Four acronyms are circulating for roughly the same thing. Before hiring anyone who uses them, it helps to know what each means.
The definitions, without the smoke
- AEO (Answer Engine Optimization): optimising to be the answer, wherever it appears: an assistant, an AI overview or a search engine with a direct answer.
- GEO (Generative Engine Optimization): the same goal, narrowed to generative engines that compose an answer from several sources.
- LLMO: optimising for the language model itself, including its trained knowledge, not only what it retrieves live.
- AI SEO: an umbrella term, often commercial, mixing all of the above with classic SEO.
In practice the work is 80% identical: a coherent entity, answer-shaped content, presence in third-party sources and systematic measurement.
Where a real difference exists
The useful distinction is not the acronym but where the answer comes from:
- Retrieved answers: the model searches the web and cites sources. Here you win with published, updated, crawlable content.
- Memorised answers: the model replies from training, without searching. Here you win with broad, consistent presence over a long time: mentions, press, directories, forums.
The first moves in weeks. The second, in quarters.
Which metric to demand from your team or agency
Whatever acronym the proposal uses, demand these four numbers:
- Mention rate over a fixed question set.
- Average position within the answer.
- Share of voice against three competitors.
- Cited sources, and which ones you are absent from.
If a proposal has no baseline and no repeated measurement, it is not a strategy: it is a wish list.
Practical conclusion
Use AEO as the working term, ignore the acronym war and focus on the only thing you can prove: whether your brand appears more often this quarter than last on the questions that matter to your business.