AI search optimisation · GEO · AIO · LLMO
Your customers are asking ChatGPT. Be the answer.
Search is moving from links to answers. AI Overviews, ChatGPT and Perplexity answer the question and name a handful of sources, and most brands are not among them. I write and restructure content so that they are, using methods I test in public every fortnight.
What is AI search optimisation?
AI search optimisation, which you'll also see called generative engine optimisation (GEO), AI optimisation (AIO) or LLM optimisation (LLMO), is the work of making your content the source an AI answer engine cites. ChatGPT, Perplexity, Claude, Gemini and Google's AI Overviews all answer the question directly and name where the answer came from. Being named is the new ranking.
Why is the click disappearing while the citation isn't?
When an AI Overview cites five or six sources, those sources inherit the trust. Everyone else is invisible, however well they rank on page one. The rules for being cited overlap with SEO but aren't the same. AI systems reward content they can extract, attribute and quote: specific, sourced, well-structured statements. Vague claims don't get cited. "Fermentation improves digestibility" gets ignored; the sentence with the percentage and the source gets lifted.
What testing has shown
Perplexity's ranking has been partly reverse-engineered: Google indexing is a prerequisite, there are curated trusted-domain lists, a quality re-ranking layer, and an early-engagement window after publishing that new content has to hit. Google's own guidance for AI-friendly content asks for multimodal pages and structured data. And researchers have shown exactly how it can be gamed, with hidden instructions and command injection. I've tested the tricks, so I know which ones are worth your time and which ones get a domain burned.
Getting cited by AI is the new link building, but AI systems need extractable, attributable information they can quote and reference.
AI Marketing School, "The great unclick", June 2025
Plan explosive early distribution within hours of publishing.
AI Marketing School, "Gaming AI search like it's 1999", August 2025
Understanding this basic principle might help you get great results with LLMO.
AI Marketing School, August 2025
- IssueThe great unclick: zero-click search and the citation strategyJun 2025
- IssueGaming AI search like it's 1999: how Perplexity ranksAug 2025
- IssueLook who's late to the party: Apple's AI searchSep 2025
The offer
What the work involves
A writer's approach rather than a plugin. Most of the gain is in the content itself: what it says, how it is structured and how it is launched.
AI visibility audit
I ask ChatGPT, Perplexity, Claude, Gemini and Google AI Overviews the questions your buyers ask, log who gets cited and why, and show you the gap between you and them. A short report, specific changes, no deck.
Citable content
New pages and rewrites built from specific, attributable claims, clear definitions, named sources and figures. The kind of sentence an AI can lift and credit.
Structure and markup
Headings that match the questions people ask, answer-first paragraphs, FAQ and structured data so AI systems can parse the page without guessing.
Authority and distribution
AI search leans on trusted domains and early engagement. I plan the launch and the third-party placements that get a new piece noticed in its first hours, not its first months.
Brand in the corpus
Getting your name, your definitions and your point of view onto the sites and formats that language models read, so you show up even when you aren't the top result.
Monthly tracking
The audit prompts, re-run every month, so you can watch citations appear and know which pages did it.
Straight talk
Ethical growth hacking
I know how the systems are gamed. I have written up the ranking layers, the hidden-instruction and command-injection methods, and what happens when they are patched. So I use the aggressive, legitimate levers fully: launch windows, early distribution, exact-match titles on emerging queries, placements on the domains the models trust, and structure the parsers handle well. What I avoid is anything that gets a domain flagged the week after it works.
It is not instant. Expect first citations within weeks for low-competition questions and a few months for crowded ones. I will say which is which before we start, and where a grey-hat method is worth testing.
It is not separate from good writing. The qualities that keep a human reader to the end, specificity, structure and a clear position, are the ones that get a piece quoted by a machine.
Find out who AI cites instead of you
Send me your site and the five questions your customers ask most. I will run them through the major AI search tools and send a short, honest assessment of where you stand.
Request an audit →Questions people ask
Questions about AI search
What is AI search optimisation?
AI search optimisation, also called generative engine optimisation (GEO), AI optimisation (AIO) or LLM optimisation (LLMO), is the practice of making your content the source that AI answer engines such as ChatGPT, Perplexity, Claude, Gemini and Google AI Overviews cite when they answer a question. It overlaps with SEO but rewards extractable, attributable, well-sourced statements, early distribution and presence on trusted domains.
How is GEO different from SEO?
SEO gets a page ranked in a list of links. GEO gets a page quoted inside an answer. AI engines cite several sources per answer, favour specific sourced claims over vague ones, lean on trusted domains, and respond to early engagement after publishing. Good SEO content is a strong starting point, but it usually needs restructuring to be cited.
How long does it take to get cited by AI search?
Weeks for low-competition questions on a site with some authority; a few months for crowded ones. I audit first and say which is which before any work starts.
Can AI search be gamed?
Yes, and I've written up how: Perplexity's trusted-domain lists and launch-window thresholds, hidden-instruction and command-injection tricks, the lot. I use the levers that work and last, and I'll test grey-hat ones with you where the upside is worth it. What I avoid is anything that gets a domain flagged the week after it's patched. Ethical growth hacking, not a sermon.
What does an AI visibility audit include?
I run the questions your buyers ask through ChatGPT, Perplexity, Claude, Gemini and Google AI Overviews, log which sources are cited and why, compare them with your pages, and deliver a short report with the specific changes and new content that would close the gap.