Generative Engine Optimisation (GEO). The new page one.

Generative Engine Optimisation (GEO) is the process of making your website's information easily retrievable, clear, and consistent for large language models (LLMs) like ChatGPT, Claude, and Perplexity. Unlike traditional SEO, which focuses on ranking links on a search page, GEO focuses on providing structured, accurate evidence so that generative AI systems confidently cite and recommend your brand when synthesising answers to user queries.

Each AI platform has unique retrieval behaviours and preferences:

• Perplexity — Prioritises real-time web retrieval, heavily favouring extreme content freshness (where citation likelihood drops after 2–3 days) and authentic community discussions from sites like Reddit.
• Claude — Favours comprehensive, in-depth content (often 3,000+ words) and places a high value on verifiable author credentials and industry databases.
• ChatGPT — Relies heavily on its underlying training data, but also uses "query fan-out" to synthesise web results, often rewarding content that front-loads its key claims.

Partially. Because both platforms utilise web search retrieval, content that ranks well on traditional search engines will generally benefit on both. However, their underlying algorithms differ significantly — Perplexity uses a complex reranking model across dozens of citation factors, while ChatGPT relies more heavily on its training data and Bing ranking signals.

An "LLM Info" page acts as a raw, structured data feed (often a markdown file) hosted on your site that explicitly lists your company's mission, pricing, features, ideal customer profile (ICP), integrations, and case studies. By giving LLMs a single, centralised source of truth, they are more likely to cite your verified facts instead of relying on outdated or scattered fragments from across the web.

To fix hallucinations or outdated facts, you must trace the misinformation back to its source — usually an old blog post, a stale review, or mismatched website details. Update, delete, or redirect this old content and replace it with clean, structured facts. Once the underlying evidence across the web is consistent, LLMs will adapt to the new narrative.

LLMs do not just look at your website; they synthesise public perception by pulling from community platforms and review sites. Reviews give AI models a sense of your brand's recent performance and reliability. Reddit is heavily cited (accounting for nearly 47% of Perplexity's citations) because it offers unfiltered user experiences. If old Reddit threads contain complaints you have since fixed, LLMs might still cite them unless you transparently respond with factual updates and evidence.

For Perplexity, content freshness is critical — content decay can begin within just 2–3 days of publication. We recommend auditing and updating your most important pages at least monthly by adding new data points or insights, and using highly visible "Last Updated" timestamps to signal freshness to the AI crawlers.

YouTube is a massive driver of AI visibility and is the most-cited domain in Google's AI Overviews. ChatGPT has also doubled its YouTube citations quarter-over-quarter. Because modern LLMs are multimodal, they actively pull videos into their answers — meaning that optimising your video titles, descriptions, and transcripts for target queries can significantly boost your visibility across multiple AI platforms.