For twenty years, being found meant ranking on Google. Companies learned the rules: keywords, links, page speed, reviews. Whole industries grew up around page one. That world has not disappeared, but it now has a rival. More and more buyers skip the search box and ask an AI assistant directly: which accountant in my area handles landlords, which dental clinic offers implants near me, which software do small law firms use for intake.
The answer they get is not ten blue links. It is a short paragraph that names two or three businesses, with reasons. If you are named, you are on the shortlist before the buyer has visited a single website. If you are not, you do not exist in that conversation. I think being named first in that answer is the new page one, and almost nobody is measuring it yet.
That belief is why my team at B4Mind built NamedFirst. It asks the major AI engines, ChatGPT, Gemini, Claude and Perplexity, the questions real buyers ask, records who gets named, and then runs fixes on a monthly loop. This essay is about what that measurement has taught us.
Four engines, four different answers
The first lesson surprised clients more than it surprised us. Ask the same question of four engines and you will often get four different shortlists. A business that is named confidently by one engine may be absent from another. The order changes. The reasons given change. Sometimes an engine names a business that has closed, or describes a service it no longer offers.
This happens because the engines are built differently. They are trained on different data, refreshed at different times, and some of them search the live web while they answer while others lean more heavily on what they learned in training. They weight sources differently, and they have different ideas about what makes a source trustworthy. There is no single algorithm to optimise for, as there was with search.
The practical consequence is that a company cannot check its AI visibility once, on one engine, and assume it knows where it stands. Visibility has to be measured engine by engine, question by question, and repeated, because the answers move. That is why NamedFirst measures monthly rather than producing a one-off audit.
Why some businesses get named
When you look at which businesses are named consistently, a pattern appears. It is not the businesses with the most keywords on their pages. It is the ones whose basic facts are clear, consistent and repeated across many independent sources.
An AI engine is, in effect, trying to answer a question with confidence. It names a business when it has seen enough agreement about what that business is, what it does, where it operates and why it is relevant. If your website says one thing, your directory listings say another, and nobody else on the web mentions you at all, the engine has very little to be confident about. It will name someone else.
This is also why some well-known businesses are surprisingly absent. Size and brand recognition help, but they are not enough if the specific facts a buyer asks about are not written down anywhere an engine can find them. A clinic that offers a particular treatment but never states it clearly, on its own site and elsewhere, will lose that question to a smaller clinic that does.
Dashboards tell you what happened. Only a model of how the buyer decides tells you what to build next.
What a company can actually do
The good news is that AI visibility is not mysterious, and it is not purely luck. Most of the levers are things a well-run company should be doing anyway.
First, state your facts plainly. Who you are, what you do, for whom, where, and what makes you different. Write it on your own site in language a person would use when asking the question, not in marketing copy. Answer the questions buyers actually ask, directly, near the top of the page.
Second, make those facts consistent everywhere. Business directories, professional registers, review platforms, partner pages, press coverage. Inconsistency is one of the most common reasons an engine hesitates. The same name, the same description, the same services, the same location.
Third, give machines a clean way to read you. Structured data on your pages, a clear site map, and a short plain-text summary of who you are and what you do all make it easier for an engine to understand you without guessing. This site has one, for exactly that reason.
Fourth, earn independent mentions. Engines trust what others say about you more than what you say about yourself. Useful writing, real expertise shared in public, appearances, credible reviews: these are the signals that tell an engine you are a reasonable answer.
And fifth, measure, then repeat. Ask the questions your buyers ask, on every major engine, every month. Record who is named and why. Change one thing, then measure again. Without measurement, all of the above is guesswork.
What not to do
Every new channel attracts shortcuts, and this one is no different. Pages stuffed with questions and thin answers, fake reviews, invented credentials, content written for machines rather than people. Some of this may move an answer briefly. In my view it is a poor bet. The engines are improving quickly at spotting manipulation, and a business that is named for the wrong reasons is one bad answer away from being named for worse ones.
There is also a more basic reason to avoid it. The questions buyers ask an AI are often the questions they would ask a trusted advisor. Being named is only valuable if, when the buyer arrives, you are what the answer said you were.
The conversation is the new search
I do not think search engines are going away. But I do think the first moment of a buying decision is moving into a conversation with an AI, and that most companies have no idea what is being said about them there. They track rankings, traffic and reviews, and never ask the one question that increasingly decides who gets the call: when the buyer asks an AI, who does it name?
Asking that question is cheap. Answering it, consistently, across engines and over time, is where the work is. The companies that start measuring now will understand this channel while their competitors are still wondering why their enquiries have changed.