What is AI visibility, and why does it matter for your business?

Ask ChatGPT or Google's AI who's good at commercial litigation in Manchester, or who does the best executive coaching for scale-up founders, and it will name a handful of businesses.

If your business isn't one of the names it says, you don't exist in that conversation. It doesn't matter if you rank first on Google for the same search.

That's AI visibility: whether the AI systems your prospective clients now ask instead of Googling actually mention you when it matters. It's a different game from search engine optimisation, and most businesses that built their reputation on being findable haven't noticed the ground has moved.

What AI visibility actually means

Search engine optimisation was about ranking. You picked keywords, built pages around them, and Google rewarded structure and authority signals with a position on a results page.

AI visibility is about recommendation. When someone asks an AI assistant "who's the best GEO consultant for a small law firm," the system doesn't hand back ten links to sift through. It synthesises an answer and names a few sources it trusts.

The unit of competition has changed. You're not competing for a spot on page one anymore. You're competing to be one of the names an AI decides is worth saying out loud.

The systems making that decision aren't reading your homepage the way a human would. They're pulling from the specific, citable content you've published, and weighing it against everyone else's.

That's the plain-language answer to "what is AI visibility": whether AI systems can find evidence that you're worth recommending, and then actually recommend you, for the questions your buyers are asking.

The pivot that proved the problem was real

I didn't arrive at this from a marketing textbook. Ghostart started as a platform to help people write in a way that sounded like them instead of sounding generic. That was the whole mission: fight the beige, keep the person in the writing.

Then business owners we were talking to raised a different problem. They were watching their Google visibility slip while their competitors' AI mentions climbed, and they had no idea what to do about it because nobody had explained the mechanism to them.

The problem wasn't abstract for them. It was clients they used to get from a Google search who'd started asking ChatGPT instead, and getting someone else's name back.

We rebuilt the product around that evidence rather than staying loyal to the original plan. That's Ghostart GEO now: the same belief about authentic writing, pointed at a business risk instead of a style preference.

I was wrong about one thing early on. I thought the case for authentic writing was mostly aesthetic, a craft argument about writing that sounds alive versus writing that sounds dead. It turned out to be a revenue argument.

The businesses losing AI visibility weren't losing it because their writing was ugly. They were losing it because it was interchangeable, and interchangeable content gives an AI system nothing distinct to cite.

The evidence behind the shift

This isn't one company's anecdote. Semrush analysed more than 300,000 AI-generated answers and found something that should worry any business that publishes only through a faceless "company voice": individual, named contributors were outciting company pages.

AI systems were more likely to reference a specific person's expertise than a brand's generic "About Us" content.

That finding lines up with what we'd already built the Beige-ometer to fix, years before AI search existed. We scored content for voice, specificity, and whether it sounded like a real person had actually said it, because generic writing failed human readers.

The Semrush data confirmed the same content now fails a different audience too: the AI systems deciding who gets recommended.

Two completely different starting points, a founder's conviction about what makes writing land, and independent research on what AI systems actually cite, point the same direction. That convergence means this isn't a trend to react to. It's a mechanism, and mechanisms can be worked with rather than chased.

Why this hits expertise-led businesses hardest

If you run a consultancy, an advisory practice, an agency, or any business where clients pick you because they trust your judgement rather than your price, this matters more for you than for almost anyone else.

Commodity businesses compete on price and footfall, so an AI recommendation engine barely changes their maths. But if your entire value proposition is "we know this better than most," and the AI your prospective client just asked has never heard of you, you've lost the pitch before you knew there was one.

The founder of a ten-person advisory firm and the marketing lead running content for a fifty-person consultancy are asking the same question, even if the scale of their problem is different: will I show up when someone asks ChatGPT who's good at this?

That question doesn't get answered by better SEO tactics or more keywords bolted onto old pages. It gets answered by whether there's enough specific, human, citable content out there proving you know what you're talking about.

What actually earns an AI citation

The instinct, when people hear "AI visibility," is to look for a lever: a bit of schema markup, a prompt trick, some technical fix that gets the machine to notice you.

I understand the instinct. It's the same one that drove a decade of SEO gaming. But AI systems are built to see through that kind of manipulation, because their entire value depends on giving trustworthy answers.

What they respond to instead is the same thing a discerning human reader responds to: content with a real point of view, written by someone who clearly knows the subject from the inside, with enough specificity that it couldn't have been written about any other business.

A generic paragraph about "the importance of client communication" could belong to anyone. A paragraph about the actual decision you made when a client pushed back on a strategy, and what you learned from being wrong, could only have come from you.

That's the same standard we apply through the Beige-ometer to human-facing writing. It's close to identical to what earns a citation from an AI model. Does this sound like a real person, and will a machine trust this enough to recommend it, are one problem wearing two hats.

What to do next

Start by asking an AI assistant the question your best clients would ask before they hired you. See who it names. If it isn't you, that's not a ranking problem to tinker with. It's a sign there isn't enough specific, human proof of your expertise out in the world yet for a machine to find.

The fix isn't a technical trick bolted onto your existing pages. It's publishing the real judgement calls, the client moments, the things only you could have written, until the systems watching have something worth citing.

Ghostart launches on 29 September and you can join the waitlist now.