Answer engines — ChatGPT, Perplexity, Google AI Overviews — write the answer themselves and choose their own sources. By Google’s own guidance, AI Overviews and AI Mode rest on the core ranking and quality systems of search: a source first has to be a page that is crawled, indexed and trusted. That is why visibility is not built with a single file or a single markup; crawlable technical structure, a clear entity definition (schema.org), sentences that stay true when lifted out of context and verifiable authority work together. On this site llms.txt is an experimental map for other readers; in Google Search it is not a signal. Generative engine optimisation means building that foundation together; we also measure separately which agent arrived on which day.
A GEO consultant builds and measures how answer engines read your site. The work has three layers: a machine-readable base (structured data, entity consistency), a citable content structure, and real visibility measurement. It differs from classic consulting in its target: the aim is for the site to be named as a source in answers.
The two share a foundation but chase different things. Technical SEO solves crawlability and ranking; generative engine optimization solves whether a model lifts your sentence into its own answer. The measured distinction sits in the AEO, GEO and SEO difference guide; in practice a site wants both, and we build both in one layer.
Answer engine optimization means building the page to answer a question: a first sentence that is correct without context, real question headings, an answer that is not hidden. Citable passage architecture starts here. We write this site’s own service pages by the same rule; the answers to the frequently asked questions below are in plain view on the page.