Generative Engine Optimization (GEO) is the practice of shaping content so large-language-model answer engines quote it verbatim inside their generated responses — a sibling discipline to AEO that focuses on the phrasing, structure, and evidence patterns that models prefer to lift.
Practitioners usually treat GEO and AEO as overlapping labels for the same job: get named inside AI answers. Where they differ in emphasis, GEO leans toward the content craft — self-contained sentences, explicit attributions, quantified claims — while AEO leans toward the machine-readability stack (schema, canonical, entity graph).
The tactics are grounded in what the models actually do at inference time. Grounded assistants like ChatGPT with browsing, Perplexity, and Google AI Overviews retrieve candidate passages and then generate an answer that cites them; Perplexity's public documentation and Google's AI-features guidance both describe grounding on retrieved web content. Passages that stand alone — a full sentence stating the fact, not a bullet fragment — retrieve and quote cleanly.
Concrete GEO moves: put the direct answer as the literal first sentence under each question heading, attribute claims by name ("Google's structured-data documentation states…") rather than passive voice, and include one date, number, or specification per paragraph so the model has something concrete to lift.
GEO is not a replacement for authority. The models still prefer to cite sources with strong entity signals and real expertise; the phrasing tricks only help once the source is trusted.