Generative Engine Optimization (GEO) is content optimization tailored for "generative engines." Traditional search engines return a list of links for users to click through, whereas generative engines (such as ChatGPT, Perplexity, and Gemini) retrieve multiple sources and use large language models to synthesize them into a single answer, thereby causing content creators to lose control over "when and how their content is displayed."

The concept of GEO was first proposed in 2023 by Aggarwal et al. in the paper GEO: Generative Engine Optimization. They defined visibility as an optimizable metric, introduced a black-box optimization framework that allows creators to adjust their content without understanding the engine's internal mechanisms, and released the GEO-bench benchmark.

A core conclusion of GEO is that "strategies vary by domain." Experiments show that appropriate optimization strategies can increase content visibility in generative answers by up to 40%, and there is no one-size-fits-all template. Subsequent research has further integrated terms like AEO and AI Visibility into the same lineage.

The significance of GEO lies in rewriting the "ranking mindset" of the SEO era into a "citation and adoption mindset." While SEO competes for ranking positions, GEO competes for being cited and adopted. For the creator economy, this relates to the redistribution of traffic and revenue; for engine developers, it raises new issues regarding citation transparency and source fairness.