Crawlability, checked properly
Including the things that are invisible locally. Hosts inject directives into robots.txt that are not in your file.
AI search ยท GEO and AEO
Getting found when the answer is written for you
People increasingly ask an assistant instead of searching. Being the source it quotes is a different job to ranking.
The short answer
Generative engine optimisation is the work of being the source an AI assistant cites when it answers a question in your category. It overlaps with SEO but is not the same job: ranking gets you into a list, citation gets you into the answer.
Three things decide it. Whether the machines can read you at all, whether your pages answer the question in the first paragraph rather than the fifth, and whether you publish anything distinctive enough to be worth quoting. The third is the one most businesses skip, and it is the one that actually works.
Answer-first rewriting
The answer in the opening forty words, because that is the chunk that gets lifted.
Structured data
FAQ, Service, Article, Organization and the entity links that let a knowledge graph know who you are.
Entity consistency
Same name, same address, same identity across every profile that mentions you.
Something worth citing
Usually original data. A number nobody else publishes beats a thousand words of commentary.
llms.txt and the plumbing
Cheap, emerging, and worth doing while it still differentiates.
Proof
We did this to our own site first.
Our robots.txt was being served with a crawl delay injected by the host that did not exist in the file itself, throttling every crawler that honours it to one request every ten seconds. Our sitemap claimed every page changed on the same date. Our Organization schema had no external identity links at all.
All three were invisible without checking what the server actually serves, as opposed to what the files say. That is most of this work: the gap between what you published and what a machine receives.
FAQ
Common questions
Is this just SEO with a new name?
No, though they overlap. SEO is about ranking a page. This is about being the source an answer engine quotes when it writes the answer itself, which depends more on being clearly structured, genuinely distinctive and citable than on where you sit in a list of ten blue links.
Does blocking AI crawlers protect my content?
It stops you being cited. There is a real argument for blocking training crawlers while allowing search ones, and the distinction matters: OAI-SearchBot and ClaudeBot fetch to answer a question now, GPTBot and CCBot may also feed training. We will set it the way you want, but blocking everything means never being the answer.
How do you measure it?
Honestly, and that means admitting the measurement is worse than for search. We track whether the assistants actually cite you for your core questions, referral traffic from AI surfaces where it is identifiable, and the things we do control: structured data coverage, answer-first formatting, entity consistency and crawlability.
How long does it take?
Answer engines re-read the web constantly, so structural work can show up in weeks rather than the months search takes. Becoming the cited source for a competitive question takes longer, because it usually requires publishing something nobody else has.
Start a project
Want to know how you look to an answer engine?
We will check what the machines actually receive, not what your files say, and tell you what is worth fixing.
- A reply within one working day
- Fixed price, agreed in writing
- No obligation, and no hard sell