🟢 🏥 In Practice Published: · 2 min read ·

arXiv:2605.25517: Study Reveals What AI Search Engines Cite — Relevance and Position Are Key

arXiv:2605.25517 ↗

Urednička ilustracija: Istraživanje otkriva koji sadržaj AI tražilice citiraju — relevantnost i pozicija ključni

Rahul Vishwakarma, Shushant Kumar, and Ratnesh Jamidar conducted 252,000 RAG experiments across six LLMs to determine which content factors determine citation in AI search engines (GEO — Generative Engine Optimization). Topical relevance and position in the list are key; explicit pricing and a fresh date provide additional advantage, while formatting has minimal impact.

🤖

This article was generated using artificial intelligence from primary sources.

What Is GEO and Why Is Citation in AI Search Engines the New SEO?

Rahul Vishwakarma, Shushant Kumar, and Ratnesh Jamidar from Sprinklr published the study “What Gets Cited: Competitive GEO in AI Answer Engines” — the first systematic analysis of factors determining which sources AI search engines cite in generated responses.

GEO (Generative Engine Optimization) is a new discipline arising from the emergence of AI search engines such as Perplexity, ChatGPT Search, and Gemini. Unlike classic SEO, which measures ranking on results pages, GEO measures citation — whether an AI search engine will include a specific page as a source when generating responses. Citation is particularly valuable because it directly builds brand visibility in AI-mediated user sessions.

How Did Researchers Measure the Impact of Content Factors?

Researchers built a controlled experimental framework with two documents in a RAG (Retrieval-Augmented Generation) test environment, where each pair of documents differs in exactly one factor. Methodological controls included:

  • Brand anonymization — eliminating the preference effect toward known brands
  • Balanced source ordering — controlling for position bias
  • 18 tested content factors — from relevance and formatting to credibility and timestamp

A total of 252,000 experiments across six LLMs were conducted, making this one of the most extensive empirical studies of the GEO phenomenon.

Which Content Factors Influence Citation and in What Order?

The study established the following priority order for GEO practice:

  1. Topical relevance and position in the list — by far the strongest factor; the document must be directly relevant to the query and as close to the top of retrieved results as possible
  2. Explicit pricing information and fresh timestamp — consistently provide an additional citation advantage
  3. Completeness and trust signals — smaller but measurable contribution
  4. Formatting — minimal impact on citation, contrary to widespread belief

Researchers published a reproducible evaluation protocol and a prioritized GEO checklist for practitioners, pilot-tested at Sprinklr with positive user feedback.

Frequently Asked Questions

What is GEO (Generative Engine Optimization) and why is it important?
GEO is the practice of optimizing content for citation in AI search engines (ChatGPT, Perplexity, Gemini), which generate answers from retrieved pages but cite only a small number of sources. Citation is the new visibility metric replacing classic SEO ranking.
Which factor has the greatest impact on citation in AI search engines?
Topical relevance and position in the list of retrieved documents have the greatest impact, with explicit pricing information and a fresh timestamp providing a consistent additional advantage.
How many experiments were conducted in this study?
The study involved 252,000 experiments across six LLMs, comparing documents that differ in exactly one of 18 tested content factors.

📬 AI news in your inbox

A daily digest built your way — pick topics, sources and cadence. One-click unsubscribe.