The answers from ChatGPT, Perplexity, and Google AI Overview don't come out of nowhere: they draw from specific sources, which vary by sector and by engine. If you don't know which sites are cited in your category, you don't even know where to start to gain AI visibility.
The method described by Gintare Rimolaityte of Trendos on Search Engine Journal starts with a systematic audit: map the sources cited for each AI engine, compare them with those of competitors, identify the gaps. From there, content and outreach priorities emerge that are not guesses, but data.
For a marketing manager or an entrepreneur, the operational point is one: before producing more content, it's worth understanding who is already occupying the space in the AI responses for your sector. This article explains how to structure that analysis and what to do with it once completed.
The problem no one has measured yet
Many companies worry about declining organic traffic without asking where that traffic is going. A growing portion ends up in AI search engine's synthetic answers: the user asks a question, gets a complete answer, and doesn't click on any site. Or rather, they only click on the sites that AIs cite as sources.
Who are these sites? In your industry, you probably don't know yet. And this is the gap to fill before any other content decision.
If you're already monitoring how the Google AI Mode reduces organic clicks , the next step is to understand exactly who is gaining the visibility you are losing. Not in the abstract: by domain name, by content type, by AI engine.
How an AI citation audit works, step by step
The approach described by Rimolaityte on Search Engine Journal is broken down into distinct phases. It's not a one-click automatable process: it requires manual work or dedicated tools, but the result is a concrete map of who dominates AI answers in your space.
The fundamental steps:
- Define relevant queries — the questions your customers actually ask, not traditional SEO keywords. AIs answer questions, not search strings.
- Querying each engine separately — Google AI Overview, Perplexity, ChatGPT, and others cite different sources. The map changes per platform.
- Record mentioned domains — for each answer, note which sites are mentioned or linked as a source. Do this on a significant sample of queries, not just one or two.
- Classify sources by type — industry media, company blogs, forums, institutional sites, aggregators. The type of source says a lot about what AIs consider authoritative in that category.
- Compare with competitors — do your direct competitors appear in citations? How often? On which topics yes and on which topics no?
This work is the basis of what in the field of AI-era SEO is called citation visibility : it's no longer enough to be indexed, you need to be considered a reliable source by language models.
Where competitor gaps are hiding
The most useful part of the audit isn't seeing who's there, but seeing who's missing. If a competitor is never mentioned on a topic they dominate with traditional content, that's a real opportunity.
The gaps are found by cross-referencing two questions:
- On which queries does the competitor appear in classic SERPs but not in AI answers?
- On which topics are no industry players cited, leaving room for generic or external sources?
The second type of gap is often the most valuable. If AIs answer questions in your industry by citing Wikipedia or general information sites, it means no vertical player has yet produced structured content substantial enough to be considered a source. Whoever arrives first with the right format occupies that space.
This connects directly to a topic we've already covered analyzing the AI visibility as a new metric for SMEs : traffic is no longer the only indicator, and those who don't measure citations are navigating blind.
From audit to operational priorities
An AI citation map is not a document to be filed away. It only becomes useful if it transforms into two types of concrete action.
Content priority. Topics where no one is cited become the first ones to focus on. Not with generic content, but with pages that directly and structurally answer the exact questions that AIs receive. The content validation workflow for AI Search is the right tool to ensure that this content has the characteristics that models are looking for.
Outreach priority. The sources most often cited are newspapers, industry blogs, and sites with high vertical authority. Appearing on those sources — with a contribution, original data, or a quote — increases the likelihood of being cited in AI responses yourself. This isn't classic link building: it's citation reputation building.
If you want to understand who is appearing in your place in AI responses today, the practical starting point is the AI citation audit that we've structured to analyze this exact scenario.
Who benefits from this work — and where it's not worth investing
Auditing AI sources makes sense for those operating in sectors where AIs are already used to search for information: professional services, technology, health, finance, education, B2B in general. In these contexts, a significant portion of information searches already go through AI search engines.
It makes less immediate sense for those selling very local or impulse products, where the user doesn't do comparative research before buying. In that case, other channels remain a priority.
The most frequent mistake is doing the audit only once and considering it final. The sources cited by AI engines change over time as models are updated and new content enters the index. Quarterly monitoring is the minimum for those who want to keep up.
To navigate this set of choices, the reference point is our section dedicated to SEO and AI visibility , where we gather updated practical analysis and methods.
Related articles
Discover more articles exploring similar topics, selected to offer you a more complete and stimulating perspective. Each piece of content is carefully chosen to enrich your experience.
