People now ask ChatGPT, Gemini, Perplexity and Google’s AI features which business to use, and those tools answer from what they can find and trust about you.
This work makes a business easier for them to understand, quote and recommend accurately. It is closer to good SEO than most of the market admits.
Most of what is sold as AI search optimisation is ordinary SEO with a new name on the invoice.
That is worth saying plainly, because the alternative is paying twice for the same work. The things that make a business visible in AI answers are the things that have always made it visible: a site that can be understood, pages that answer real questions, genuine expertise on the page and credible references from elsewhere.
What has genuinely changed is narrower than the noise suggests. These systems answer rather than list, so they quote a passage rather than send a click. They assemble an answer from several sources at once, so what other people say about you matters more. And they need to be confident about who you are before they will recommend you, which punishes businesses that describe themselves differently in every place they appear.
Those three changes are real, and they do need work. They are not a reason to start again.
AI tools answer by assembling what they can find and trust about a business across the web, not by reading one page in isolation.
That work starts with describing the business the same way everywhere: the website, structured data, profiles and listings. Inconsistent descriptions give these systems no clear answer to repeat.
AI systems extract the passage that answers the question rather than summarising a whole page.
Service and insight pages need the answer near the top, written plainly, with the supporting detail below it. A page that builds to its point over five paragraphs is harder to quote.
These tools favour sources that show real experience: named people, specific methods, original data and detail that could not be written by someone who had not done the work.
Generic explainers of the kind anyone could produce give an AI system no reason to choose you over the hundred other pages saying the same thing.
Comprehensive, interlinked coverage of one subject signals expertise more reliably than scattered pages across many.
That means strengthening a cluster of related pages and the links between them rather than adding new pages in the hope that volume helps.
Much of what AI tools say about a business comes from sources the business does not control: directories, roundups, trade press, community answers.
Being present and described accurately in those places influences the answer more than anything on the website alone.
AI visibility is measurable, but not through rankings. The useful measure is whether the business appears for the questions its buyers ask, and whether the description is accurate.
That means running a fixed set of queries, recording what comes back, and repeating it over time.
You will see this work described as Answer Engine Optimisation, Generative Engine Optimisation or AI SEO. The terms are used loosely and often interchangeably.
AEO usually refers to being the answer rather than a result. GEO usually refers to being cited inside an AI-generated response. AI SEO tends to mean whatever the person selling it wants it to mean.
The distinctions matter less than the overlap. All three describe the same underlying job: being clear enough to understand, useful enough to quote and credible enough to recommend. If a proposal treats them as three separate services with three separate fees, that is a reason to ask questions.
The first step is a baseline. I run the questions your buyers actually ask through the tools they actually use, and record what comes back: whether you appear, what is said about you, and which sources the answer is built from.
That baseline usually explains the problem on its own. Businesses are often missing entirely from the questions that matter, or present but described from an outdated profile somewhere they had forgotten about.
From there the work divides into what is on the site and what is not. On the site: clearer descriptions, structured data that agrees with the page, and pages rewritten so the answer comes first. Off the site: making sure the places these tools draw from describe the business correctly.
Then the same questions are re-run at intervals. Not because the number is the point, but because it is the only way to tell whether anything is working.
It will not get you into an AI answer on demand. These systems are not rankings with a route in, and they change without notice or explanation.
It will not replace search traffic you have lost to AI summaries. Where a question is now answered without a click, that click is unlikely to return, and the useful response is to be the source being quoted rather than to chase the visit.
It will not work on a site with nothing distinctive on it. These tools are selecting between sources, and a page that reads like every other page gives them no basis to choose it.
If the website has technical problems, thin service pages or no clear description of what the business does, AI visibility is not the first thing to fix. Those same weaknesses are the reason the business is not being cited, and they are better addressed as SEO and content strategy work.
If you are not yet sure which questions matter to your buyers, an SEO audit or a wider marketing audit will tell you more than an AI visibility exercise will.
And if the honest answer is that your customers are not using these tools to choose a supplier in your market yet, I will say so. It is a reasonable thing to prepare for without treating it as urgent.
This suits businesses whose buyers research carefully before they choose, where being shortlisted matters and where there is real expertise to show.
It works best when there is something specific to say: a method, a dataset, a point of view or experience that a generic competitor cannot copy.
It needs access to the people who hold that knowledge, and a willingness to publish it rather than keep it in a proposal document.
If you want the longer argument first, I have written about what to prioritise for AI search and why the foundations still come first.
If you want to know how AI tools currently describe your business, or whether they mention it at all, tell me a little about the business and the questions your buyers ask.
I’ll read it personally and let you know whether AI search work is the right fit, or whether something else would help more first.