From Google to AI: how to get your company cited by ChatGPT and AI search

What is GEO (Generative Engine Optimization)?

GEO (Generative Engine Optimization) is the discipline of optimising a website's content so that it gets cited as a source in the answers generated by artificial intelligences such as ChatGPT, Perplexity and Google's AI Overviews. Where SEO aimed to climb the list of ten blue links, GEO aims to be the source the AI reads, summarises and cites when a user asks it a question.

The shift is concrete and already under way. More and more people no longer open Google to choose between ten results: they ask an assistant and read a single conversational answer. If your company is not among the sources of that answer, for that user it does not exist. We set the scene for this when we wrote about the end of traditional SEO: GEO is the operational chapter of that story — the answer to "so what do I actually do now?".

How does AI search choose its sources?

AI search selects its sources by favouring content that answers the question directly, is well structured, comes from an authoritative domain and is up to date. It is not magic: it is an understandable logic, and one you can largely influence.

Engines like Perplexity and Google's AI Overviews / AI Mode work in a similar way: they interpret the question, retrieve relevant pages from the web and build a synthetic answer from them, linking to the sources used. ChatGPT, when it browses the web, does something comparable. The key point: AI does not reward keyword stuffing, but clean, easy-to-extract information. Five factors weigh most:

• Direct answers: the page answers the question immediately and explicitly, without forcing the reader to read between the lines.
• Clear structure: headings that are real questions, short paragraphs, bullet lists, tables. Content a machine can "unpack".
• Recognisable entities: company, people, product, place and concept names used consistently, so the AI understands what — and who — you are talking about.
• Authority: the domain is cited elsewhere, has a reputation and demonstrates real expertise on a topic.
• Freshness: the content is current, with visible dates and information that is not outdated.

What makes content "quotable" by an AI?

Content is quotable by an AI when it offers a clear, self-contained answer that the machine can extract and report without having to interpret or correct it. It is the principle of this article: every section opens with a sentence that answers its heading on its own.

In practice, quotable content combines a few ingredients. The answer comes before the explanation. The language is specific: numbers, proper names and sharp definitions beat vague phrasing. Claims are verifiable: real data, no invented statistics — AI, like a good journalist, trusts what it can check. And above all the content shows genuine expertise: it is no coincidence that we set out which AI content-creation tools truly work for an SME: AI can help you write more, but it is substance — the company's real experience — that makes you a source worth citing.

What is the llms.txt file and what is it for?

The llms.txt is a text file published at the root of a website that gives AI models a curated map of that domain's most important and reliable content. For AI it is the equivalent of what the sitemap and robots.txt are for traditional search engines: a way of saying "here is what really matters".

The standard involves two complementary files: llms.txt, a concise index with links to the key resources, and llms-full.txt, an extended version presenting the content in clean form, without the noise of menus, banners and code. For an AI this is valuable material: information already ready to read.

We are not saying this in the abstract: a126.it already publishes its own llms.txt and llms-full.txt files. You can see them at www.a126.it/llms.txt — the very technique we recommend to clients, applied to our own site. The standard is young and not every engine uses it yet, but the cost of adoption is minimal and it points in the right direction: making your company knowledge easy for a machine to understand.

Do structured data and online reputation still matter?

Yes: structured data and online reputation remain two fundamental levers, because they help AI understand who you are and trust you. GEO does not cancel established good digital practices: it makes them more important.

Structured data (the Schema.org standard) are labels invisible to the reader but perfectly readable by machines, explicitly declaring "this is a company", "this is an article written by this author on this date", "this is an FAQ". They help the AI interpret the page without ambiguity and connect entities: the silent infrastructure that makes content machine-readable.

Reputation matters because AI, like people, trusts those who are cited by others. Brand mentions on media outlets, industry directories, authoritative sites and reviews build the authority that leads a generative engine to treat you as credible. This is the territory of digital PR: no longer just backlinks for Google, but citations and mentions that teach AI your name is trustworthy. We discussed why online brand reputation is fundamental: in the era of generated answers, it becomes a direct ingredient of your visibility.

How do you measure AI-driven traffic?

AI-driven traffic is measured by isolating, in your analytics tools, the visits arriving from ChatGPT, Perplexity, Gemini and AI search, and by observing how often your domain appears as a source in the answers. It is a young kind of analysis, but already concrete. You work on three fronts:

• Referrals from assistants: in Google Analytics 4 (or equivalent tools) you segment the sources that point back to the AI assistants' domains, to see how many visitors arrive and what they do on the site.
• Presence in the answers: you check, manually or with dedicated tools, whether and how the brand is cited when you ask the AI the typical questions your customers ask.
• Quality over quantity: AI traffic is often lower in volume but more qualified, because the user arrives having already received an answer and wants to dig deeper. So measure conversions, not just visits.

The golden rule is not to expect the decimal-point precision of classic SEO: we are in a transition phase and the tools mature month after month. What matters is starting to look at this data now.

In summary

GEO (Generative Engine Optimization) is the discipline that makes a company's content quotable by artificial intelligences such as ChatGPT, Perplexity and Google's AI Overviews. These AI searches choose the sources that answer directly, are well structured, authoritative and up to date: to get cited you therefore need to open every piece of content with a clear, self-contained answer, use question-headings and lists, name your entities consistently, mark up pages with structured data (Schema.org) and build reputation through digital PR. A practical tool is the llms.txt file — which a126.it already publishes — offering AI a clean map of your content. AI-driven traffic is measured by isolating assistant referrals in Google Analytics 4 and checking how often the brand appears in the answers: low in volume, but qualified. It is not magic: it is method, and those who apply it today build a visibility their competitors will have to chase.

At A126 we help SMEs turn their website into a source that AI reads and cites: we analyse how your brand currently appears in generative answers, rewrite content with a GEO mindset, implement structured data and llms.txt files and set up the measurement of AI-driven traffic. To understand where you stand and where to start, contact us for a free consultation.

A126 Corporate Advisors — your company, among the sources AI chooses to cite.