AEO focuses on the answer itself. The aim is to make the single best response to a question easy for a machine to lift cleanly: a concise lead sentence, a clear question-and-answer structure, and supporting detail underneath.
AEO predates the current wave of generative engines. It grew up around featured snippets and voice assistants, where one answer is read aloud or shown in a box. Those same habits (lead with the answer, use plain language, add structured data) now help large language models quote you accurately.
In practice AEO and GEO overlap heavily. AEO is the narrower, format-level discipline (write extractable answers); GEO is the broader program that also covers authority, entity signals, and cross-engine measurement.