AEO and SEO Glossary: Key Terms for Website Owners

Search is changing. Alongside the familiar list of blue links, people now get written answers from AI systems — Google's AI Overviews, ChatGPT, Perplexity, Microsoft Copilot — and those answers cite websites as sources. A new vocabulary has grown around getting your site mentioned in them, and much of it is unfamiliar even to people comfortable with SEO.

This glossary explains the terms you are most likely to meet in plain language. It covers both the newer AI search vocabulary and the classic SEO terms that still matter, because in practice the two overlap a great deal. You do not need to read it to start to finish — use the contents list to jump to the term you came for.

Where a term is disputed in the industry, or where a tactic has no proven effect, this article says so. Some widely repeated advice about AI search is not supported by what the search platforms themselves publish, and knowing the difference will save you wasted effort.

AI Search Basics

Start here. These ten terms underpin everything else in this glossary.

Term	What it means
Answer engine	Any system that replies to a question with a written answer instead of a list of links. Google's AI Overviews and AI Mode, ChatGPT search, Perplexity, Claude, and Microsoft Copilot are all answer engines. Most of them cite the websites they drew the answer from.
AEO (Answer Engine Optimization)	The practice of making your website easy for answer engines to find, understand, trust, and cite. Where traditional SEO aims to win a click from a list of links, AEO aims to win a mention and a citation inside an answer the AI writes.
GEO, LLMO, AI SEO	Other names for roughly the same practice. GEO stands for Generative Engine Optimization; LLMO for Large Language Model Optimization. The industry has not settled on one term, and you will see all of them used interchangeably. They do not describe meaningfully different activities.
AI Overviews	Google's AI-written summary that appears above the usual search results, with links to the sources it used. It draws on Google's existing search index, so there is no separate place to submit your site.
AI Mode	Google's separate conversational search experience, where you can ask follow-up questions. Google has said AI Mode may use different models from AI Overviews and show a different set of links, so the two are worth thinking about separately.
Citation	The source links an AI answer attaches to something it says. Being cited is the AEO equivalent of ranking — it is how people find their way from the answer back to your site.
Mention	Your business being named in an answer without a link back. A mention still builds awareness, but it sends no traffic. Mentions and citations are counted separately by most tools, and it is easy to confuse the two when reading a report.
Zero-click search	A search that ends without the person visiting any website, because the answer appeared on the results page. Zero-click searches have been rising for a decade — they were already common before AI answers arrived, and AI has accelerated the trend rather than started it.
Hallucination	A confident, fluent, but factually wrong statement from an AI system. For a business, this matters practically: an AI answer can state the wrong price, the wrong opening hours, or an outdated policy. Keeping accurate information clearly published on your own site is part of the defense.
Share of voice	How often does your business appear in AI answers across a set of questions, compared with your competitors. Sometimes called AI share of voice or mention share.

How Do AI Answers Get Built?

You do not need to understand the technology in depth, but a rough picture of how an answer is assembled explains why some advice works and some does not.

Term	What it means
Retrieval	The step where the system searches for and collects source pages before writing anything. This is the gate: if your page is never retrieved, nothing else about it matters — however well written it is.
Grounding	Tying an AI's answer to specific source documents it has just retrieved, rather than letting it answer from memory alone. Grounded answers are the ones that carry citations.
Index	The searchable store of web pages a system looks through when answering. Google's AI features use Google's ordinary search index, which is why being indexed in Google Search is the starting point rather than a separate task.
Query fan-out	Where the system quietly turns your one question into several related searches and looks at the results of each. Google has confirmed its AI features do this. It explains why a page can be cited in an answer even though it does not rank for the question as the person typed it.
Passage	A section of a page that gets picked up and used on its own, rather than the whole page. Google has confirmed it identifies individual passages within pages. In practice this is why sections that make sense when read out of context tend to do better.
Prompt set	A fixed list of questions you ask AI assistants repeatedly, so you can track how often your business appears over time. It is the AI equivalent of a keyword list in rank tracking.
Citation decay	A page being cited for a while, then dropping out as answers refresh and sources change. A large share of cited sources turn over between one answer and the next, so losing a citation is not necessarily a sign that anything is wrong.
Snippet	The short piece of page text shown below a search result. Your page needs to be eligible to show a snippet before it can be used in Google's AI features.

Because answers vary between runs, asking an AI about the same question twice and comparing the results tells you truly little. If you want to measure your visibility, use the same set of questions, check them repeatedly over several weeks, and look at the pattern rather than any single result.

AI crawlers and controlling access

AI companies use automated programs, called crawlers or bots, to collect web pages. The single most useful thing to understand in this whole glossary is that these bots do different jobs, and blocking them has very different consequences depending on which job the bot does.

Term	What it means
Crawler (bot)	An automated program that visits and reads web pages. Search engines and AI companies both use them. Each one identifies itself with a short name called a user agent.
User agent	The name a bot gives when it visits your site — for example Googlebot, GPT Bot or Perplexity Bot. You use these names to write rules about which bots may visit which parts of your site.
Training crawler	A bot that collects pages to help train future AI models. Blocking one does not affect whether you appear in AI answers today. It only keeps your content out of future model training.
Retrieval crawler	A bot that collects pages for the index of an AI product searches when answering a question. Blocking one of these does remove you from that product's answers and citations. This is the kind with a genuine cost.
User-triggered fetcher	A bot that fetches a single page because a person specifically asked about it. Several AI companies have said that site rules may not apply to this kind of visit, because a human requested it.
robots.txt	A plain text file at the root of your site that tells crawlers which parts they should not visit. It is a request rather than a lock — well-behaved crawlers follow it, but it does not physically prevent access. Anything you genuinely need to keep private needs a password, not a robots.txt rule.
Google-Extended	A setting you can add to robots.txt that controls whether Google may use your content to train and support its Gemini AI models. It does not remove you from AI Overviews, and Google has said it has no effect on your search rankings.
Search generative AI control	A setting in Google Search Console that excludes your site from AI Overviews, AI Mode, and AI features in Google Discover. Google has said it is not used as a ranking signal for the rest of Search. This is the setting that controls AI Overviews — robots.txt and Google-Extended do not.
noindex	An instruction telling search engines not to include a page in search results at all. It is a blunt tool: it removes the page from ordinary search as well as from AI features.
llms.txt	A proposed file listing the pages on your site you would most like AI systems to read. It is a community proposal rather than an agreed standard, and no major AI company has publicly committed to using it. Google has stated that Google Search does not use these files. Publishing one does no harm, but do not expect it to change your visibility, and do not treat it as a way of controlling AI access — it grants and blocks nothing.

Warning: A common and costly mistake: if you add a rule for one named bot to your robots.txt file, that bot stops following the general rules you set for all bots. Anything you still want to apply must be repeated inside the named bot's own section.

Warning: Blocking a retrieval crawler removes your site from that product's AI answers. Before blocking anything, check which job the bot does — several widely shared block lists mix training and retrieval bots together, and following one can quietly cost you the visibility you meant to keep.

Structured data and your business identity

Structured data is extra code that describes what a page is about in a form machines can read reliably. It is worth adding, but it is often oversold — see the note at the end of this section.

Term	What it means
Structured data (schema markup)	Code added to a page that labels what the content is — that this is a product; this is its price; this is the business behind the site. It does not change what visitors see.
schema.org	The shared vocabulary that structured data is written in, used by all the major search engines.
JSON-LD	The format Google recommends for structured data: a small block of code placed in the page, separate from the visible content.
Rich result	An enhanced search listing built from structured data — star ratings, event dates, product prices shown directly in the results. Also called a rich snippet.
Organization schema	Structured data describing your business: name, website, logo, address, contact details, and links to your official profiles. This is the most valuable type for most small businesses, because it helps search, and AI systems recognize your business as a distinct, known thing.
Local Business schema	A more specific version of the above for businesses with a physical location or service area.
sameAs	A property inside structured data that lists other pages which clearly identify your business — your Wikipedia entry if you have one, your official social profiles. It helps systems confirm that different mentions across the web refer to the same business.
Entity	A distinct thing that searches and AI systems recognize and store information about — a business, a person, a product. Becoming a recognized entity is a matter of being described consistently in enough places.
Knowledge panel	The information box about a business or person that sometimes appears beside search results. These are generated automatically; you cannot apply for one, but if one exists for your business, you can claim it and suggest corrections.
E-E-A-T	Experience, Expertise, Authoritativeness and Trustworthiness — four qualities described in Google's guidelines for the people who assess search quality. Google has stated that trust is the most important of the four, and that E-E-A-T is not a ranking factor. There is no E-E-A-T score.
NAP	Name, Address and Phone number — shorthand for keeping those three details identical everywhere your business is listed. Inconsistency makes it harder for systems to tell that the listings refer to one business.

Important: Structured data is worth adding rich results to help systems recognize your business. But Google has said plainly that no special structured data is needed to appear in its AI features, and independent testing has repeatedly found that facts placed only in structured data are not picked up by AI assistants. If a fact matters, put it in the visible text of the page as well.

Classic SEO terms you will still see

AI search has not replaced search engine optimization. Google's own guidance describes optimizing its AI features as still being SEO, and most of what helps one helps the other. These are the established terms you will keep meeting.

Term	What it means
SEO (Search Engine Optimization)	Making your site easier for search engines to find, understand, and rank.
Crawling	A search engine bot fetching your pages.
Indexing	The engine storing a fetched page so it can appear in results. Being crawled and being indexed are not the same thing — a page can be visited and still not included.
Rendering	Running a page's JavaScript to see the finished content, as a browser would. Content that only appears after scripts run can be missed by some systems.
Keyword	A word or phrase people type into a search engine, to allow them to find what they are looking for.
Backlink	A link to your site from another website.
Internal linking	Links between pages on your own site. They help both visitors and crawlers find their way around.
Anchor text	The visible, clickable words of a link. Descriptive anchor text helps everyone. All websites should avoid phrases like "click here".
Canonical URL	The version of a page you have declared as the main one, when the same content is reachable at more than one address.
XML sitemap	A machine-readable list of your website's site pages, submitted to search engines so they can discover everything. It aids discovery; it does not grant or restrict permission.
Featured snippet	The answer box that quotes a single page at the top of some Google results. It has existed for years and is a different feature from AI Overviews.
Core Web Vitals	Google's measurements of how quickly and smoothly a page loads and responds.
Google Business Profile	Your business listing in Google Search and Maps. For a local business serving walk-in customers, keeping this accurate and complete will usually do more for visibility than any amount of content optimization.

Measuring what is working

Measurement for AI search is still immature, and it is worth knowing where the gaps are before reading reports.

Term	What it means
Google Search Console	Google's free tool showing how your site performs in Google Search — which queries you appear for, how often, and how many people click.
Impression	Your page being shown in results. Search Console now reports impressions for AI Overviews and AI Mode.
Click-through rate (CTR)	Clicks are divided by impressions. Note that Search Console reports impressions for AI features but not clicks, and counts AI clicks inside the overall web search totals — so a click-through rate for AI features specifically cannot currently be worked out.
Organic traffic	Visitors arriving from unpaid search results. Google counts clicks from AI Overviews and AI Mode here, rather than separating them out.
Referrer	The page a visitor arrived from. It is how analytics knows a visit came from ChatGPT or Perplexity rather than from a search engine.
UTM parameter	A tag added to the end of a link that tells analytics where the click came from. ChatGPT, for example, adds one identifying itself, though not consistently.
Direct traffic	Visits where no referrer was passed, so the source is unknown. Links opened from inside a mobile app often land here — which means some of your AI traffic may be sitting in Direct rather than where you expect it.
AI Assistant channel	Grouping in Google Analytics 4 for visits from ChatGPT, Gemini, Copilot, and similar tools. Worth knowing it does not include Google's AI Overviews or AI Mode, which Google counts as organic search instead.

Common misunderstandings

A few claims circulate widely enough to be worth correcting, because acting on them wastes time.

Term	What it means
"You need a special file or markup for AI"	Google has stated the opposite: no special files, markup, or structured data are needed to appear in its AI features. Clean, crawlable, genuinely useful pages are the requirement.
"You need to chunk your content for AI"	Google has said it identifies the relevant sections of a page itself, and that restructuring content into chunks for AI is not necessary. Writing sections that stand on their own is good practice for readers anyway, but it is not a technical requirement.
"Longer articles get cited more"	Analysis of large numbers of AI answers has found almost no relationship between word count and being cited. More than half of the cited pages are under a thousand words. Write what the topic needs.
"AI search has replaced Google"	Not yet, and not close. Across large samples of websites, traffic arriving from AI assistants is still around one percent of all visits, while Google sends very substantially more. The bigger practical change is to the clicks you already get, not a new channel replacing the old one.
"Blocking AI bots protects my content without cost"	It depends entirely on which bot. Blocking a training bot costs you nothing in visibility. Blocking a retrieval bot removes you from that product's answers.

Where to start

If the vocabulary is new, the useful first steps are ordinary ones:

Check that your site is being indexed in Google Search, using Google Search Console.
Check your robots.txt file is not accidentally blocking crawlers you want — security plugins and content delivery networks sometimes add rules you did not set.
Make sure the facts that matter — what you sell, where you are, what you charge, how to reach you — appear in the visible text of your pages, not only in code or images.
Keep your Google Business Profile accurate if you serve customers locally.
Add Organization structured data describing your business if your SEO plugin does not already.

Did you find this article helpful?