Customer Impact

SEO & GEO

ChatGPT SEO: how to get found in ChatGPT

Copy for AI

ChatGPT SEO is the work of getting your brand named in ChatGPT’s answers.

More and more people no longer ask Google, but ChatGPT. Instead of ten blue links they get one answer, often with a handful of brands and sources in it.

The verdict: ChatGPT SEO takes a different approach than classic search engine optimization. A clear entity, citable content and off-site mentions decide whether you’re named, or your competitor.

This is the core of generative engine optimization (GEO): not optimizing for a click, but for a mention in the answer an AI model generates.

How does ChatGPT choose its sources?

From two places: what it learned in training, and what it fetches live.

ChatGPT works differently than a search engine.

Part of what the model “knows” sits in its training data: text from the whole web it was trained on.

The live search layer

On top of that, newer versions fetch live information via a built-in search layer.

That happens when a question is current or specific.

OpenAI’s OAI-SearchBot crawler is what surfaces websites in those search answers.

So don’t block it in robots.txt.

Here’s what an open robots.txt looks like on our own site:

The robots.txt file of customerimpact.be with one rule that allows every crawler plus the sitemap line, so OAI-SearchBot and other AI bots can read the site for ChatGPT SEO Our own robots.txt, October 2026: one rule for all crawlers, so every AI bot, training and search, may read the site.

One rule for all crawlers, plus the sitemap. Our guide to letting GPTBot and other AI crawlers in covers the per-bot rules.

How that live layer leans on a search index is explained in how ChatGPT uses the search index.

What that means for your brand

Two things follow for you as a brand:

  • Training data: the model builds a picture of who’s linked to a topic.
  • Live queries: it picks sources that are clear, reliable and well structured.

If your brand consistently shows up in many sources next to a theme, the chance rises that the model names you when someone asks about it.

For live queries, it prefers sources it can easily distil an answer from.

Not the top-ranking page

So ChatGPT doesn’t look for “the page that scores highest on a keyword”. It looks for “the source that gives the most reliable and clear answer to this question”.

Ahrefs found that only 12% of links cited by ChatGPT, Gemini and Copilot rank in Google’s top 10 for the same prompt.

Here’s that gap in one chart:

AI CITATIONS VS GOOGLE TOP 10 Most cited links don’t rank top 10 Also in Google’s top 10 12% Not in Google’s top 10 88% Links cited by ChatGPT, Gemini and Copilot, same prompt. Source: Ahrefs AI search overlap study.

The study covered 15,000 prompts. Perplexity was the outlier, with 28.6% of its cited URLs in Google’s top 10.

That difference determines your entire approach.

How do you make your entity crystal clear?

By describing yourself the same way, everywhere, so a model can’t misread you.

An AI model thinks in entities: people, brands, products, topics and how they’re connected. Before you get found in ChatGPT, the model must understand who you are and what you stand for.

Our explainer on entity SEO covers the concept itself.

Consistency first

That starts with consistency. Use exactly the same brand name everywhere, the same description of what you do and the same core terms.

Call yourself a “B2B growth agency” in one place and something quite different elsewhere? Then a model struggles to connect the pieces.

Also put the logical relationships explicitly on your site: which problem you solve, for whom, and with which approach.

What happens when you’re vague

Our experience: if you can’t state your positioning clearly, the model fills the gap. That’s where it starts to hallucinate about you.

Three concrete moves

Concretely it helps to:

  • write a plain “about us” that says what you do and for whom
  • make one strong page per topic instead of five half ones
  • link your expertise to a recognizable face

At Customer Impact, founder Tanguy De Keyzer, author of the book The End of Search, is the voice behind our vision on GEO.

Such a consistent link between person, brand and topic helps a model place you. His take is in The End of Search.

The sharper your entity, the easier an AI model recognizes you as the right answer.

How do you write citable content?

Put the answer first and state facts so plainly that a model can quote them.

Citable content is content an AI model can take over literally, without guessing what you mean.

That’s a different goal than content that guides people through a long funnel.

Begin with the answer

Put the core at the top and only elaborate afterwards.

Models like to pick up the first, clear formulation.

A question as heading, followed by a direct answer in one or two sentences, works excellently.

Be explicit with definitions

Write “GEO stands for generative engine optimization and ensures AI models name your brand in their answer”.

Don’t spread a vague description over three paragraphs. The less interpretation is needed, the more safely a model can cite you.

Our guide to citable content for AI has more patterns like this.

Structure for readability

Short paragraphs, clear subheadings, lists and tables make it easy for a model to find the right passage.

The same structure that helps a human scan helps a model extract. That’s why AI cites tables and lists more often than running text.

Don’t make things up

Write claims you can substantiate and that are correct.

A model that judges your source as reliable names you more often.

Inaccuracy is counterproductive.

Rule of thumb: if you can’t point to where a number comes from, cut it.

How do you build off-site signals?

By getting named next to your topic in sources you don’t own.

ChatGPT doesn’t form its picture only from your own website.

What others say about you weighs heavily.

If your brand shows up consistently next to a topic in independent sources, that strengthens the connection the model makes. Think trade media, directories, forums and third-party articles.

Think broader than classic backlinks here.

Mentions count even without a link.

Being named in an overview article about GEO agencies, a podcast transcript or a trade magazine feeds the picture a model has of you.

Our piece on unlinked brand mentions explains why that still works.

Rule of thumb: the more often your brand appears in a credible context next to your core topic, the stronger your position in AI answers.

Review sites and overview pages

Watch out for sources AI models often consult for structure and facts. Overview pages, comparison articles and well-maintained trade communities are typical examples.

Here’s an example of such a source in action. We asked ChatGPT which CRM suits small B2B teams:

ChatGPT answer naming HubSpot, Pipedrive and Attio for small B2B teams, with G2 cited as a source, an example of off-site signals in ChatGPT SEO ChatGPT (free), 17 October 2026: three CRM suppliers named in one answer, with a review platform cited as a source.

Note the review platform among the sources. Being named in sources like that has a big effect, because they weigh relatively heavily.

Quality and relevance count more than volume. Ten strong, relevant mentions say more than a hundred loose ones outside your market.

Build a reputation, not a trick

Practically you actively build your visibility:

  • guest contributions in trade media
  • presence on platforms where your audience is
  • correcting existing mentions that are outdated

It isn’t a trick. It’s building a reputation a model can pick up.

Our guide to digital PR for AI citations shows where to start.

From our own data: in our study of which marketing agencies AI recommends, over half the answers cited an agency directory such as Sortlist.

ChatGPT SEO: how do you measure your mentions?

By asking ChatGPT your customers’ questions and logging whether you’re named.

With classic SEO you measure positions and traffic. With getting found in ChatGPT you measure whether and how you’re named. That takes a new habit.

Start from real customer prompts

Write a list of prompts that fit your market, such as:

  • “which agency helps with GEO in the Benelux”
  • “how do I get found in ChatGPT”

Our guide to prompt research for ChatGPT helps you build that list from real customer questions.

Then ask those questions yourself. Note each time: are you named, in which place in the answer, and in which context?

Here’s one of our own test runs:

ChatGPT answer on 17 October 2026 naming a ranked shortlist of three AI visibility tools, each with a label, the kind of answer you log when you measure ChatGPT SEO mentions Our own ChatGPT test on 17 October 2026. Note who’s named, in which order and with which label.

Order and label matter as much as the mention itself. A step-by-step routine is in measuring AI visibility without tools.

Track a trend, not a snapshot

Do this repeatedly over time, so you see a trend instead of a snapshot.

Also watch your competitors.

Who’s named where you’re missing, and why? That points you to the topics and sources where you still have work.

Expect answers to vary

Keep in mind that AI answers vary.

The same model can name different sources for the same question.

In a SparkToro study, ChatGPT and Google’s AI returned the same brand list less than 1% of the time across repeated runs of one prompt.

So look at the pattern across multiple checks, not at one chance answer. More on why in the probabilistic nature of AI answers.

Check the visits it sends

Mentions are one half. The other half is the visits ChatGPT sends you.

In GA4 you can filter Traffic acquisition on the source chatgpt.com:

GA4 traffic acquisition report for customerimpact.be filtered on chatgpt.com as source, the visits that ChatGPT SEO sends to the site GA4 Traffic acquisition in our own property, last 90 days, October 2026: 58 sessions from chatgpt.com.

That’s 58 sessions in 90 days for our site. Our guide shows how to track ChatGPT traffic in GA4 step by step.

Don’t forget your SEO basics

ChatGPT SEO doesn’t stand apart from your SEO.

A strong, technically healthy site with clear content is the foundation on which you build your GEO. The search layer ChatGPT uses leans partly on the same signals of findability and authority.

Neglect your SEO and you make your AI visibility unnecessarily hard. For the difference and the overlap, read GEO vs SEO.

What we saw on our own site

Our experience: when our own content engine published about 1,000 pages in three weeks, we counted around 1,000 ChatGPT citations. Volume on a solid SEO base works.

Mentions are a route, not a scoreboard

The real goal isn’t the mention itself, but what it delivers. Think more leads, more revenue and more people who see you as the right choice.

Visibility in ChatGPT is a route to leads, not a scoreboard.

Getting started with ChatGPT optimization

Getting found in ChatGPT isn’t a matter of luck. It’s the result of four things:

  • a clear entity
  • citable content
  • strong off-site signals
  • consistently measuring your mentions

Start small: choose one core topic, make the sharpest source about it on the web, and check every month whether ChatGPT names you.

Want to move faster? See how we make brands visible in AI answers via our approach to getting found in ChatGPT.

And if you just want to know where you stand now, get in touch for an honest conversation about your AI visibility.

Free website scan

Enter your website and get an automatic scan within minutes, with concrete technical and SEO improvements. No sales pitch.

Where should we send your report?

We only use your details for your scan. No spam, unsubscribe anytime.