LLM SEO
LLM SEO is the practice of improving how large language model (LLM) products such as ChatGPT, Gemini, Perplexity and Google’s AI features find, cite and describe your content and brand. Where those products search the web before answering, LLM SEO is mostly SEO for their retrieval systems; where they answer from training data, it depends on how widely and consistently your brand is discussed on the web.
How do LLM products decide what to say?
There are two paths. Retrieval: the product searches an index at query time and grounds its answer in the results. Google says its generative AI features rely on retrieval-augmented generation (“also known as grounding”) over its Search index.[1] Model knowledge: the model answers from what it learned in training. Training crawlers are controlled separately from search crawlers, for example GPTBot versus OAI-SearchBot at OpenAI,[2] and Google-Extended for Gemini training and grounding in Gemini Apps.[3]
Which crawlers matter for LLM SEO?
| Token | Owner | Purpose (per vendor docs) |
|---|---|---|
| OAI-SearchBot | OpenAI | Surfaces sites in ChatGPT search; opted-out sites “will not be shown in ChatGPT search answers”[2] |
| GPTBot | OpenAI | Crawls content that may be used to train foundation models[2] |
| PerplexityBot | Perplexity | Surfaces and links sites in Perplexity search; “not used to crawl content for AI foundation models”[4] |
| Googlebot | Google Search including all Search features (so AI Overviews and AI Mode)[3][5] | |
| Google-Extended | Control token for Gemini training and grounding in Gemini Apps and Vertex AI; does not affect Search inclusion or ranking[3] |
What LLM SEO tactics are supported by evidence?
- Retrieval rank: in C-SEO Bench, being placed first in the model’s context improved citation rank far more than any rewriting method.[6]
- Brand mentions: branded web mentions correlated 0.664 with AI Overview brand visibility in Ahrefs’ 2025 study (correlation only).[7]
- Clear, sourced content: Google’s guidance favors unique, people-first, non-commodity content.[1]
What LLM SEO tactics are not supported?
- Keyword stuffing: little to no improvement even in the original GEO paper.[8]
- Adding statistics or quotes by rewriting: lowered or did not change citation rank in most C-SEO Bench settings.[6]
- llms.txt for Google: Google Search ignores it, though it is fine to keep for other systems.[1]
- Inauthentic mentions: Google says these are not as helpful as they might seem.[1]
Frequently asked questions
Is LLM SEO the same as GEO?
Effectively yes. LLM SEO, LLMO, GEO and AEO describe the same goal with different emphasis; GEO is the term used in academic research.
Should I block GPTBot?
That is a business decision about training use. Blocking GPTBot does not remove you from ChatGPT search, which is governed by OAI-SearchBot, according to OpenAI.
Does llms.txt help LLM SEO?
Google says its Search does not use llms.txt. Some other systems may read it, but no engine has documented it as a ranking or citation factor.
See also
References
Sources accessed October 14, 2026 unless another date is given. See all sources.
- “Optimizing your website for generative AI features on Google Search”. Google Search Central. accessed October 14, 2026.
- “Overview of OpenAI Crawlers”. OpenAI. accessed October 14, 2026.
- “List of Google's common crawlers (Googlebot, Google-Extended)”. Google for Developers. accessed October 14, 2026.
- “Perplexity Crawlers”. Perplexity. accessed October 14, 2026.
- “AI features and your website”. Google Search Central. accessed October 14, 2026.
- “C-SEO Bench: Does Conversational SEO Work?”. arXiv (Puerto, Gubri, Green, Oh, Yun). June 2025; NeurIPS 2025 Datasets & Benchmarks.
- “An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied)”. Ahrefs. May 26, 2025.
- “GEO: Generative Engine Optimization”. arXiv (Aggarwal, Murahari, Rajpurohit, Kalyan, Narasimhan, Deshpande). v1 submitted November 16, 2023; published at KDD ’24.