GEO: Generative Engine Optimization - How to Rank in ChatGPT, Perplexity, and Claude

GEO vs Traditional SEO

Search Engine Optimization (SEO) has a 30-year-old playbook: earn backlinks, target keywords, optimize for crawlers, and compete for blue links. Generative Engine Optimization (GEO) targets a different output: being cited inside an AI-generated answer.

The difference is fundamental. In SEO, you appear in a list and users choose to click. In GEO, an AI system synthesizes your content and presents it as part of its answer - often with attribution, but sometimes without. Your content becomes an ingredient in someone else's output. Getting cited means your brand appears in front of the user; not getting cited means your competitor's content shapes the answer instead.

The Princeton GEO Paper

The foundational academic work on GEO comes from a Princeton-led research team (arXiv:2311.09735). They tested 6 content strategies against multiple generative engines and measured citation rates. Their findings:

1. Statistics and quantitative data - Adding specific numbers, percentages, and measurements to content significantly increases citation rates. AI systems prefer specific, verifiable facts over vague claims.
2. Quotation and authority signals - Quoting recognized experts or citing authoritative institutions increases likelihood of being included.
3. Fluency and readability - Well-structured, clearly written content outperforms dense or jargon-heavy writing. AI models prefer passages they can extract cleanly.
4. Authority markers - Explicit mentions of credentials, publication venues, institutional affiliations.
5. Easy-to-understand format - Summaries, definitions, and direct answers perform better than nuanced essays when the query calls for a clear response.
6. Citations to sources - Content that itself cites its sources tends to be more trusted by AI systems as a reliable source.

How Perplexity Chooses Sources

Perplexity's answer engine pulls from the Bing index (plus its own crawlers), weighted heavily toward:

• Freshness - Recently published or updated content ranks higher in Perplexity's source selection
• Authority - Domains with strong backlink profiles and brand recognition
• Bing index status - If you're not indexed in Bing, you won't appear in most Perplexity answers. Submit to Bing Webmaster Tools.
• Structured, scannable answers - Content that directly addresses the query in the first paragraph

The Perplexity Publishers Program offers revenue sharing to sites that are frequently cited - another reason getting into AI answers has real business value.

Getting Cited in ChatGPT

ChatGPT's browsing feature (when enabled) uses Bing search results. Standard SEO principles apply: rank on Bing, get cited. For ChatGPT's base model (no browsing), you're dealing with training data - content from high-authority domains that was included in the training corpus.

For the browsing case, the same Perplexity advice applies: Bing indexing, fresh content, authoritative domain, direct answers.

Getting Cited in Claude (Anthropic)

Claude's knowledge comes from training data and, when tools are enabled, real-time web search via Brave Search or other providers. The same principles apply for web-enabled Claude: strong domain authority, structured content, direct answers.

For base Claude (no tools), high-quality content on authoritative domains that was available before the training cutoff has the best chance of being part of the training distribution.

Monitoring AI Mentions

Track whether AI systems are citing your content using:

• Manual queries in Perplexity for your target topics
• Perplexity API (paid) to automate mention monitoring
• Google Alerts for your brand name appearing in AI-generated contexts

Links: GEO paper arXiv:2311.09735 | Perplexity Publishers Program