GEO
GEO: generative engine optimisation
Generative engine optimisation is the work of shaping how a business is described when an AI model generates an answer about it. Where answer engine optimisation asks whether a business is cited at all, GEO asks whether the description that follows is accurate. Both are decided by the same set of sources.
What the model is doing when it describes you
A model does not hold a profile of a business. When asked, it retrieves sources, reconciles them, and writes a description in its own words. Every adjective in that description came from somewhere.
This is why two businesses with similar offerings can be described very differently. The difference is rarely the offering. It is which sources exist, what they emphasise, and whether they agree.
Where the description comes from
Six kinds of source, which together make up what we call an AI fingerprint: the website, the content written to answer customer questions, the structured data that states the facts in machine-readable form, the data and facts themselves, the statements the models already make about the business, and the earned media that corroborates all of it.
Three of those are owned outright and can be corrected immediately. Two are signals a business can shape but not control. One is earned.
What you can and cannot influence
You cannot edit an AI answer. You can edit what it is assembled from, which makes this a question of sources rather than persuasion.
In most categories a small number of domains account for most citations, so the useful question is narrow: which sources do the models rely on here, and is the business present on each with the same facts?
Measuring a description, not just a mention
Counting mentions misses the point of GEO. A business can be named often and described badly.
So alongside visibility, share of voice and position, we read every claim a model makes about a business against its verified facts, and trace each claim to the source it came from. That is what makes a wrong description fixable rather than merely annoying.
Frequently asked questions
Is GEO different from AEO? They overlap heavily and we run them as one practice. AEO is usually about citation, GEO about description. The same six sources decide both.
Does structured data improve how AI describes a business? It helps machines read the facts unambiguously, and we use it for that. It is not a shortcut to being described well: a controlled test by Ahrefs on 1,885 pages that added structured data found no gain in AI Overview visibility. Consistency across sources does far more.
How quickly can a wrong description be corrected? It depends where the claim came from. A fact on a page the business controls can be corrected immediately and usually propagates within weeks. A claim repeated across third-party listings takes longer, because each source has to be corrected or outweighed.