New term: GEO

2025-05-28 · Publisher : Andreessen Horowitz · Domain : Artificial Intelligence

GEO (generative engine optimization) aims to get a brand or content cited in AI assistants' answers, the way SEO aims for ranking in search engines.

### Definition **GEO** (*generative engine optimization*) covers practices aimed at increasing a piece of content's visibility in the answers of generative engines: ChatGPT, Perplexity, Gemini or the AI answers of search engines. It is also called AEO (*answer engine optimization*). ### Origin The term was born in research: the paper "GEO: Generative Engine Optimization", by Pranjal Aggarwal and researchers including a Princeton team, was posted on arXiv on November 16, 2023 and presented at the KDD conference in August 2024. Its move into marketing came later. The date chosen here is May 28, 2025, when Zach Cohen and Seema Amble (Andreessen Horowitz) published the widely shared "How Generative Engine Optimization (GEO) Rewrites the Rules of Search": "Traditional search was built on links. GEO is built on language." ### What the term covers - In the Princeton study, citing sources and adding statistics and quotations raised content visibility in answers by up to 40%, while keyword stuffing did not help. - For a16z, the reference rate in answers replaces click-through rate as the visibility metric. - Tools such as Profound, Goodie or Daydream track brand mentions in model answers. ### Criticism and limits The original study relies on a specific benchmark, not on everything now filed under the acronym. GEO also opens the door to manipulation of generative engines, which researchers are starting to study. ### Key takeaway Being found is no longer enough: content must be picked up and cited by models, which rewards reliable, sourced and well-structured material.

Tags : Vocabulaire, Recherche, Marketing, Web