How to Map the Sources Shaping AI Answers in Your Industry: A Practical Audit Framework

A brand can appear in Google search results without appearing in the sources an AI answer cites. Understanding which pages actually inform the answers your buyers see requires a structured audit, not guesswork.

In a recent webinar hosted by Search Engine Journal, Gintare Rimolaityte, Chief Commercial Officer at Trendos, explained how to map those sources by industry and engine, then turn the gaps into a practical worklist. Her presentation covered consumer products and retail, IT and solution services, and communication services, drawing on Trendos’ analysis of citation patterns observed in mid-2026.

The central argument was not to abandon SEO. It was to stop treating search rankings as a complete picture of AI visibility. Rimolaityte contrasted a page of search results with an AI answer whose sources may require an extra click to inspect. Monitoring whether your brand appears in the answer is useful, but it does not explain which sources contributed to that response. Looking at those sources reveals a different set of decisions: whether to improve owned content, build a presence on relevant review sites, contribute to community discussions, or pursue editorial coverage.

Industry-specific citation patterns

The analysis showed that community and user-generated sources were prominent across all three industries examined. The remaining mix varied significantly, making a universal list of sites to target a poor substitute for examining your own category.

Datos clave
Speaker	Gintare Rimolaityte, Chief Commercial Officer, Trendos
Core insight	AI citation sources vary by industry and engine; a universal site list is ineffective
Recommended starting point	10-20 customer questions reflecting the buying decision
Key caution	One engine’s decline in a source does not mean it has disappeared from all engines

Rimolaityte interpreted the retail pattern partly through content structure: product pages often make specific facts easy to locate. For other businesses, she suggested clear question-and-answer content rather than assuming every page should look like an ecommerce listing. She emphasised that this is a starting hypothesis to test, not evidence that short answers alone cause citations.

Engine-level differences matter

A change in one engine does not establish a trend across all AI search. Rimolaityte used Reddit as an example: reduced prominence in ChatGPT did not mean it had disappeared from the other engines she examined. She also described differences in the balance of community, editorial, and brand-owned sources across Google AI Overviews, Perplexity, ChatGPT, and Gemini. In her examples, Gemini’s source mix was particularly dependent on industry.

The practical implication is to keep engine-level results visible in your reporting. A combined total can obscure where a source remains useful and where it no longer contributes as much.

There is also a reason to participate in communities beyond AI citations. If customers already discuss your category on Reddit, those conversations can be valuable in their own right. Rimolaityte recommended honest participation from a company or employee voice rather than disguised accounts or a promise that presence can simply be bought.

How to conduct a citation audit

Rimolaityte’s audit begins with 10 to 20 questions customers might ask while choosing a product or provider. Topics and keyword research can help develop that set, but the questions need to reflect the buying decision you want to understand.

Run the questions in the engines relevant to those buyers, record the cited sources, and group them by type. Then inspect the pages where competitors appear and your brand does not.

Rimolaityte explained why those gaps deserve attention: “If your competitor appears in those sources and you don’t, this is your opportunity to get the mention on the source because this source is already being used to get the answer, right?” The opportunity is more specific than a general request for more brand mentions: you have a source URL, a customer question, and an existing competitor reference to investigate. It still requires judging whether your business genuinely belongs on that page and what a useful contribution would be.

Do not give every discovered domain equal priority. Rimolaityte recommended examining citation share and which source categories drive answers for your industry, rather than chasing thousands of editorial sites indiscriminately.

Video content and tracking

During the Q&A, attendees asked whether cited YouTube content had to come from users rather than a company or its founders. Rimolaityte said brand and founder content could also be useful; her recommendation was to make it insightful and clearly structured. She emphasised concise question-and-answer formats and accurate transcripts.

When asked how to track a planned series of video FAQs, she suggested checking whether the channel or its content appeared among the sources for the prompts being monitored. That gives a video project a more concrete measurement plan than publishing a batch and hoping for visibility. Identify the customer questions first, make the answers useful, and track the relevant sources over time. A transcript supports that work, but does not guarantee a citation.

An agency attendee asked what to do when they could recommend changes but could not implement them. Rimolaityte advised preparing a concise, client-specific report that identifies the gaps and gives the client a to-do list.

For brands and agencies in India tracking AI visibility, this framework offers a way to move beyond vanity metrics and focus on the sources that actually influence buyer answers. The full session includes detailed industry comparisons and audience Q&A.

Source: Search Engine Journal, “How To Find The Sources Shaping AI Answers In Your Industry” by Heather Campbell, published September 21, 2026. Available at: https://www.searchenginejournal.com/how-to-find-the-sources-shaping-ai-answers-in-your-industry/589756/