AI SEO, LLM SEO and GEO describe one job: getting a business named or cited in AI answers. The names differ in origin and in demand. In US Google Ads data for November 2026, "ai seo" is searched about 6,600 times a month, "generative engine optimization" about 4,400, "answer engine optimization" about 2,400 and "llm seo" about 7201.
How the four names compare
Nobody with standing defines AI SEO, LLM SEO or answer engine optimization. A definition of each from a standards body, a dictionary or a paper is NF. Two search engines do speak to the terms, and one paper introduced GEO. Table 1 sets the four names side by side, with US searches a month and cost per click from the same Google Ads data.
Table 1. Four names for the same job
| Term | Where it comes from or who defines it | Searches a month and cost per click | What people who use it mean |
|---|---|---|---|
| AI SEO | No standards body, dictionary or paper definition (NF) | 6,600 and $41.451 | An umbrella phrase; no source defines its scope (NF) |
| GEO, generative engine optimization | Introduced in a 2023 paper2; defined in Bing's guidelines3 | 4,400 and $35.041 | Bing: content eligibility for grounding and reference in AI responses3 |
| AEO, answer engine optimization | Spelled out in Google's guide as a term others use4; no definition of its own (NF) | 2,400 and $21.311 | Google: work on improving visibility in AI search experiences4 |
| LLM SEO | No standards body, dictionary or paper definition (NF) | 720 and $40.331 | No source defines it (NF) |
Answer engine optimization is searched more than three times as often as LLM SEO (2,400 / 720), which is why it has a row here. Three nearby phrases are searched too: "ai seo agency" about 1,600 times a month at $32.60 a click, "ai search optimization" about 1,300 at $52.23 and "llm optimization" about 720 at $27.601.
Marketers use the word GEO. The CMO Survey reports that 41.5% of 188 respondents said they use AI for GEO, a use it described to them as getting content to appear in AI-generated search results5. The CMO Survey reports a sample of marketing leaders at U.S. for-profit companies, 308 of 2,111 invited5.
What Google and Bing say about the terms
Google's guide to optimizing for generative AI features, last updated July 10, 2026, spells out both terms and says they describe work on visibility in AI search experiences4. It then states Google's position: from Google Search's perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO4. The guide describes the terms as ones you may see used and does not adopt them.
An earlier Google page, last updated December 10, 2025, says a page needs no technical requirements for AI Overviews or AI Mode beyond being indexed and eligible to be shown in Search with a snippet6. The same page says specific optimization is not required and that existing SEO fundamentals remain worthwhile6.
Bing goes further. Its Webmaster Guidelines define Generative Engine Optimization as work on content eligibility for grounding and reference in AI responses, and set it apart from SEO3. Microsoft's Bing blog post of February 12, 2026 describes GEO as the practice of understanding how content participates in AI-driven experiences7. The same guidelines say Bing and Copilot search experiences rely on the same crawling, indexing and ranking foundation as traditional search3. Microsoft defines AI grounding as using content as input to inform AI-generated responses, and says indexing and ranking in traditional search is not grounding8.
What the AI companies document
OpenAI, Anthropic, Perplexity and Google each publish the crawlers they run and what each is for. Table 2 gives each company's own description.
Table 2. Crawlers and tokens as their companies describe them
| Company | Name | Purpose, per the company |
|---|---|---|
| OpenAI | OAI-SearchBot | Surfaces websites in ChatGPT's search features9 |
| OpenAI | GPTBot | Crawls content that may be used to train its models9 |
| OpenAI | ChatGPT-User | User-initiated actions; robots.txt rules may not apply9 |
| Anthropic | ClaudeBot | Collects web content that could contribute to training10 |
| Anthropic | Claude-SearchBot | Indexes content for search optimization10 |
| Perplexity | PerplexityBot | Surfaces and links sites in Perplexity results; not used for AI foundation models11 |
| Perplexity | Perplexity-User | Fetches at a user's request; generally ignores robots.txt11 |
| Google-Extended | A token for Gemini training and grounding; no effect on Search12 |
Anthropic lists a third robot, Claude-User, for user-directed retrieval10. The controls differ by company. OpenAI says its settings are independent, so a site can allow OAI-SearchBot to appear in search while disallowing GPTBot to keep content out of training9. A site opted out of OAI-SearchBot is not shown in ChatGPT search answers but can still appear as a navigational link9. To be eligible, OpenAI's help page says to allow OAI-Searchbot and to confirm the host or content delivery network allows OpenAI's published IP addresses13. Anthropic says disabling Claude-SearchBot may reduce a site's visibility and accuracy in user search results10.
Google keeps its control inside Search. It says robots.txt directives for Googlebot are the control for how a site is crawled for Search, because AI is built into Search6. It lists nosnippet, data-nosnippet, max-snippet and noindex as the controls that limit what Search shows6, and its robots meta specification, last updated March 24, 2026, says nosnippet also keeps content from being used as direct input for AI Overviews and AI Mode14.
Microsoft added two preferences, AI Training and AI Grounding. Its FAQ says they were announced on October 30, 2026 and that Microsoft begins honoring them on February 1, 20278. AI Grounding cannot be set in robots.txt, only per page by meta tag or HTTP header, and both preferences default to Allow8. They replace NoArchive and NoCache for training and grounding8.
The llms.txt file is a proposal by Jeremy Howard, first published September 3, 2024 and last modified August 10, 202615. It describes a markdown file at a site root that gives brief background and links to LLM-friendly content15. Its own page calls the specification open for community input, with a GitHub repository hosting an informal overview15. Google says you do not need llms.txt or other AI text files to appear in Google Search, including its generative AI features, because Search does not use them4. Google adds that creating one for other services is fine and will neither harm nor help visibility or rankings in Google Search4.
What the evidence says the job involves
Whatever the work is called, three kinds of finding bear on it. Each comes from a different publisher with its own keyword set or sample, so we give each with its source and join none of them. Ahrefs† (an SEO software vendor), BrightEdge† (an SEO platform vendor) and DataForSEO† (a search data provider) measure their own panels.
Citations in AI answers overlap little with top rankings. Ahrefs† reports that 37.9% of URLs cited in Google AI Overviews also appeared within the first 10 blocks of the results for the same query, across 863,000 keywords and 4 million cited URLs16. In a separate study of 15,000 long-tail prompts, Ahrefs† reports that on average 12% of links cited by ChatGPT, Gemini and Copilot appeared in Google's top 10 for the same prompt, and 28.6% of Perplexity's cited URLs did17. BrightEdge† reports that about 17% of sources cited in AI Overviews also ranked in the organic top 10 between February and July 202518.
Assistants cite different sites and name different businesses. DataForSEO† reports that YouTube is first in Google AI Overviews with 51.9 million citations and 19th in ChatGPT with 211,00019, and says the lists are ranked separately and cannot be compared on absolute volume20. BrightEdge† reports that overlap in cited sources between pairs of five engines ranged from 16% to 59%, and overlap in named brands from 36% to 55%21. In our own study, we put 199 local questions to ChatGPT, Gemini and Perplexity on November 10, 202622. Only 4.1% of the businesses named were named by all three assistants and 79.7% were named by only one22. Of the websites cited, 88% were cited by only one assistant22. Each question was asked once, and answers through an API are a baseline, not what every signed-in user sees22.
Answers change between asks. In our repeat study, which put 63 local questions to the three assistants 5 times each, a business named in one answer was in another answer to the same question from the same assistant 55% of the time: 35.6% for ChatGPT, 50.5% for Gemini and 78.9% for Perplexity23. Ahrefs† reports a 70% chance that an AI Overview changed from one observation to the next, across more than 43,000 keywords over one month24. A separate Ahrefs† study of about 37,500 citations across five AI tools reports that at least half of the citations disappeared by the next day25.
What raises the odds of being named is less settled. The paper that introduced GEO reports that its methods raised a source's visibility in generative engine responses by up to 40%, that the effect varies across domains, and that content creators have little to no control over how generative engines display their content2. Ahrefs† reports that among 75,000 brands, branded web mentions showed the strongest correlation with AI Overview visibility at 0.664, against 0.218 for backlinks, and states that correlation is not causation26.
We go further than the sources, on our own judgment as people who do this work. What an assistant says about a business is largely dictated by what is distributed about that business online, and sites with more authority are trusted more. We also believe the one-sidedness of the record matters. A little-known business with the same thing said about it everywhere acquires that association quickly. A large one, with many opinions and facts about it in many places, changes slowly. That second point is our belief, and nobody we have read has measured it.
The size of the prize is a separate question. Ahrefs† reports that all AI chatbots combined sent 0.28% of total web traffic across 74,752 websites in March 202627. BrightEdge† reports that AI platforms accounted for 2.25% of search activity in September 2026, with Google at 86.3%28. The two measure different sets of sites in different months.
Which term to use and when
We think the scope matters more than the name, and a name tells a buyer little about what is included. On this site we use AI SEO for all of it, since it is the most searched of the four1. GEO is the term to use when you are reading Bing's guidelines or the 2023 paper, because both use it in a defined way2,3. Answer engine optimization appears in Google's guide next to GEO4. LLM SEO has the smallest demand of the four at about 720 searches a month1.
Ask any provider, whatever term it uses, to state in writing which assistants and which Google features it covers, which questions it will test, how many times and on which dates. The answers above change from one run to the next, so one test says little.
What we could not find
A definition of AI SEO, LLM SEO or answer engine optimization from a standards body, a dictionary or a paper is NF. The nearest evidence is a September 2026 IETF working group draft, which defines AI Training as using an asset to modify the learned parameters of a generative AI model29. It is about usage preferences and does not name any of the three terms.
A measure of whether any of this work causes a local service business to be named in AI answers is NF. The nearest evidence is the GEO paper's gain of up to 40% in its own tests2, and Ahrefs† says that its correlations are not causal26. Neither covers local service businesses. A study of AI citations for those businesses specifically is NF as well; the nearest evidence is that Ahrefs† reports AI Overviews appeared for 7.9% of queries categorized as local searches, against 22.8% of non-local queries30.
An independent test of whether an llms.txt file changes how often AI assistants cite a site is NF. The nearest evidence is Google's statement that Google Search ignores the file4.
Where our service fits
Our AI SEO work sits inside the plans. Visibility starts at $999 a month for one location, month to month, plus a $499 one-time onboarding fee, and includes a monthly AI answer check31. Managed marketing is $1,800 a month on every term and includes content written to be cited by AI assistants31. We promise no ranking and no mention, since the evidence above shows assistants change their answers.
People also ask
Is AI SEO the same as GEO?
The two names cover the same job, getting a business named or cited in AI answers, and no standards body defines either one. GEO has a source: Bing defines it as work on content eligibility for grounding and reference in AI responses3. AI SEO has no definition we could find (NF).
What is LLM SEO?
LLM SEO is a name for work aimed at having large language model assistants name or cite a business. No standards body, dictionary or paper defines it (NF). It is searched about 720 times a month in the US, far below AI SEO at about 6,6001.
What is answer engine optimization?
Answer engine optimization, or AEO, is a term Google's guide spells out and describes as one you may see used for work on visibility in AI search experiences4. Google adds that, from its perspective, optimizing for generative AI search is still SEO4.
Does Google treat GEO as different from SEO?
No. Google's generative AI guide, last updated July 10, 2026, says optimizing for generative AI search is optimizing for the search experience, and thus still SEO4. Bing takes the other position and defines GEO in its guidelines as a separate focus from SEO3.
Do I need an llms.txt file?
Google says you do not need one to appear in Google Search, including its generative AI features, because Search does not use it. Creating one for other services is fine and will neither harm nor help your visibility in Google Search4.
Which AI crawlers can a website allow or block?
Each company publishes its own. OpenAI runs OAI-SearchBot for search and GPTBot for training, and says the settings are independent9. Anthropic runs ClaudeBot, Claude-User and Claude-SearchBot10. Perplexity runs PerplexityBot and Perplexity-User11.
How often do people search for AI SEO, GEO and LLM SEO?
In US Google Ads data for November 2026, "ai seo" is searched about 6,600 times a month, "generative engine optimization" about 4,400, "answer engine optimization" about 2,400 and "llm seo" about 7201.
How to read this page: † a vendor’s own figure, NF not found, ( ) our arithmetic on cited figures, and a small number points to a source below.
Sources
- Technical Marketing. What service businesses search for: 2026 demand data. Technical Marketing, 2026-11. technicalmarketing.ai/reports/service-business-search-demand/ Our own data.
- Pranjal Aggarwal, Vishvak Murahari, Tanmay Rajpurohit, Ashwin Kalyan, Karthik Narasimhan, Ameet Deshpande. GEO: Generative Engine Optimization. arXiv, 2023. arxiv.org/abs/2311.09735
- Microsoft Bing. Bing Webmaster Guidelines. bing.com/webmasters, 2026-11. bing.com/webmasters/help/webmaster-guidelines-30fba23a
- Google Search Central. Optimizing your website for generative AI features on Google Search. developers.google.com, 2026-07-10. developers.google.com/search/docs/fundamentals/ai-optimization-guide
- The CMO Survey (Duke Fuqua School of Business, Deloitte, American Marketing Association). The CMO Survey Topline Report - 2026 (35th edition). cmosurvey.org, 2026-01. cmosurvey.org/wp-content/uploads/2026/03/The_CMO_Survey-Topline_Report Secondary source.
- Google Search Central. AI features and your website. developers.google.com, 2025-12-10. developers.google.com/search/docs/appearance/ai-features
- Microsoft Bing (Jordi Ribas). Elevating the Role of Grounding on the AI Web. blogs.bing.com, 2026-02-12. blogs.bing.com/search/2026/2/Elevating-the-Role-of-Grounding-on-the-AI
- Microsoft Bing. FAQ - New AI Preferences. bing.com/webmasters, 2026-10-30. bing.com/webmasters/help/new-ai-preferences-2t1cc16d
- OpenAI. Overview of OpenAI Crawlers. platform.openai.com, 2026-11. platform.openai.com/docs/bots
- Anthropic. Does Anthropic crawl data from the web, and how can site owners block the crawler?. privacy.claude.com, 2026-11. privacy.claude.com/en/articles/8896518-does-anthropic-crawl-data-from-
- Perplexity. Perplexity Crawlers. docs.perplexity.ai, 2026-11. docs.perplexity.ai/guides/bots
- Google Search Central. Google's common crawlers (Google-Extended). developers.google.com, 2026-10-30. developers.google.com/search/docs/crawling-indexing/google-common-craw
- OpenAI. Searching the web with ChatGPT. help.openai.com, 2026-11. help.openai.com/en/articles/9237897-chatgpt-search
- Google Search Central. Robots meta tag, data-nosnippet, and X-Robots-Tag specifications. developers.google.com, 2026-03-24. developers.google.com/search/docs/crawling-indexing/robots-meta-tag
- Jeremy Howard (llmstxt.org). The /llms.txt file, v2. llmstxt.org, 2024-09-03 to 2026-08-10. llmstxt.org/
- Louise Linehan. Update: 38% of AI Overview Citations Pull From The Top 10. Ahrefs, 2026. ahrefs.com/blog/ai-overview-citations-top-10/ Secondary source.
- Louise Linehan. Only 12% of AI Cited URLs Rank in Google's Top 10 for the Original Prompt. Ahrefs, 2025-07. ahrefs.com/blog/ai-search-overlap/ Secondary source.
- BrightEdge. AI Overviews at the One-Year Mark: Presence, Size, and What They're Citing. BrightEdge, 2025-02 to 2025-07. brightedge.com/resources/weekly-ai-search-insights/ai-overviews-one-ye Secondary source.
- DataForSEO. Search Insider: 2026 Edition, p. 21. DataForSEO, 2025-07 to 2026-09. Secondary source.
- DataForSEO. Search Insider: 2026 Edition, p. 48. DataForSEO, 2025-07 to 2026-09. Secondary source.
- BrightEdge. Why AI Engines Cite Different Sources but Recommend the Same Brands. BrightEdge, published 2026-04-24; dates of prompts not stated. brightedge.com/resources/weekly-ai-search-insights/ai-search-same-bran Secondary source.
- Technical Marketing. Who AI recommends locally: a study of local answers. Technical Marketing, 2026-11. technicalmarketing.ai/reports/ai-local-recommendations/ Our own data.
- Technical Marketing. Does AI recommend the same businesses twice?. Technical Marketing, 2026-11. technicalmarketing.ai/reports/ai-answer-consistency/ Our own data.
- Louise Linehan. AI Overviews Change Every 2 Days (But Never Change Their Mind). Ahrefs, 2025. ahrefs.com/blog/ai-overview-change/ Secondary source.
- Mateusz Makosiewicz. How Stable Are AI Citations? 10 Things We Learned Tracking 37,500 Citations. Ahrefs, 2026. ahrefs.com/blog/how-stable-are-ai-citations/ Secondary source.
- Louise Linehan. An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied). Ahrefs, 2025. ahrefs.com/blog/ai-overview-brand-correlation/ Secondary source.
- Mateusz Makosiewicz. AI Chatbot Traffic: What It Is, and How to Get More. Ahrefs, 2026-03. ahrefs.com/blog/ai-chatbot-traffic/ Secondary source.
- BrightEdge. BrightEdge AI Market Pulse. BrightEdge, 2026-09. brightedge.com/brightedge-ai-pulse Secondary source.
- IETF AI Preferences Working Group (Keller, Thomson). draft-ietf-aipref-vocab-08, A Vocabulary For Expressing AI Usage Preferences. datatracker.ietf.org, 2026-09. datatracker.ietf.org/doc/html/draft-ietf-aipref-vocab-08
- Ryan Law. What Triggers AI Overviews? 86 Factors and 146 Million SERPs Analyzed. Ahrefs, 2025-09. ahrefs.com/blog/ai-overview-triggers/ Secondary source.
- Technical Marketing. Pricing. Technical Marketing, 2026-11. technicalmarketing.ai/pricing/ Our own data.