---
name: seo-aeo-geo-optimizer
description: Review a specific page, draft, or piece of content for traditional search engine optimization (SEO), answer engine optimization (AEO, for featured snippets and "People Also Ask"), and generative engine optimization (GEO, for citation by AI tools like ChatGPT, Perplexity, and Google AI Overviews), then produce a prioritized, sourced list of fixes. Trigger this when the user shares a URL or a content draft and asks for an SEO review, wants to know if something will get picked up by AI search, or asks about search or AI visibility for a specific page. Not for keyword research from scratch with no page or draft in hand, and not for paid search or ads.
---

# SEO / AEO / GEO Optimizer

Traditional SEO, AEO, and GEO are not three separate disciplines needing three separate audits. They largely share the same underlying work: content that is easy to crawl, clearly answers a specific question, and is written in a way that's easy to extract and cite. This skill evaluates a page or draft across all three lenses in one pass and gives one prioritized fix list, rather than three redundant ones.

## Definitions to work from

- **SEO** targets ranking position on traditional search engine results pages. Signals: crawlability, on-page relevance, keyword usage, backlinks, technical health (page speed, mobile-friendliness), authority.
- **AEO** targets winning the direct answer: featured snippets, "People Also Ask" boxes, voice results. Signals: concise, direct answers to specific questions, question-format headings, scannable structure, FAQ schema.
- **GEO** targets being cited, quoted, or recommended inside AI-generated answers (ChatGPT, Perplexity, Google AI Mode/Overviews, Claude). Signals: clear entity identification, authoritative and specific claims, content structured so a generative model can accurately summarize and attribute it, and technical scaffolding like structured data.

AEO and GEO overlap heavily; the tactics that work for one largely work for the other. Treat them as one combined lens with SEO as a separate, complementary lens, not three independent checklists.

## Workflow

1. **Get the input.** Accept a live URL (fetch it directly) or a pasted draft. If it's a live URL, pull the actual page content and, where visible, the page's title tag, meta description, heading structure, and any structured data in the source.

2. **Check the SEO layer.**
   - Is there a clear, specific title tag and meta description matching what a searcher would type?
   - Is the heading structure (H1, H2s) logical and does it reflect real subtopics, not just formatting?
   - Does the content target a specific, identifiable search intent, or is it generic?
   - Note technical items you can observe (missing alt text, broken structure) but don't fabricate performance data (page speed, backlink counts) you can't actually see. If asked about those, say a dedicated tool (e.g. Google Search Console, PageSpeed Insights) is needed.

3. **Check the AEO/GEO layer.**
   - Does the content directly and concisely answer a specific question somewhere near the top, in a form that could stand alone as a snippet or AI-generated answer?
   - Are claims specific and attributable (named, dated, sourced) rather than vague, since generative engines favor content they can confidently cite?
   - Is there a clear entity (organization, person, product) that the content is clearly "about," so an AI system can tell what and who this content represents?
   - Would FAQ-style question headings improve extractability, and are any already present?
   - Is there structured data (schema.org markup, visible in page source) supporting this content? Note if absent.

4. **Prioritize, don't just list.** Identify the single highest-impact fix first, then 2-4 secondary fixes. Prioritize based on what's most clearly missing relative to the content's apparent goal, not a generic checklist run in order.

5. **Self-check before delivering:**
   - Every observation traces to content or source code actually read, not a general assumption about what pages "usually" need
   - No performance claim (rankings, traffic impact, citation likelihood) is stated as certain; frame likely impact as likely, not guaranteed
   - No em dashes
   - If the URL couldn't be fetched or the draft is too thin to assess meaningfully, say so rather than producing a generic checklist

## Output format

- **What this page is trying to do:** 1 sentence, your read of its intent
- **SEO:** 2-4 bullet observations
- **AEO/GEO (AI and answer visibility):** 2-4 bullet observations
- **Fix this first:** 1-2 sentences, the single highest-priority change and why
- **Then:** short bulleted list of secondary fixes

## Notes

- If the user wants ongoing tracking of AI citation performance (whether a brand actually gets cited by ChatGPT/Perplexity/etc. over time), that requires a dedicated monitoring tool; this skill evaluates a single page's readiness, not live citation tracking.
- Keep the combined SEO/AEO/GEO framing in the output. Splitting it into three full separate sections overstates how different the underlying work actually is.