[The Engines]
How to Ensure Your Brand is Cited by AI Search Engines
A new analysis of 6.8 million AI citations shows 86% trace back to sources brands already control: their own websites and business listings. Here is what changed on October 9, 2025, and what to fix first.
The short answer
On October 9, 2025, Yext published an analysis of 6.8 million AI citations across ChatGPT, Gemini, and Perplexity showing that 86% trace back to brand-controlled sources: first-party websites (44%) and business listings (42%). Forums like Reddit accounted for just 2%. If you have been chasing forum mentions for AI visibility, the data says to redirect that effort into the site and listings you already own.
What changed in how AI engines cite brands?
Yext analyzed 6.8 million AI citations collected across ChatGPT, Gemini, and Perplexity between July and August 2025, and the finding reorders how marketers should prioritize AI visibility work: 86% of citations trace back to sources a brand already controls. Not a Reddit thread. Not a stray forum post. A brand's own website or its business listings.
The breakdown, reported by Search Engine Land on October 9, 2025, splits out to 44% from first-party websites, 42% from business listings, 8% from reviews and social media, and just 2% from forums. That last number is the headline: the widely repeated assumption that Reddit drives the bulk of AI answers does not hold up against a dataset this size.
The scale matters here. Yext's methodology drew on roughly 1.6 million queries per model, spanning four intent quadrants (branded and unbranded, objective and subjective) across four industries, and it touched a long tail of over 20,000 unique domains. This was not a narrow sample built to confirm a thesis. It is closer to a census of how three major AI engines actually cite sources.
Who does this affect, and where does it vary by platform?
Any brand investing in AI search visibility is affected, because the study reframes where that investment should go. That includes B2B SaaS and tech growth teams trying to show up when someone asks an AI engine for 'best [category] tool,' and it includes local, multi-location, and service businesses trying to show up for '[service] in [city]' prompts.
The mix is not identical across engines. Gemini leans hardest on websites, at 52.1% of its citations. OpenAI's ChatGPT leans more on listings, at 48.7%. Perplexity spreads further across secondary aggregator sources, including MapQuest and TripAdvisor, rather than concentrating on one type.
Industry shapes the mix further. Retail brands drew 47.6% of their citations from websites. Healthcare drew 52.6% from listings, likely reflecting how often people ask AI engines for a provider near them. Food service was the outlier: reviews and social content accounted for 13.3% of its citations, well above the average, because people ask AI engines what a restaurant is actually like before they ask what it serves.
The practical read: a SaaS company optimizing for comparison prompts should weight first-party content higher, while a multi-location service business should weight listing accuracy higher. The 86% figure is an average across very different behavior underneath it.
| Source type | Before: common industry assumption | After: Yext data (Jul to Aug 2025) |
|---|---|---|
| Brand websites | One tactic among many, competing for budget with PR, links, and forum outreach | 44% of all AI citations, the single largest category |
| Business listings | Treated as local SEO housekeeping, not an AI visibility lever | 42% of all AI citations |
| Forums (e.g. Reddit) | Assumed to be a dominant, hard-to-influence source AI models favor | 2% of all AI citations |
| Reviews and social media | Assumed to carry outsized weight because AI 'sounds like' it reads reviews | 8% of all AI citations |
How should you respond to the shift toward brand-controlled citations?
Stop treating forum seeding as a primary AI visibility tactic and redirect that budget into the two channels the data says already win: your website and your listings. The table below lays out the shift between what marketers commonly assumed drove AI citations and what Yext's data actually shows.
On the website side, that means structured, crawlable, frequently updated first-party content built to directly answer the questions your buyers are asking an AI engine right now, not just the questions a Google ranking factors checklist assumes. On the listings side, it means auditing every place your business is described (directories, maps, review platforms, industry-specific aggregators) for accuracy and completeness, since 42% of citations come from exactly that layer.
Reviews and social content still matter, just less than the industry narrative implied, and unevenly: food service brands should weight them heavily, most others should treat them as a secondary channel worth maintaining but not the center of the strategy.
This is the same mechanic behind what we call Citation Engineering: publishing authoritative, well-structured, first-party content at a scale and cadence that AI engines can find and trust, then tracking Citation Share (the percentage of relevant AI answers in a category that cite you) as the metric that actually reflects the work. LeadHaste, Omnicite's own site, went from 0 to 1 million impressions and 200-plus AI citations, with Domain Rating moving from 1 to 24, over 4 months by doing exactly this: scaling first-party publishing rather than chasing forum mentions.
What should you check first this week?
You do not need to overhaul a content strategy overnight. Start with the assets the data says already carry 86% of the weight.
- Audit your core business listings across directories, maps, and industry aggregators for accuracy, completeness, and consistency.
- Check that your key pages are crawlable and structured (clean headings, direct answers near the top, schema markup where relevant), since AI engines cite what they can parse quickly.
- Publish first-party content that directly answers category and comparison questions your buyers are already asking an AI engine, rather than content built only for a Google ranking.
- If you are in food service or another review-heavy category, keep review volume and freshness current instead of ignoring it.
- Track Citation Share and Answer Presence, not just keyword rankings, so you can see whether the shift is working.
- Deprioritize forum seeding as a primary tactic. It accounted for 2% of citations in this dataset, not the dominant share the prior narrative assumed.
Key takeaways
- 86% of AI citations across ChatGPT, Gemini, and Perplexity trace back to brand-controlled sources: 44% from websites, 42% from business listings.
- Forums like Reddit accounted for just 2% of citations in this 6.8 million-citation dataset, undercutting the assumption that forums dominate AI answers.
- The mix varies by platform: Gemini favors websites (52.1%), ChatGPT leans on listings (48.7%), Perplexity spreads across secondary aggregators.
- The mix varies by industry: healthcare leans on listings (52.6%), retail on websites (47.6%), food service on reviews and social (13.3%).
- The response is not a Reddit strategy. It is structured, crawlable first-party content plus accurate, complete business listings.
- Track Citation Share, the percentage of relevant AI answers in your category that cite you, to see whether the shift is working.
Omnicite Editorial. "AI Citations: How Brands Get Cited by AI Search" The Citation Report, Omnicite. https://omnicite.co/blog/how-to-ensure-your-brand-is-cited-by-ai-search-e/
Sources
86% of AI citations across ChatGPT, Gemini, and Perplexity trace back to brand-controlled sources; breakdown of 44% websites, 42% listings, 8% reviews/social, 2% forums; study covered 6.8 million citations from July to August 2025 Yext, 2025-10-09
Reporting on the Yext study, including platform variation (Gemini 52.1% websites, ChatGPT/OpenAI 48.7% listings, Perplexity diversifying to MapQuest and TripAdvisor) and industry variation (retail 47.6% websites, healthcare 52.6% listings, food service 13.3% reviews/social) Search Engine Land, 2025-10-09
Frequently asked questions
What did the Yext AI citations study actually find?
That 86% of AI citations across ChatGPT, Gemini, and Perplexity come from sources a brand already controls: 44% from first-party websites and 42% from business listings, based on 6.8 million citations collected in July and August 2025.
Does this mean Reddit does not matter for AI search anymore?
Reddit and other forums accounted for just 2% of citations in this dataset, far below the dominant role the prior industry narrative assumed. Forums still show up, especially on Perplexity, but they are not where most brands should focus first.
Which AI engines were included in the study?
ChatGPT, Gemini, and Perplexity. Yext analyzed roughly 1.6 million queries per model across four industries and four intent quadrants, covering more than 20,000 unique domains.
How is an AI citation different from a search ranking?
A ranking places you on a results page a person scrolls through. A citation is a source an AI engine names directly inside its answer. Citation Share, the percentage of relevant AI answers in a category that cite you, measures the second thing, and it is the metric this study speaks to.
What should a local or multi-location business prioritize first?
Listing accuracy and completeness across directories, maps, and industry aggregators. Listings accounted for 42% of citations overall and 52.6% in healthcare specifically, well above the average.
What should a B2B SaaS company prioritize first?
First-party content that directly answers category and comparison prompts buyers put to an AI engine, since websites drove 44% of citations overall and 52.1% on Gemini specifically.