What 176,000 AI Citations Reveal About Where ChatGPT and Google AI Get Their Answers

For years, search engine optimization followed a familiar playbook: target high-volume keywords, build authoritative backlinks, and earn mentions in prominent news publications. However, as generative engines transform how consumers discover products, evaluate services, and make buying decisions, that traditional playbook is being rewritten.

To understand how AI engines synthesize information and choose their sources, Mencoro conducted a data study analyzing AI answers in Spain between July 1 and September 30, 2026. Tracking 42 brands across eight core commercial sectors, including health, finance, retail, automotive, and technology, the study evaluated 23,466 AI answers generated across four major platforms: ChatGPT, Google AI Overviews, Google AI Mode, and Perplexity.

The study examined the most cited domains in AI answers, capturing 176,332 total citations across 22,193 distinct domains. The findings show how AI search operates today and that each engine relies on fundamentally different source profiles to inform its responses.

1. The Core Distribution: Markets Beat Digital Press

A common assumption in digital strategy is that securing coverage in top-tier news media guarantees visibility across AI search engines. The data tells a very different story.

• 72.6% of citations go directly to company websites (including direct competitors, specialized service providers, merchants, and manufacturers).
• 11.6% of citations point to industry directories, review hubs, and online marketplaces.
• 4.3% of citations go to news outlets, magazines, and digital press.
• 4.8% of citations point back to the specific brand’s own domain being queried.

When users ask AI engines commercial questions about what to buy, who to hire, or where to go, the models prioritize direct market infrastructure over traditional media commentary. AI engines turn primarily to functional, transactional, and category-specific business websites to construct their recommendations.

Furthermore, a brand’s owned website accounts for under 5% of total citations. Winning in AI visibility requires optimizing your broader digital presence across the web, rather than relying solely on your primary domain.

2. No Universal Database: The Hyper-Specific Web

Another significant insight from the dataset is the fragmentation of source selection across different queries.

57.4% of citations point to domains that appeared for only one brand.

There is no golden list of 100 universal websites that guarantee visibility across every industry. Outside of a small cluster of dominant horizontal platforms, AI engines dynamically crawl and cite hyper-specific, niche-relevant web nodes tailored to each query and vertical.

For brand marketers and SEO strategists, this means visibility strategies cannot be copy-pasted across sectors. Understanding the specific digital ecosystem of your immediate vertical local directories, specialized suppliers, trade publications, and category-focused platforms is far more critical than acquiring generic placement on high-authority websites.

3. Dissecting the Engines: Specialized Source Profiles

The study highlights sharp operational differences in how individual AI models gather information. Each platform relies on a distinct mix of sources to formulate answers.

Google AI Overviews & AI Mode: Social-First and Visual Discovery

Google’s AI cites social networks far more than ChatGPT does: Instagram, Facebook, and TikTok citations come almost entirely from AI Overviews and AI Mode:

• Instagram was cited in answers for 40 out of the 42 tracked brands.
• 97.6% of all Instagram citations came directly from Google AI Overviews and Google AI Mode.
• Platforms like Facebook and TikTok also featured prominently within Google’s generative ecosystem.

Brands maintaining active, highly engaging visual and social channels are significantly more likely to surface in Google’s AI-generated summaries.

ChatGPT: Press, Official Records, and Deep Text Analysis

ChatGPT demonstrates a completely different retrieval pattern, leaning heavily on authoritative written records, major news publications, and official public documentation:

• ChatGPT accounted for 84.1% of all citations to Spain’s major newspaper, El País.
• ChatGPT generated 94.4% of all citations to BOE (Spain’s official state gazette).
• Public administration websites and government bodies made up 11.7% of ChatGPT’s total citation pool.
• TikTok received zero citations from ChatGPT throughout the entire study period.

ChatGPT prioritizes structured, highly verified textual information. When answering commercial inquiries that touch on regulation, market context, or established brand history, OpenAI’s model turns to traditional press and primary governmental records.

Perplexity: Real-Time Web and Video Reference

Perplexity shows a strong reliance on video content as a primary reference source. Specifically, Perplexity accounts for 191 of the 246 YouTube citations (78%) in the study, making video content its most distinctive source category.

What This Means for Brand Strategy

The data shows AI search isn’t a monolithic channel. Optimizing for AI visibility requires a multi-faceted approach tailored to how different engines retrieve information:

1. Focus on Vertical Footprints: Because over half of all citations go to domains that are unique to a single brand, conduct category-specific research to identify the exact directories, comparison hubs, and secondary sites AI engines draw from in your niche.
2. Diversify Beyond Owned Media: With owned websites receiving only 4.8% of citations, digital authority relies on third-party coverage across regional platforms, partner networks, and industry hubs.
3. Align Channels with Platform Goals:
   • To influence Google AI Overviews and AI Mode, maintain consistent activity on social and visual channels like Instagram.
   • To influence ChatGPT, prioritize authoritative digital PR, whitepapers, official documentation, and coverage in major news outlets.
   • To influence Perplexity, optimize video content on platforms like YouTube where it actively draws references.

Understanding these platform-specific citation patterns allows brands to build targeted strategies that ensure their products and services remain visible wherever customers ask questions.