Ai Seo

---
name: Ai Seo
description: Use when the user wants content cited by AI systems (ChatGPT, Perplexity, AI Overviews, Gemini, Copilot) — AEO, GEO, LLMO, AI visibility, or AI citations.
category: SEO & Content
version: 2.0.0
tools: []
---

# AI SEO

Optimize content so it gets discovered, extracted, and cited by AI systems — Google AI Overviews, ChatGPT, Perplexity, Claude, Gemini, and Copilot. The goal is citation, not just ranking.

## Before Starting
Search the Knowledge Base for existing product-marketing context before asking questions, and only ask about what isn't already covered. Otherwise gather:

1. **Current AI visibility** — Does the brand already appear in AI-generated answers? Have ChatGPT, Perplexity, or Google AI Overviews been checked for key queries? What queries matter most?
2. **Content & domain** — Content types produced (blog, docs, comparisons, product pages), traditional SEO/domain strength, existing schema markup.
3. **Goals** — Get cited as a source, appear in AI Overviews for specific queries, out-cite named competitors, optimize existing content or create new AI-optimized content.
4. **Competitive landscape** — Top competitors in AI search results, and where they're cited but the user isn't.

## How AI Search Works

| Platform | How it works | Source selection |
|---|---|---|
| Google AI Overviews | Summarizes top-ranking pages | Strong correlation with traditional rankings |
| ChatGPT (with search) | Searches the web, cites sources | Draws from a wider range than just top-ranked results |
| Perplexity | Always cites sources with links | Favors authoritative, recent, well-structured content |
| Gemini | Google's AI assistant | Pulls from Google index + Knowledge Graph |
| Copilot | Bing-powered | Bing index + authoritative sources |
| Claude | Brave Search (when enabled) | Training data + Brave search results |

Traditional SEO gets a page ranked; AI SEO gets it **cited**. A page ranking #2 or #3 can still get cited over the #1 result if it's better structured — AI systems weigh content quality and extractability, not just rank position.

**Key stats:** AI Overviews now appear in ~45% of Google searches and cut click-through to sites by up to 58%. Brands are 6.5x more likely to be cited via third-party sources than their own domain. Optimized content is cited roughly 3x more often than non-optimized content, and adding statistics/citations alone lifts visibility 40%+.

## AI Visibility Audit

1. **Test 10-20 priority queries** across platforms — "What is [category]?", "Best [category] for [use case]", "[Brand] vs [competitor]", "How to [problem]", "[category] pricing." Log, per query, whether an AI Overview/ChatGPT/Perplexity answer exists and who's cited.
2. **Analyze citation gaps** — where competitors are cited and the user isn't, compare content structure/extractability, authority signals (citations, stats, expert quotes), freshness, schema markup, and third-party presence (Wikipedia, Reddit, review sites).
3. **Content extractability check**, per priority page: clear definition in the first paragraph; self-contained answer blocks that work without surrounding context; sourced statistics; comparison tables for "X vs Y" queries; an FAQ section in natural-language phrasing; schema markup (FAQ/HowTo/Article/Product); author attribution; updated within the last 6 months; headings that mirror query phrasing; AI bots allowed in robots.txt.
4. **AI bot access check** — confirm robots.txt doesn't `Disallow` GPTBot/ChatGPT-User (OpenAI), PerplexityBot, ClaudeBot/anthropic-ai (Anthropic), Google-Extended (Gemini/AI Overviews), or Bingbot (Copilot). Blocking prevents both AI training and citation on that platform; a common middle ground is blocking training-only crawlers like CCBot (Common Crawl) while leaving the search-facing bots open.

## Optimization Strategy: Three Pillars

**1. Structure — make content extractable.** AI systems extract passages, not whole pages; every key claim should work as a standalone statement. Use definition blocks for "What is X" queries, step-by-step blocks for "How to X," comparison tables for "X vs Y," pros/cons blocks for evaluation queries, FAQ blocks, and cited statistic blocks. Lead every section with the direct answer instead of burying it, keep key answer passages to 40-60 words (the sweet spot for snippet extraction), use H2/H3 headings phrased the way people search, and prefer tables and numbered lists over prose for comparisons and processes.

**2. Authority — make content citable.** Princeton's GEO research (KDD 2024, studied on Perplexity.ai) ranked nine optimization methods by visibility boost: citing sources (+40%), adding statistics (+37%), adding quotations (+30%), authoritative tone (+25%), improved clarity (+20%), technical terms (+18%), unique vocabulary (+15%), fluency optimization (+15-30%) — while keyword stuffing actively *hurts* visibility (-10%). Fluency plus statistics together give the largest combined boost, and low-ranking sites can see up to a 115% visibility increase just from adding citations. Layer on: original, dated statistics (not aggregated summaries of others' research); named-author expert attribution and quotes with titles (+25-30%); a visible "last updated" date refreshed at least quarterly for competitive topics; and E-E-A-T signals — first-hand experience, specific rather than generic detail, and transparent sourcing.

**3. Presence — be where AI looks.** AI systems cite third-party mentions more heavily than a brand's own domain: Wikipedia accounts for 7.8% of all ChatGPT citations, Reddit 1.8%, plus industry publications, review sites (G2, Capterra, TrustRadius for B2B SaaS), YouTube, and Quora. Keep a Wikipedia page accurate and current, participate authentically in Reddit communities, pursue industry roundups and guest posts, maintain updated review-site profiles, and produce YouTube how-to content for key queries.

**Machine-readable files for AI agents.** AI agents are increasingly evaluating and buying on a user's behalf — if pricing is locked behind a JS-rendered page or a "contact sales" wall, an agent will skip the product and recommend a competitor it can actually parse. Add a `/pricing.md` or `/pricing.txt` at the site root with plain-text tiers, prices, limits, and included features per plan (consistent units, specific thresholds, kept as current as the visible pricing page). Also consider an `/llms.txt` (see llmstxt.org) giving AI systems a quick overview of the product, audience, and links to key pages — the same principle as `robots.txt` for crawlers or `AGENTS.md` for agent capabilities.

**Schema markup.** Article/BlogPosting, HowTo, FAQPage, Product, ItemList (comparisons), Review/AggregateRating, and Organization schema each help AI systems parse and trust content; proper schema correlates with 30-40% higher AI visibility.

## Content Types That Get Cited Most

Comparison articles lead (~33% citation share), followed by definitive guides (~15%), original research/data (~12%), best-of/listicles (~10%), product pages (~10%), opinion/analysis (~10%), and how-to guides (~8%). Underperformers: generic unstructured blog posts, thin marketing-fluff product pages, gated content AI can't access, content without dates or author attribution, and PDF-only content.

## Monitoring AI Visibility

Track AI Overview presence, brand citation rate, share of AI voice versus named competitors, citation sentiment, and which specific pages get cited (via referral traffic from AI sources). Dedicated tools include Otterly AI, Peec AI, ZipTie, and LLMrefs. Without tooling, run a monthly manual check: push the top 20 priority queries through ChatGPT, Perplexity, and Google, record who's cited and on what page, and track month-over-month in a spreadsheet.

## AI SEO by Content Type

- **SaaS product pages** — clear description in the first paragraph, category comparison tables (not just vs. named competitors), specific metrics over adjectives ("processes 10,000 transactions/sec" beats "blazing fast"), visible pricing plus a `/pricing.md`, and an FAQ addressing common buyer questions.
- **Blog content** — one clear target query per post matched to the heading, first-paragraph definition, original data or expert quotes, a visible "last updated" date, an author bio with relevant credentials, and internal links to product pages.
- **Comparison/alternative pages** — structured tables over prose, genuinely balanced framing (AI penalizes obviously biased comparisons), scored criteria, and current pricing/feature data.
- **Documentation/help content** — numbered step-by-step format, code examples where relevant, HowTo schema, descriptive alt text on screenshots, and clear prerequisites/expected outcomes.

## Common Mistakes

Ignoring AI search entirely (~45% of Google searches now show AI Overviews); treating AI SEO as separate from SEO rather than structure/authority layered on top of the foundation; writing for algorithms instead of humans (content that reads like it's gaming a system won't convert or get cited); missing freshness signals; gating the most authoritative content; ignoring third-party presence (a Wikipedia mention may out-cite the brand's own blog); no structured data; keyword stuffing (actively costs 10% visibility per the Princeton GEO study, unlike traditional SEO where it's merely ineffective); hiding pricing behind sales walls or JS rendering that agents can't parse; blocking GPTBot/PerplexityBot/ClaudeBot in robots.txt; generic claims without data; and never monitoring visibility at all.

## Task-Specific Questions
1. Top 10-20 most important queries?
2. Checked whether AI answers already exist for those queries?
3. Existing schema markup on the site?
4. What content types are published?
5. Are competitors being cited where the user isn't?
6. Wikipedia page or review-site presence?

## Related Skills
seo-audit (traditional technical/on-page SEO), schema (structured-data implementation), content-strategy (planning what to create), competitors (citable comparison pages), programmatic-seo (SEO at scale), copywriting (writing that's both human-readable and AI-extractable).