To get cited by ChatGPT, your website needs three things: content structured around the specific questions people ask, clear factual claims that AI models can lift and attribute without misreading them, and enough independent signals elsewhere on the web to convince the model your business is a trustworthy source on the topic. Get those three right and citation follows. Get them wrong and you can have the best product in your market and still never appear in an answer.
That’s the short version. Most businesses stop there and wonder why nothing changes. The reason is that “structure your content properly” isn’t a task, it’s a framework, and without one you’re guessing at what a language model actually rewards.
Why This Is Different From SEO
Traditional SEO optimises for a ranking algorithm working through keywords, backlinks and page authority. ChatGPT and other AI engines don’t rank pages, they synthesise answers. That means the unit of value shifts from “does this page rank” to “does this page contain a clean, extractable fact that answers a specific question.” A page can rank on Google and be invisible to ChatGPT. A page can also barely rank on Google and still get pulled into an AI answer, because the model found one paragraph that answered the query precisely.
This is the gap most agencies haven’t caught up with. They’re applying 2015 SEO thinking to a 2026 problem.
The OPTIMUM Framework
OPTIMUM is the audit methodology I built at State of the Art Digital to diagnose exactly where a business stands on AI visibility, and where the opportunity sits. It scores a site across seven dimensions, each rated on where it is now and where it could realistically get to, with the gap between the two showing you which dimension is worth fixing first.
The dimensions cover things most businesses have never audited before: whether your factual claims are stated in a form a model can extract cleanly, whether the right entities (your business, your people, your credentials) are unambiguously identified across the web, whether your content answers real user intent or just targets a keyword, and whether independent third parties are corroborating what you say about yourself. A business can have a beautifully designed website and score badly on every one of these, because good design and AI-readiness are unrelated problems.
The value of scoring it this way is prioritisation. Nobody has the budget to fix everything at once. OPTIMUM tells you which dimension has the biggest gap between where you are and where you could be, so the work you commission actually moves the number that matters.
Fan-Out Query Mapping
Here’s the part that catches almost everyone out. When someone asks ChatGPT a question, the model rarely answers from a single search. It breaks the question into several related sub-queries, runs each one, and blends what it finds into one response. This is called query fan-out, and it’s the reason a business can rank well for its main keyword and still be absent from the AI answer entirely.
Take “best accountant near me.” A model won’t just search that phrase. It will fan out into sub-queries covering qualifications, pricing structure, specialisms, client reviews, and response times, then decide which businesses show up consistently across those separate searches. If your website only ever addresses the headline query and says nothing concrete about pricing or specialism, you vanish the moment the model fans out into those directions, even though your homepage was built for exactly the customer asking.
Fan-Out Query Mapping is the process of working out what the real sub-queries are for a given topic and making sure your content answers each one directly, in language a model can extract without ambiguity. It’s not about guessing keywords. It’s about thinking the way the model thinks: breaking one question into the five or six it will actually go and check.
Where to Start
If you want to know exactly where your business stands, an OPTIMUM audit gives you a Now Score, a Future Score and a clear list of what closes the gap between them, dimension by dimension. That’s the honest starting point. Everything else is guesswork dressed up as strategy.
Steve Coulter, a U.K. based AI Search Consultant and creator of OPTIMUM which identifies the hidden SEO, GEO and entity issues preventing companies from appearing prominently in AI-driven discovery. Local search and competitive industry expert.