffic away from your pages.
---
## Symptom profile
- Organic CTR on >=5 queries dropped 20-60% in a 2-4 week window
- SERP screenshots show AI Overview block above organic results
- Impressions stayed flat or rose; clicks fell
- Page still ranks top-3 organically but gets fewer clicks
## Phase 1 — Measure the damage
### 1.1 Identify affected queries
From `~~search console`: filter last 28d vs previous 28d, export queries with CTR change < -20% AND impressions change > -10%. Flag top 20 queries by click loss.
### 1.2 Confirm AI Overview is the cause
For each target query (incognito, location-matched): Is AI Overview present? Does it cite your page? Visible or hidden citation? How many organic results above fold?
### 1.3 Segment the 20 queries
| Segment | Pattern | Recovery strategy |
|---------|---------|-------------------|
| A | AI Overview + cites you | Low priority — optimize carousel ordering |
| B | AI Overview + does NOT cite you | HIGH — lost ranking surface AND citation |
| C | AI Overview + cites competitor | HIGH — identify what competitor has |
| D | No AI Overview, intent shifted | Different problem — re-check ranking |
Focus on segments B and C.
## Phase 2 — Diagnose why AI Overview skipped you
### 2.1 Answerability check
- [ ] First 100 words contain a direct answer (not just the keyword)?
- [ ] Single `<h2>`/`<h3>` rephrases the query as a statement?
- [ ] 3-5 ordered points answering sub-questions?
- [ ] Answer is quote-able — standalone sentence that makes sense out of context?
### 2.2 Freshness check
- [ ] Last-updated date visible in HTML and rendered
- [ ] Published/updated within last 12 months
- [ ] Statistics with year qualifier: "In 2026, X..."
- [ ] Reference links to pages updated in last 12 months
### 2.3 Authority check (E-E-A-T)
- [ ] Author bio visible with credentials
- [ ] Author's Knowledge Graph entity exists
- [ ] External citations from authoritative sources
- [ ] Brand entity recognized by Wikidata / Google KG (see `entity-registry`)
### 2.4 Structural / crawlability check
- [ ] Content rendered in initial HTML (not post-hydration)
- [ ] `robots.txt` allows `Googlebot`; do not require `Google-Extended` unless the site intentionally allows Google AI training/product use
- [ ] Schema present: `Article` / `HowTo` / `FAQPage`
- [ ] No paywall blocking first answer
## Phase 3 — Rewrite targets
### 3.1 Answer-first rewrite template
```
**[Direct query-answering sentence in <=30 words.]**
[2-sentence expansion with key specifics — numbers, proper nouns, date.]
[Transition sentence.]
[Jump-linked TOC or 3-5 H2s covering sub-questions.]
```
### 3.2 Quotable-chunk insertion
Add 2-3 standalone sentences (<=30 words each) containing a specific number/year/qualified claim, flanked by attribution.
### 3.3 FAQ appendix
Append `<h2>FAQ</h2>` with 6-10 PAA-matching questions. Each answer 40-60 words. Mark up with `FAQPage` schema.
### 3.4 Structured data that supports citation
Add markup that genuinely matches the page: `Dataset` (real survey/benchmark data you publish),
`HowTo` (actual step-by-step), `Article` with author linked to a KG entity. Only add `Review` /
`aggregateRating` when the page has real first-party reviews — self-assigned ratings to win a rich
result are structured-data spam under Google's review policy and risk a manual action. Schema must
describe what is on the page, never decorate it with claims that aren't.
## Phase 4 — Monitor recovery (T+7, T+14, T+28)
- T+7: SERP AI Overviews changed? (often no)
- T+14: 2-3 pages appearing in citations?
- T+28: CTR recovered to >50% of pre-drop baseline?
If no recovery at T+28:
- Segment C: run `competitor-analysis` on cited page
- Segment B: query likely answered from training data; shift to long-tail + adjacent queries
## Red flags — stop and re-scope
- AI Overviews affecting queries you shouldn't target — that traffic wasn't valuable
- AI Overviews cite 3+ competitors and you're DR <30 — run `entity-registry` first
- AI Overviews on branded queries cite competitor comparisons — knowledge-graph gap; hand to `entity-registry`
## Handoff
- **Status**: DONE | DONE_WITH_CONCERNS
- **Objective**: "AI Overview recovery plan for [N] affected queries"
- **Key Findings / Output**: segmentation table, per-query rewrite priority, estimated recovery timeline
- **Evidence**: GSC export ref, SERP screenshots, pre/post CTR deltas
- **Open Loops**: monitoring windows (T+7/14/28), pages pending rewrite, entity gaps
- **Recommended Next Skill**: `content-writer`, `entity-registry`, `rank-tracker`
## See also
- [entity-registry](../../../../protocol/entity-registry/SKILL.md)
- [content-writer](../../content-writer/SKILL.md)
- [technical-seo-checker LLM crawler guide](../../../tune/technical-seo-checker/references/llm-crawler-handling.md)
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |