isarcode

AI

Why the AI never mentions you

At a glance

  • More and more people put their question straight to Claude, ChatGPT, Perplexity or Google’s AI overview and read the single answer, rather than working down ten blue links. If the machine does not know your firm, you are simply not in that answer.
  • GEO (Generative Engine Optimization) is the craft of being cited as a source inside AI answers. It does not replace classic SEO; it is built on top of it.
  • The lever is the self-contained morsel: an AI quotes individual, freestanding statements, not whole pages.
  • isarcode measures visibility in search and AI answers with Hawkeye, a monitoring system built in-house, so the advice rests on evidence rather than on a hunch.

A friend who runs a small joinery outside Rosenheim told me recently, with the particular pride of a convert, that he no longer googles anything. He asks. The other week he wanted someone to wire his warehouse software into his accounts, so he put the request to ChatGPT and let it name the firms worth calling. It offered three. His usual supplier, a perfectly capable outfit he has used for years, was not among them. Not because the work is poor. Because the artificial intelligence had never heard of it.

And the story barely changes with the setting. City or countryside, a ten-person workshop or an international group, the same handful of names surfaces, and the same firms are quietly left off.

This is not some distant scenario. It is happening now, daily, and most firms never notice, because nothing shows up in a report that simply failed to take place.

The search is shifting, quietly

Every week the industry unwraps another model, and every week we are told which one is cleverer, faster, or harder to jailbreak. Gemini, then GPT, then Grok: a procession of announcements that is, by now, mostly marketing. Meanwhile, with far less fanfare, the more consequential thing is happening one layer down. It is a change in how people search at all, and in how the answers they receive are put together.

For years the ritual was familiar. You googled, you were handed a list, you clicked your way through it. Whoever sat at the top tended to win. Classic SEO is the discipline of sitting at the top.

Now a second layer has arrived. People ask Claude, ChatGPT or Perplexity, or the AI summary that Google now sets above everything else, and they are handed a finished answer. Often with two or three named sources, frequently with no click at all. That answer is the new first place, and it does not hold ten results. It holds a handful of names.

How sharply this bites is captured in an analysis of some sixty ChatGPT sessions (Suganthan, 2026). ChatGPT frequently writes brand names into its own search query before it fetches a single page. Those early-named brands turned up in the final answer 68.9 per cent of the time. Brands that only entered later, during the search itself, managed 2.1 per cent, roughly thirty-three times less often. The sample is small, though the mechanism takes seconds to reproduce yourself. The lesson is blunt: if you are not already on the machine’s mind at the first pass, you are, for practical purposes, out.

What GEO is, in a sentence

GEO stands for Generative Engine Optimization, and it is the ability to be cited as a source by the machines that now answer our questions. You will also see it called Answer Engine Optimization, or AEO: the same craft under a different label. The prize is no longer merely a good rank in Google’s list, but a mention inside the answer that ChatGPT, Claude, Perplexity or Gemini serves up directly, ideally with your name attached.

A necessary word before anyone panics: GEO does not replace SEO. It rests on it. Google itself says, in so many words, that AI search is still, at bottom, search. Get the fundamentals right, sound structure, genuine content, a clear shape, and you hold the better cards in the AI answer too. GEO is the extra layer laid over the top, not a demolition of what came before.

Why classic SEO no longer suffices on its own

Classic SEO optimises a whole page towards a good position in a list. An AI, however, does not lift the whole page. It lifts individual statements. It goes hunting for the one clean sentence that supports its answer, and sets it in.

Which means a page can sit third in Google and still never surface in an AI answer, because its claims are too vague, too tangled, or too dependent on the paragraph before to stand alone. Conversely, a crisp, self-supporting passage gets quoted gladly, even where the page around it is unremarkable. The contest is no longer really about content. It is about quotability.

The first three moves

You do not need to rebuild half your marketing for this. Start where the return is largest:

  1. Write the answer first. Open each section with the complete answer in a single sentence, then explain. No throat-clearing, no slow approach. An AI reads that first line and decides whether it works as a standalone morsel.
  2. Make your facts machine-legible. Clear definitions, concrete numbers, named examples, a tidy FAQ. Add structured data (schema markup), the small labels in the code that tell a machine what it is looking at. Technical, yes, but pleasantly finite.
  3. Bind your name to the claim. Not “this is how automation works”, but “this is how isarcode does automation”. Otherwise the AI quotes the point and forgets the source.

Sounds technical? The interesting part is not. It is a matter of clarity rather than trickery, and that same clarity, as it happens, serves your human readers rather well too.

How we know

GEO is young, and the field is thick with hunch. We would rather measure. isarcode runs Hawkeye, a monitoring system we built ourselves, which watches how a site fares in search and in AI answers and how quotable its content actually is. We run it exclusively for our clients, to plan their content with some precision rather than by candlelight. What that monitoring looks like, we are happy to show you over at isarcode.io.

And yes, all of this sits close to the question of how these systems “think” in the first place. If that intrigues you, read What is an LLM?, where we explain the engine behind Claude and ChatGPT in plain terms.

Local by nature, international by practice

isarcode is a Munich agency with an office in the Chiemgau. Which means, for you, a partner within reach who nonetheless plays in the same league as the larger houses and their international work. Local proximity, international standard, whether you sit by the Chiemsee, in Rosenheim, Munich or a good deal further afield.

Frequently asked

Is GEO the same as SEO? No, though they are relatives. SEO aims at a good position in Google’s list of results. GEO aims to have you named as a source inside the finished answer from AI systems such as Claude, ChatGPT or Perplexity. GEO builds on a sound SEO foundation.

Should I stop optimising for Google? By no means. The bulk of search still runs the classic way, and AI search draws on the very same fundamentals. GEO is added on top; it replaces nothing.

How do I tell whether an AI names me? By measuring it: put your typical customer questions to Claude, ChatGPT, Perplexity and Google’s AI overview, and see who gets named. That is precisely the check isarcode automates with Hawkeye, rather than leaving it to the odd manual glance.

Is this worth it for a small, regional firm? Especially then. In many regional niches almost no one is yet optimised for AI visibility. Whoever is soundly set up now gets named, while the competition is still asleep.

And now?

Curious whether Claude, ChatGPT and their cousins already know your firm, and what it would take for them to say your name? Let’s talk. We will look at your visibility with real figures, and tell you where the effort pays and where it does not, yet.