How Businesses Appear in ChatGPT and AI Search

Businesses increasingly ask a version of the same question: why does a competitor get mentioned by ChatGPT, or by Google's AI Overviews, when they don't? There's no single switch to flip, and no platform publishes a complete answer. But there is a useful way to think about what's actually happening, and what a business can reasonably influence.

The starting point: retrieval before generation

Appearing in an AI-generated answer isn't quite the same process as ranking in a traditional search results page, though the term "ranking" still applies in a modified sense. OpenAI states that ranking in ChatGPT Search is based on a number of factors designed to help users find reliable, relevant information, while also confirming there is no way to guarantee top placement. OpenAI states that when ChatGPT Search is used, ChatGPT can retrieve current information from the web and return linked sources, going beyond information available in its built-in model knowledge. It may also rewrite a user's request into one or more targeted search queries when working with search providers.

Google's AI Overviews and AI Mode work differently again: Google has said these features can issue multiple related searches, sometimes described as "query fan-out," to identify supporting webpages before generating a response. Microsoft documents Copilot Studio agents generating web queries sent to Bing to help ground certain answers.

The important distinction is that AI visibility isn't usually represented by one stable, visible ranking position in the way a conventional search results page is. Search and retrieval systems may still rank or prioritise candidate information internally, but those sources are then used within a generated response. That means a business can be considered, cited, or mentioned differently depending on the question, platform, and context.

Appearing doesn't always mean the same thing

It's worth separating out what "appearing" in an AI answer can actually mean, since these outcomes aren't interchangeable:

Mentioned — the business appears by name within an answer.

Cited — the business's website, or another source about it, is linked as supporting evidence.

Recommended — the business is actively presented as an option in response to a commercial or comparative query.

Described — the system answers a question specifically about the business, by name.

A business's website could be cited as an informational source without the business itself being actively recommended. A business could be mentioned based on information drawn from a third-party source rather than its own website at all. Treating all of these as the same thing, "are we showing up in ChatGPT?", tends to obscure what's actually happening and makes it harder to work out where to focus.

A framework for thinking about visibility

At Echeva, we use six conditions as a practical way of thinking about AI visibility: Accessible, Understandable, Relevant, Credible, Corroborated, and Measurable. These aren't published ranking factors or a formula used by any one AI platform. They're a framework for organising the different conditions that can affect whether information about a business can be found, interpreted, and used. Applied specifically to appearing in ChatGPT and similar AI search experiences, they look like this:

Accessible. For a website's content to be discovered and used within ChatGPT Search summaries and snippets, OpenAI says site owners should allow OAI-SearchBot and ensure their hosting or CDN permits traffic from OpenAI's published IP ranges. Blocking legitimate crawlers can materially restrict how a site's content is retrieved and used in ChatGPT Search, though a URL may sometimes still surface as a link and title via other sources even when blocked.

Understandable. The system needs to be able to work out who the business is and what it does. A page that's vague, generic, or inconsistent about the business's identity and offering is harder to use confidently, even if it's technically accessible.

Relevant. The business's content needs to actually address the kind of question being asked. A business can be well known in general and still be a poor match for AI retrieval if the information available about it doesn't clearly address the specific need or question being asked.

Credible. When systems retrieve information from the web, the quality and reliability of the underlying sources matter. Exactly how different platforms assess source quality is proprietary, so credibility is best treated as a principle rather than a universal published ranking factor.

Corroborated. Information that's consistent across a business's website and credible third-party sources is easier to verify than information that conflicts from one source to another. Corroboration is useful for reducing ambiguity, even though platforms don't publish a universal weighting for it.

Measurable. Without some way of checking whether and how a business is appearing, it's difficult to know whether any of this is working, or where the actual gaps are. See our guide on measuring AI search visibility for more on this.

Why one business may appear while another doesn't

There's rarely one identifiable cause. A difference in visibility can reflect whether relevant information is available to the retrieval system, how closely that information matches the question, how clearly the business and its services are described, which sources the platform retrieves, and the platform's own selection and generation processes.

The six conditions above are best used as areas to investigate, rather than as a checklist where satisfying every item guarantees inclusion.

An example

Common mistakes

Assuming a good website is enough. Design quality and user experience matter for visitors, but they don't substitute for the underlying clarity and consistency that AI visibility can also depend on.

Overlooking third-party information. A business's own website is only one part of its wider online representation. Relevant directories, review sites, and business profiles can provide additional information about the organisation, although platforms don't publish a universal weighting for how individual sources affect AI visibility.

Treating this as a single project rather than an ongoing practice. Business details change, competitors improve their own visibility, and platforms update how they retrieve and use information. A one-off fix doesn't stay effective indefinitely.

Expecting a guarantee. Any provider promising guaranteed ChatGPT or AI Overview mentions is making a claim no one can actually back, since the platforms themselves control what they surface, and OpenAI is explicit that there's no way to guarantee top placement.

Checking how your business currently appears

A simple starting point is to try representative, non-branded questions that resemble how a genuine customer might search, for example, asking for providers of a particular service in a relevant location rather than asking specifically about the business by name.

These checks should be treated as observations rather than definitive rankings. AI responses can vary by platform, wording, location, timing, and other contextual factors, so a single test doesn't provide a reliable measure of overall visibility. More systematic measurement looks at patterns across a representative set of queries and platforms, rather than whether a business appeared in one response.

It's also worth checking Google's Search Console, which has begun rolling out a dedicated Generative AI performance report to a subset of websites, announced in June 2026, showing impressions for a site's pages within AI Overviews and AI Mode. This data is also included within the standard overall performance report.