SearchVisibility.wikiVisibility in search and AI answers

Generative engine optimization (GEO)

From Search Visibility Wiki · By the · Published · Last reviewed · How we research

Generative engine optimization (GEO) is the practice of improving how often and how prominently content is used, cited or named in answers produced by generative AI search systems such as ChatGPT search, Perplexity, Gemini and Google AI Overviews. The term was introduced in a November 2023 research paper (arXiv 2311.09735) that reported visibility gains of up to 40% in a simulated engine; a later independent benchmark found most of those tactics did not improve citation ranking.

Where does the term GEO come from?

“GEO” was coined in GEO: Generative Engine Optimization by Pranjal Aggarwal, Vishvak Murahari, Tanmay Rajpurohit, Ashwin Kalyan, Karthik Narasimhan and Ameet Deshpande, first posted to arXiv on November 16, 2023.[1] The v1 paper lists affiliations with Princeton University, Georgia Tech, the Allen Institute for AI and IIT Delhi.[2] It was published at KDD ’24, the 30th ACM SIGKDD Conference, in Barcelona in August 2024.[3] See the GEO paper explained for full detail.

The authors define generative engines as systems that retrieve sources and use language models to write a response “grounded on the sources,” with inline attribution, and GEO as a creator-side framework “to aid content creators in improving their content visibility in generative engine responses.”[1]

What did the original GEO paper find?

The engine in the main experiments was a simulation: the top five Google results for each query fed to gpt-3.5-turbo.[1]

Do GEO tactics hold up in later research?

Only partly. C-SEO Bench (Puerto et al., 2025) tested the GEO paper’s eight content-rewriting methods plus two new ones across six domains and four models (gpt-4o-mini, Claude 3.5 Haiku, o3, o4-mini), measuring change in citation rank.[4] It found that “most current C-SEO methods are not only largely ineffective but also frequently have a negative impact on document ranking.” Of 54 cases, only three showed significant gains, and none of those used the original GEO methods. The Statistics method lowered rankings in 19 of 24 settings.[4] Meanwhile, moving a document to the top of the model’s context produced far larger gains than any rewrite, which the authors read as evidence that “traditional SEO strategies remain critical.”[4] Details: GEO replications and critiques.

The two papers measure different things: GEO mainly measured how many words of the answer relate to a source; C-SEO Bench measured how early the source is cited. The C-SEO authors argue the results “do not contradict each other” once the metric difference is accounted for.[4]

What do Google and other engines say about GEO?

Google’s guide addresses the term directly: “From Google Search’s perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO.” It adds that many suggested AEO and GEO “hacks” aren’t effective, and lists things you can ignore for Google, including llms.txt files, “chunking” content, rewriting content just for AI and seeking inauthentic mentions.[5] OpenAI and Perplexity document crawlers that must be allowed for a site to appear in their search answers (OAI-SearchBot, PerplexityBot).[6][7]

What does GEO practice look like in 2026?

A defensible GEO program, based on the evidence above, looks like this:

  1. Access: allow the search crawlers of the engines you care about (Googlebot, Bingbot, OAI-SearchBot, PerplexityBot). See AI crawlers.
  2. Retrieval first: the strongest independent finding is that retrieval rank drives citation, so classic SEO is the base layer.[4]
  3. Citable content: clear answers, original information, sourced facts. Google says unique, compelling content will likely matter “more than any of the other suggestions” in its guide.[5]
  4. Brand corroboration: authentic third-party mentions; see brand visibility.
  5. Measurement: track citations and mentions on a fixed prompt set and read first-party reports. See measurement.

How is GEO different from AEO and SEO?

In practice the terms overlap heavily. AEO predates generative AI and was first used for featured snippets and voice answers; GEO names the same goal for LLM-generated answers. See AEO vs GEO vs SEO and GEO SEO.

Frequently asked questions

Who invented generative engine optimization?

The term was introduced by Aggarwal, Murahari, Rajpurohit, Kalyan, Narasimhan and Deshpande in a paper first posted to arXiv on November 16, 2023 and published at KDD 2024.

Does GEO really increase AI visibility by 40%?

The 40% figure is the paper’s best result on its own word-count-based metric in a simulated engine using gpt-3.5-turbo. An independent 2025 benchmark measuring citation rank found most of the same tactics ineffective or harmful.

Is GEO just SEO?

Google says optimizing for its generative AI features is “still SEO.” Research also finds retrieval rank dominates citation. GEO adds measurement of citations and mentions, and attention to engines beyond Google.

References

Sources accessed October 14, 2026 unless another date is given. See all sources.

  1. “GEO: Generative Engine Optimization”. arXiv (Aggarwal, Murahari, Rajpurohit, Kalyan, Narasimhan, Deshpande). v1 submitted November 16, 2023; published at KDD ’24.
  2. “GEO: Generative Engine Optimization (v1 PDF, author affiliations)”. arXiv. v1, November 16, 2023.
  3. “GEO: Generative Engine Optimization”. Proceedings of the 30th ACM SIGKDD Conference (KDD ’24). August 2024.
  4. “C-SEO Bench: Does Conversational SEO Work?”. arXiv (Puerto, Gubri, Green, Oh, Yun). June 2025; NeurIPS 2025 Datasets & Benchmarks.
  5. “Optimizing your website for generative AI features on Google Search”. Google Search Central. accessed October 14, 2026.
  6. “Overview of OpenAI Crawlers”. OpenAI. accessed October 14, 2026.
  7. “Perplexity Crawlers”. Perplexity. accessed October 14, 2026.
Categories: Core concepts