What Gets Content Cited in AI Search

TL;DR

Content gets cited more in AI search when it names sources, quotes directly and gives specific statistics, each raising visibility up to 40% in a Princeton and IIT Delhi study.

  • Keyword stuffing did worse than no edits
  • The size of the gain shifts by topic
  • Results held on the live Perplexity tool

Ask an AI tool such as ChatGPT or Perplexity a question, and the answer often comes with sources attached. Which sources get picked is not random. Researchers at Princeton and the Indian Institute of Technology Delhi tested nine ways of writing the same content. They ran each version through 10,000 real searches and measured which changes made an AI answer cite the page.

For each search, the researchers pulled the five sources a normal search engine would return. They fed those sources to a system that writes answers and cites them the way AI search tools do. Before rewriting anything, they counted how many words of each answer traced back to each source, and how prominently it appeared. That score is the source’s visibility. Then they rewrote each source one technique at a time and reran the searches.

The three techniques that stood out were naming the source behind a claim, adding a direct quotation and backing a claim with a specific statistic. Each raised visibility by up to 40% over unedited text. The paper calls this kind of writing for AI answers Generative Engine Optimization.

An AI search engine reads several pages and writes a fresh answer, choosing which source to credit for each sentence. A fact stated on its own competes with thousands of similar sentences across the web. Each of the three techniques ties that fact to something specific, which gives the engine a passage to lift and credit.

Citing a source does not guarantee a place in an AI answer. The tests used one set of questions built by one research group, and the size of the gain shifts by topic. Statistics, for instance, helped most with questions about law and government, and less with others. Still, the results held when the researchers reran the techniques on the live Perplexity tool.

For two decades, search marketers leaned on keyword stuffing, the old trick of repeating a target phrase over and over. Older search engines counted keyword matches and ranked pages accordingly. A page packed with repeated phrases gives an AI engine nothing distinct to quote. In these tests, it scored below the same page with no changes at all.

Lead smarter with AI in just 5 minutes a week

Join 50K+ leaders getting our AI For Leaders newsletter every Wednesday: deep analysis, frameworks, and strategies.

No spam ever • Unsubscribe anytime

What Gets Content Cited in AI Search | Chief AI Leader