AI SEO: AEO vs GEO vs SEO, and What Actually Changed with AI

What Is AI SEO?

AI SEO is the work of getting a site found, used and cited by AI search: Google's AI Overviews and AI Mode, ChatGPT search, Perplexity, Microsoft Copilot and Gemini. Answer engine optimization (AEO), generative engine optimization (GEO), large language model optimization (LLMO) and AI optimization (AIO) are other names for the same job.

AI SEO starts from the same basics as classic SEO: a site that search engines can reach, index and trust. What's different is the prize. A page competes to be named as a source inside a written answer, not for a spot in a list of links.

"AI SEO" is also used for something else: using AI tools to do SEO work, such as AI keyword research or AI-drafted content. That is a question of who does the work, which the Agentic SEO guide covers as AI-assisted and agentic SEO. This guide is about what the work is aimed at.

Two more meanings sit outside this guide. Making a site easy for AI agents that browse and buy on someone's behalf is its own subject, and the Agentic SEO Commerce & Product Discoverability guide goes deeper on it. And GEO here has nothing to do with geography: local search in AI answers belongs to the Local SEO for AI Search guide.

AEO vs. GEO: Two Names for the Same Job

AEO and GEO are competing names for one practice: making content visible in AI-generated answers. Google's own guide treats them as a pair, describing both as "terms you may see used to describe work specifically focused on improving visibility in AI search experiences."

Wikipedia's entry on generative engine optimization notes that the terms, along with LLMO, AIO and AI SEO, have no agreed definitions and are "frequently used interchangeably". Profound, a company that sells AI visibility tracking, argues outright that AEO and GEO are "the same thing".

Some ranking pages draw a line anyway: AEO for featured snippets and voice assistants, GEO for chatbots. In this guide's reading, no engine draws that line. None of the engine documentation cited here separates the two, and Google files both under one description.

GEO entered the research literature through a 2023 paper by researchers led by Princeton University, later presented at the 2024 Conference on Knowledge Discovery and Data Mining (KDD). Venture firm Andreessen Horowitz (a16z) made the marketing case for the term in its 2025 essay "GEO over SEO".

What Changed in Search: From Ranked Links to Cited Answers

Search changed from handing back a ranked list of links to handing back a written answer, built from pages the engine retrieves and cites. So a page now competes for something new: not only a position on a list, but a place among the few sources an answer draws on.

	Classic search	AI search
What people ask	Shorter queries	Longer, more specific questions, often with follow-ups
What comes back	A ranked list of links	A written answer that cites a few sources
How a page gets used	Ranked and shown as a link	Retrieved and used as evidence for the answer, which may involve many related searches run at once
What a page competes for	A position on the list	Being one of the sources cited
How success shows up	A click	A citation or mention, often with no click

Microsoft's Bing team put the shift in two sentences in May 2026: "Search indexing was built to help humans decide what to read. Grounding indexing is being built to help AI systems decide what to say."

Two terms from Google's own guide explain how this works. Retrieval-augmented generation (RAG), which Google also calls grounding, means the AI model looks up relevant pages at the moment of the question and builds its answer from them, rather than relying only on what it learned in training.

Query fan-out means the engine splits one question into several related searches, runs them at once and combines the results. Google describes AI Mode as "issuing a multitude of queries simultaneously on your behalf." The How AI Search Engines Work guide goes deeper on how each engine retrieves, ranks and cites.

What longer questions and fan-out mean for keyword research is the subject of the SEO Content Fundamentals guide.

What the table can't show, in this guide's reading, is how much harder the edges are. On a list, the eighth result is still visible and clickable. In an answer, an uncited page is simply absent, and a cited page may still get no visit: the answer already did the click's job, the pattern usually called zero-click search.

The two formats also coexist. Google says AI Overviews appear only when its systems judge them "additive to classic Search", and that they "often don't trigger".

Is AI SEO Still SEO? What Google, Microsoft, OpenAI and Perplexity Say

Mostly, yes: Google treats optimizing for its AI features as ordinary SEO, OpenAI says sites that block its search crawler aren't shown in ChatGPT search, and Perplexity runs a separate search crawler. Where the engines differ is whether the work on top of those basics deserves a new name.

Engine	What it says	What that means for a site
Google (AI Overviews, AI Mode)	"From Google Search's perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO." Its AI features are "rooted in our core Search ranking and quality systems".	A page must be indexed and eligible to show with a snippet, and the site must be included in Search Console's "Search generative AI" setting.
Microsoft Bing	Names GEO outright, as "the practice of understanding how content participates in AI-driven experiences", and says visibility "is no longer defined only by rankings or clicks."	Being used in answers and citations is treated as its own thing to optimize and measure, with its own guidance.
OpenAI (ChatGPT search)	Runs OAI-SearchBot to surface sites in ChatGPT search, separate from GPTBot, which collects training data.	A site that blocks OAI-SearchBot is not shown in ChatGPT search answers, whatever it decides about GPTBot.
Perplexity	PerplexityBot is "designed to surface and link websites in search results on Perplexity" and "is not used to crawl content for AI foundation models."	Allowing PerplexityBot is a search visibility choice, not a training one.

Read side by side, the positions agree on more than the naming fight suggests. The entry requirements are classic SEO: crawl access (for ChatGPT search and Perplexity, that includes their own search crawlers), indexing and quality. Crawling and indexing belong to the Technical SEO Fundamentals guide.

What changed is the competition and the scoreboard. Bing's definition of GEO, built around answers and citations rather than only rankings and clicks, describes that new scoreboard. Forrester analyst Nikhil Lai drew the same line in November 2025: "AEO is significantly, but not fundamentally, different from SEO."

The strongest case for a new discipline is a16z's "GEO over SEO" essay, which argues that being referenced by language models replaces ranking as the thing to win. It's an investor's thesis, not engine documentation, and it reaches past search into business models, contrasting ad-funded search with subscription AI assistants.

AI search didn't replace the basics of SEO: a page still has to be crawlable, indexed and trustworthy before any engine will use it. What changed is the contest and how it's scored. Pages now compete to be cited inside an answer, and success shows up as citations and impressions, often without a click.

What makes a page trustworthy to Google and to AI answers is the subject of E-E-A-T in SEO.

What Google Says You Don't Need for AI Search

Google's guide names several things sites don't need to appear in its AI features, each owned by a more specific topic:

• New machine-readable files such as llms.txt. Google Search "doesn't use them". These files belong to the AI SEO Crawlers & Agent Accessibility guide.
• Splitting content into chunks or rewriting it for AI. Google says neither is needed. Formatting is where the On-Page SEO & AI Answers Optimization guide picks up.
• Special markup. "Structured data isn't required for generative AI search, and there's no special schema.org markup you need to add." The Structured Data & AI Entity Optimization guide goes deeper on markup.
• Inauthentic mentions. Google says planted mentions don't help. Earned mentions belong to the Third-Party Presence & RAG Visibility guide.

Microsoft leans the other way: its October 2025 guidance on inclusion in AI answers says long walls of text make it "harder for AI to separate content into usable chunks", credits schema markup with helping AI systems understand content, and tells writers to "Make answers snippable".

The engines don't fully agree on how much AI-specific work pays off, and that disagreement is worth knowing rather than smoothing over.

What the GEO Research Measured, and What It Didn't

The GEO paper by researchers led by Princeton University (KDD 2024) found that some content changes raised a source's visibility in generated answers by up to 40%. That figure comes from a benchmark the authors built, GEO-bench, run on a simulated generative engine, and the paper says results varied by domain.

Evidence that writing affects citation, not a measured traffic gain. In this guide's reading, the paper shows that how a page is written can change whether it gets cited. It is not a measured lift on Google or ChatGPT, although the 40% figure is often repeated as a real-world result.

AI engines lean on earned media. A September 2025 preprint from the University of Toronto, not peer-reviewed at the time of writing, found AI search engines leaned heavily toward earned, third-party sources over brand-owned and social content, unlike Google. The engines also differed from each other.

A GEO statistic is only as good as what it measured. Before treating a number as a result a site can expect, check whether it came from a lab benchmark or from real search traffic, and on which engine.

Why AI SEO Matters Now: Scale, Clicks and New Reports

AI SEO matters now because AI answers reach billions of people a month, they change how often people click through to sites, and since 2026 site owners can see and control their presence in them in Google's and Bing's own tools.

Scale. Alphabet's June 2026 investor presentation reported that AI Overviews have "over 2.5 billion users each month" and that AI Mode, launched a year earlier, has passed 1 billion monthly users.

Clicks. How much AI answers cut clicks is openly disputed, and three widely quoted figures come from different kinds of source:

Source	Data from	Finding	Type of source
Pew Research Center	March 2025, browsing data from 900 US adults	Users clicked a result link on 8% of visits with an AI summary vs. 15% without, and a link inside the summary on 1%. 18% of searches showed a summary.	Independent research
Ahrefs	December 2025, 300,000 keywords	The top result's click-through rate, the share of searchers who click it, was 58% lower when an AI Overview showed, up from 34.5% in its April 2025 study.	Vendor research, method published
Google	Claim made August 2025	Total organic click volume "relatively stable" year over year, with "slightly more quality clicks".	Google's claim, no data published

In this guide's reading, the figures disagree partly because they count different things. Pew and Ahrefs measure clicks per search where an AI summary appears. Google reports total clicks across Search, which can hold steady even as each search sends fewer clicks, if the number of searches grows.

New reports. Search Console's generative AI performance report shows a site's impressions in AI Overviews, AI Mode and the generative AI features in Google Discover. Google rolled it out to all sites worldwide on August 31, 2026.

A companion "Search generative AI" control lets a site stay included, the default, or exclude its content, which also stops that content being used to ground AI answers. Bing announced AI performance reporting in Bing Webmaster Tools in February 2026, calling it "a first window into a broader GEO toolset".

What each of these reports counts, and what it leaves out, is the subject of SEO Measurement.

These reports matter because they come from the engines themselves. Google warns that "No third-party tool has access to our internal ranking or AI systems." Tracking these numbers over time is where the AI Visibility, Mentions & Prompt Monitoring guide picks up.

In this guide's view, being able to see and control a site's presence in AI answers makes AI search measurable, controllable work, not speculation about a future channel.

Is SEO Dead?

SEO isn't dead: Alphabet reported in June 2026 that search queries had reached an all-time high, and Google's AI features draw on the same ranking systems as its classic results. Earlier, in Pew's March 2025 panel, more than four in five searches showed no AI summary at all.

The click studies show fewer clicks per search where an AI summary appears, not the end of search traffic.

What's Left Open: How Engines Pick Sources, Strategy and What to Let AI Use

Knowing what AI SEO is leaves three questions open: how each engine chooses the sources it cites, where AI search fits a brand's strategy, and what a site should let AI systems use.

The first is the natural next step from the retrieval and query fan-out definitions above. The AI SEO Strategy & Positioning guide goes deeper on the second.

The third, including whether to exclude a site from AI features and how search use differs from training use, belongs to the Governance, Compliance & Risks of AI Use guide. Acting on any of it is separate work, done on a real site with its own data.

Sources

Sourced claims:

• Optimizing your website for generative AI features on Google Search, Google Search Central. Cited for Google's description of AEO and GEO, its "still SEO" position, AI features rooted in core ranking systems, eligibility requirements, the list of things sites don't need, and the warning about third-party tools.
• AI features and your website, Google Search Central. Cited for AI Overviews appearing only when additive to classic Search, and often not triggering.
• Top ways to ensure your content performs well in Google's AI experiences on Search, Google Search Central Blog, May 21, 2025. Cited (paraphrased) for people asking longer, more specific questions and follow-ups in AI search.
• AI in Search: Going beyond information to intelligence, Google, May 20, 2025. Cited for AI Mode's query fan-out.
• Elevating the Role of Grounding on the AI Web, Bing Search Blog, Jordi Ribas, February 2026. Cited for Bing naming and defining GEO, visibility no longer being defined only by rankings or clicks, and Bing Webmaster Tools AI performance reporting as "a first window into a broader GEO toolset".
• Evolving role of the index: From ranking pages to supporting answers, Bing Search Blog, May 6, 2026. Cited for the search indexing vs. grounding indexing quote.
• Optimizing Your Content for Inclusion in AI Search Answers, Microsoft Advertising blog, October 8, 2025. Cited for Microsoft's guidance that walls of text make content harder for AI to split into usable chunks, that schema markup helps AI systems understand content, and that answers should be "snippable".
• Overview of OpenAI Crawlers, OpenAI. Cited for OAI-SearchBot vs. GPTBot and for sites that block OAI-SearchBot not being shown in ChatGPT search answers.
• Perplexity Crawlers, Perplexity. Cited for what PerplexityBot does and that it isn't used to crawl content for AI foundation models.
• SEO's Hype-Fueled Move To The Center Of The Marketing Mix, Forrester, Nikhil Lai, November 24, 2025. Cited for "AEO is significantly, but not fundamentally, different from SEO."
• GEO over SEO, a16z, May 28, 2025. Cited as the investor case for GEO as a new discipline, with being referenced replacing ranking, and for its contrast between ad-funded search and subscription AI assistants.
• GEO: Generative Engine Optimization, Aggarwal et al., KDD 2024. Cited for GEO-bench, the "up to 40%" visibility result on a simulated generative engine, results varying by domain, and the term's entry into the research literature.
• Generative Engine Optimization: How to Dominate AI Search, Chen et al., University of Toronto, September 2025 (preprint). Cited for AI search engines' bias toward earned media over brand-owned and social content, and differences between engines.
• Generative engine optimization, Wikipedia. Cited only to show the terms lack agreed definitions and are used interchangeably.
• AEO vs. GEO, Profound, June 29, 2025. Cited as a vendor's argument that AEO and GEO are the same thing.
• Alphabet investor presentation, June 2026, Google, June 3, 2026. Cited for AI Overviews' and AI Mode's monthly users, and search queries at an all-time high.
• Google users are less likely to click on links when an AI summary appears in the results, Pew Research Center, July 22, 2025. Cited for the 8% vs. 15% and 1% click figures and for 18% of searches producing an AI summary.
• Update: AI Overviews Reduce Clicks by 58%, Ahrefs, February 4, 2026, and the earlier April 2025 study. Cited for the 58% and 34.5% click-through rate figures.
• AI in Search is driving more queries and higher quality clicks, Google, Liz Reid, August 6, 2025. Cited for Google's claim that organic clicks are "relatively stable" with "slightly more quality clicks".
• Introducing Search Generative AI performance reports in Search Console, Google Search Central Blog, June 3, 2026, and the report's help page. Cited for what the report shows and its worldwide rollout on August 31, 2026.
• Search generative AI control, Search Console Help. Cited for the include (default) and exclude setting, and exclusion also stopping use in grounding AI responses.

Argued, not sourced (this guide's reasoning from the sources above):

• The core claim: AI SEO shares classic SEO's entry requirements, with a different competition and scoreboard.
• That no engine draws a line between AEO and GEO.
• The reading of the GEO paper's 40% as benchmark evidence rather than a field result.
• Why the click figures disagree, and that an uncited page is absent from an answer.
• The conclusion that AI search is now measurable, controllable work.