To rank in ChatGPT, your business needs to be clearly and consistently described across the web, backed by third-party corroboration, and structured so AI engines can read you without ambiguity. It is not about keywords or backlinks in the traditional sense. It is about giving an engine enough unambiguous information about your business that it can recommend you without hedging. The businesses that appear in AI answers have built that picture. The ones that do not appear have not.
Here is the practical version: what to do, in what order, and why the things most people try first do not work.
Why most tactics do not get you into ChatGPT answers
ChatGPT does not work like Google. Google ranks web pages by matching keywords and measuring authority signals. You optimise a page and it climbs the list. The cause-and-effect is fairly direct.
ChatGPT, Claude and Perplexity do something different. They do not return a ranked list of pages. They read across many sources, build a picture of your business, and return a short shortlist of named businesses as an answer. The question is not “which page matches these words?” It is “which business should I recommend, and am I confident enough to name them?”
That second question is where most tactics fail. Tactics like adding more keywords to your homepage, updating your meta descriptions, or publishing three new blog posts do not address whether an engine is confident about your business. They address visibility to Google’s crawler. Those are related but not the same thing.
When we measure why one business gets named and a near-identical competitor does not, the same four signals come up every time: entity clarity (does the web agree on who you are), consistency across sources (are you described the same way in multiple places), third-party corroboration (do other credible sources mention you), and structured, readable content an engine can extract information from. None of those reduce to keyword optimisation.
This matters because your customers are already asking. They type something like “who are the best [your service] in [your area]” and three to five businesses come back named, silently, by an engine that has already formed its view of who is worth recommending. You never see the question asked. You see only its consequence: the enquiries that did not arrive.
The fix starts with knowing where you stand.
Step 1: Measure your baseline citation rate first
This is not optional, and it is not just a nice-to-have before starting. Measuring first is the whole methodology. Without a baseline, you have no way to know which changes are working, which are not, and where your effort is best spent.
Here is how to do it yourself. Write down the ten to twelve questions your customers would actually type into an AI engine. Not abstract queries like “ai search uk services” but the specific, natural-language questions a person in your target market asks when they want what you sell: “who are the best [your service] in [your town]”, “I need a [your service], who do you recommend”, “which [your service] companies are worth using”.
Run each question through ChatGPT, Claude, Perplexity and Google AI Overviews. Record: whether your business is named, where it appears in the answer, and what the engine says about you. Do this for your top three competitors as well.
The result is your citation rate: the percentage of relevant questions where you are named. Most businesses that have never done this work score in the low single figures or zero. Your competitors may score higher. The gap between the two is the gap the rest of these steps close.
Our guide on how to measure AI visibility walks through a more structured version of this audit. Alternatively, a Visibility Briefing, our entry-point engagement, does this measurement across all four engines and hands you a prioritised gap analysis without you having to run it yourself.
Step 2: Make your entity information unambiguous
“Entity information” sounds technical. The practical version is straightforward: does the web clearly agree on what your business is, what it does, and where it operates?
Answer engines are cautious. They will not recommend a business they are unsure about. If your website describes your services in one set of words, your Google Business Profile uses a different set, your directory listings name your business differently, and your social profiles have an outdated address, the engine reads a contradictory picture. Contradiction breeds hesitation. The engine names the competitor whose picture is unambiguous.
The fixes here are mostly unglamorous:
- Audit how your business is described across your website, directory listings, and profiles. Name, address, phone number and service description should be consistent everywhere.
- Write a clear, factual “about this business” paragraph that describes what you do in plain, specific terms, and use it as the source version that all other descriptions derive from.
- Add structured data (Schema.org markup) to your website so machines can extract your business type, location and services without guessing.
- Fix any contradictions you find. An outdated address on a directory listing is not a minor housekeeping issue in this context; it is a consistency break that signals uncertainty to an engine.
This is not the exciting part of the work. It is the foundation. Every step after this one depends on it.
Step 3: Build a well-formed llms.txt file
An llms.txt file is a structured, machine-readable document that tells AI crawlers exactly who you are and what you do. Think of it as a covering letter for your business, written for an engine rather than a person.
A well-formed llms.txt file names your business clearly, describes your core services without ambiguity, and points the crawler to the most important pages on your site. It removes one of the most common barriers to being understood correctly: the engine misreading your homepage copy because it is written for humans, not machines.
This is a high-leverage technical step. It does not replace the broader work of consistency and corroboration, but it removes a consistent source of misreading at the point where the engine first encounters your site.
Step 4: Earn third-party corroboration
An engine building its picture of your business is not just reading your own website. It is reading what the rest of the web says about you. Reviews, directory listings, published articles, press mentions, podcast appearances, association memberships, industry roundups that name you. Any place where another credible source describes your business.
“Credible” does not mean only major publications. A well-regarded industry directory, a local business association, a trade body membership, a review platform relevant to your sector, a customer case study published on a third-party site: all of these add corroboration. The engine trusts a business the wider web agrees on.
This is where the AI visibility work overlaps with what used to be called digital PR. The difference is the frame: you are not trying to earn links for a ranking algorithm. You are trying to build a body of external evidence that an AI engine can read and say, with confidence: yes, this business is what it claims to be.
Practical starting points: ensure you are listed and accurately described on the main directories for your sector, actively collect reviews on the platforms that matter for your audience, and look for opportunities where a third party might write about or mention your business in a way that adds to the corroborated picture.
Step 5: Publish content that adds new information
Not all content helps. Content that adds new information about your business, your specific expertise, and your point of view does help. Generic content that uses the same phrases as every other business in your category adds almost nothing.
The content that tends to work: case studies that describe specific problems you solved and how, articles that explain your methodology or process in specific terms, expert commentary on topics directly relevant to your service area, Q and A formats that match the natural-language questions people ask AI engines.
Format matters too. Answer Engine Optimisation (AEO), the practice of making your business visible to AI search engines, rewards content that is structured clearly, written with authority, and cites specific evidence. A clear question-and-answer structure, concrete numbers, and referenced sources are all signals an engine weights positively when deciding whether to quote a piece of content.
The how to get cited by ChatGPT guide goes deeper on content formatting for AI citation. The short version: write like an expert explaining something real, structure it so an engine can extract the answer, and make sure it adds information about your business that was not there before.
What does not work (and wastes your time)
Being specific here is useful, because the tactics that do not work are often the ones most heavily marketed.
Keyword-stuffing your homepage does not cause an engine to recommend you. The engine is not counting keyword occurrences. It is assessing whether the whole picture of your business is clear and trustworthy.
Publishing content at high volume without substance does not help. Thin content that restates what is already on your site or repeats sector-level generalities does not add corroboration. It adds noise.
Paying an agency for monthly “AI optimisation reports” that track rankings on Google but do not measure your citation rate across actual AI engines is not AI visibility work. It is repackaged SEO with a new name.
Waiting for a tool to do it automatically is not a plan. The engines change their retrieval and synthesis methods regularly. The methodology that compounds across quarters is the one built around understanding what the engines actually reward, measuring that, and adjusting. Tools help with measurement. They do not replace structured thinking about what to measure.
The order matters: measure, fix, re-measure
The sequence is not incidental. Measuring first is the principle at the core of how Qyliq works: show the evidence before recommending a course of action, not the other way around.
A business that skips measurement and jumps straight to “fixing” may spend time on the wrong thing. A business that measures first knows exactly which questions it is not appearing on, which competitors are appearing instead, and which of the four signals (entity clarity, consistency, corroboration, content quality) is the weakest link. The fix is then targeted, not a scatter of activity.
Re-measuring after the engines have had time to re-index closes the loop. It tells you whether the work moved the needle, by how much, and what to do next. That quarterly cycle is what produces a documented, improving citation rate rather than a one-off effort that may or may not have worked.
This sequence produces measurable results when it is followed in order. Our work with a dance and fitness business ran exactly these steps, and the ChatGPT citation rate went from zero to 92 per cent on the buyer questions that matter, with 60 per cent across all four major engines combined. The full methodology behind that result is set out on our methodology page.
If you want to understand where your business stands before committing to any of the work above, that is exactly what a Visibility Briefing is for. It measures your current citation rate across ChatGPT, Claude, Perplexity and Google AI Overviews, identifies the gap, and tells you what to prioritise. Diagnostic, not a sales call.
Request a Visibility Briefing and we will show you exactly where you stand.