How to Get Cited by ChatGPT, Perplexity, and Google AI Overviews

SEO is not dead. But it is no longer sufficient. In 2026, a growing percentage of search queries are answered by AI-generated summaries that cite sources — and the rules for getting cited are fundamentally different from the rules for ranking.

Welcome to GEO: Generative Engine Optimization. The practice of optimizing your content to be cited by AI systems — ChatGPT, Perplexity, Google AI Overviews, Claude, and the growing ecosystem of AI search tools that generate answers instead of listing links.

After implementing GEO strategies across 52 websites over the past three months, here is what actually works for getting cited — broken down by platform.

How AI Citations Actually Work

Traditional search engines rank pages. AI search engines synthesize answers from multiple sources, then cite the sources they drew from. The critical difference: you are not competing for a position on a list. You are competing to be included in a generated answer.

AI citation decisions are based on several factors:

• Authority signals — E-E-A-T indicators, domain authority, author credentials
• Content clarity — concise, quotable statements that can be extracted without losing meaning
• Data specificity — specific numbers, dates, and facts that add precision to the AI's answer
• Source diversity — AI systems prefer citing multiple sources, so niche expertise can win over broad authority
• Structured data — schema markup, especially SpeakableSpecification, that flags quotable content
• Freshness — recent content with current data points gets cited over outdated content

Platform-by-Platform Strategy

Google AI Overviews

Google's AI Overviews pull from the same index as regular search, but the citation logic is different. Overviews favor content that provides a direct, concise answer to the query — not content that ranks well for traditional SEO.

What works for Google AI Overviews:

• Answer the question in the first 2-3 sentences of your content
• Use the exact phrasing of common queries as subheadings
• Include specific data points — Google Overviews cite statistics at a disproportionate rate
• Add SpeakableSpecification markup to flag your most quotable paragraphs
• Publish FAQ schema for common questions in your niche

What does not work: Keyword-stuffed content, thin articles that circle a topic without answering it, pages that bury the answer below the fold.

ChatGPT (Browse Mode)

When ChatGPT browses the web to answer a question, it visits pages, reads the content, and decides what to cite. Its citation behavior differs from Google's in several ways:

What works for ChatGPT:

• Comprehensive, well-structured content that covers a topic thoroughly
• Clear headings that match common question patterns
• An llms.txt file at your site root that briefs the model on your expertise
• An agent-card.json in your .well-known directory
• Long-form content (1,500+ words) with clear section breaks — ChatGPT prefers depth

What does not work: Content behind paywalls (ChatGPT cannot access it), pages that require JavaScript rendering to display content, sites that block GPTBot in robots.txt.

Perplexity

Perplexity is the most citation-heavy AI search engine. It cites sources aggressively and links directly to them. This makes it the highest-value platform for traffic from AI citations.

What works for Perplexity:

• Original research and data that cannot be found elsewhere
• Clear attribution of your own claims — Perplexity trusts sources that cite their data sources
• Recent publication dates — Perplexity strongly favors fresh content
• Topical authority — publishing multiple related articles on the same subject cluster
• Structured data and clean HTML — Perplexity's crawler prefers well-structured pages

What does not work: Aggregated content that summarizes other sources without adding original analysis, content without dates, pages with heavy ad loads that obscure the content.

Claude

Claude's approach to web browsing and citation is more conservative. It tends to cite fewer sources but with higher confidence.

What works for Claude:

• Authoritative, well-sourced content with inline citations
• Content that demonstrates genuine expertise through specific examples and case studies
• Clear author attribution with verifiable credentials
• Sites that allow Anthropic's crawler in robots.txt

The Content Format That Gets Cited Everywhere

After testing across all platforms, one content format consistently outperforms all others for AI citations: the definitive guide with embedded data.

The structure:

1. Opening statement (2-3 sentences) — directly answers the core question with a specific claim or data point
2. Context section — explains why this matters with relevant trends or statistics
3. Detailed breakdown — covers the topic with specifics, examples, and actionable steps
4. Data table or comparison — structured data that AI can extract and reference
5. Conclusion with a clear takeaway — a single quotable statement that summarizes the article

This format works because AI systems are looking for two things simultaneously: a concise answer to cite and a comprehensive source to validate that answer. The opening gives them the citation. The body gives them the confidence.

Practical GEO Upgrades for Your Next Audit

Here is the checklist we used across our 52-site network:

Structural upgrades:

• Add SpeakableSpecification schema to all blog post templates
• Deploy llms.txt at the site root with topic authority declarations
• Deploy agent-card.json with expertise and capability metadata
• Ensure robots.txt allows GPTBot, Google-Extended, and anthropic-ai crawlers

Content upgrades:

• Audit your top 20 posts — does each one answer the core question in the first two sentences?
• Add specific data points to every article — percentages, dollar amounts, timeframes
• Convert opinion statements to evidence-based claims with sources
• Add an "article summary" paragraph with the .speakable class to each post

Publishing strategy:

• Publish at least one piece of original research per month — data nobody else has
• Update evergreen content quarterly with current data points
• Build topical clusters — 5-10 related articles on the same subject area
• Cross-reference your own content with internal links (AI agents follow internal links to assess topical depth)

The Compounding Effect

GEO is a compounding strategy. Each article with proper GEO optimization increases your site's overall authority signal. AI systems that cite you once are more likely to cite you again — they build internal trust profiles for domains that consistently provide accurate, well-structured information.

We started seeing meaningful AI citation traffic by the third month of implementation. By month six, AI-referred visits accounted for 12-18% of total traffic across the network — traffic that did not exist before because we were not optimized for these platforms.

The websites that invest in GEO now will compound that advantage. The ones that wait will spend years catching up.

This strategy is covered in more depth in The W-2 Trap — which demonstrates how to build authoritative content that both humans and AI systems recognize as expert-level. Buy The W-2 Trap on Amazon.