SEO, GEO and AEO terms, in plain words

GEO (generative engine optimization)

Work that makes a business more likely to be named in answers written by AI tools such as ChatGPT, Gemini, Perplexity and Google's AI Overviews. It builds on SEO rather than replacing it.

In our audits: Across 1.5k+ business websites, the average AI visibility score was 68.6/100, against 83.4/100 for SEO.

AEO (answer engine optimization)

Shaping a page so a direct answer can be lifted from it: the answer comes first, headings are phrased as the questions people ask, and each section makes sense on its own.

In our audits: 35.7% of the sites we checked had a questions-and-answers section on any page we read.

LLMO, AIO and AI SEO

Other names for GEO and AEO. LLM optimization, AI optimization (AIO) and AI SEO describe the same work under newer labels, not a separate service.

AI Overview

The AI-written summary Google shows above the normal results for some searches, with links to pages that support it. A page has to be indexed and allowed a snippet to be one of those links.

Query fan-out

How Google says its AI Overviews and AI Mode may answer: by issuing several related searches across subtopics and combining what they find. A page can be cited for one sub-question it answers well.

AI visibility audit

A check of whether ChatGPT, Gemini, Perplexity and Google's AI answers can reach a website, understand it, and have reason to mention the business.

AI crawler

A program that fetches web pages for an AI company: to train its models, to build a search index, or to read a page a user asked about. Each one announces itself by name.

In our audits: 3.9% of the sites we checked shut out an AI crawler in robots.txt.

GPTBot

OpenAI's crawler for collecting training data. It is separate from OAI-SearchBot, the crawler that decides whether a site can appear in ChatGPT search.

In our audits: 14.4% of the sites we checked refused GPTBot.

OAI-SearchBot

OpenAI's crawler that finds pages to show in ChatGPT search. Blocking it keeps a site out of those answers; blocking GPTBot does not.

In our audits: 5.1% of the sites we checked refused OAI-SearchBot.

ClaudeBot, Claude-SearchBot and Claude-User

Anthropic's three crawlers. ClaudeBot collects content that may be used for training, Claude-SearchBot indexes pages for Claude's search results, and Claude-User fetches a page when a user asks Claude a question.

In our audits: 13.2% of the sites we checked refused ClaudeBot, and 4.8% refused Claude-SearchBot.

PerplexityBot

Perplexity's crawler that finds and links pages in Perplexity search; Perplexity says it is not used for training. Perplexity-User visits a page to answer a user's question.

In our audits: 7.2% of the sites we checked refused PerplexityBot.

Google-Extended

A robots.txt setting, not a separate crawler. It controls whether Google may use a site for training and grounding Gemini, and Google says it does not affect a site's inclusion in Google Search.

Blocked by name

A server or firewall refusing a request because of the crawler name it carries, while serving the same request under another name. robots.txt can say yes while the server says no.

In our audits: 19.5% of the 1357 sites we could test refused an AI crawler this way, against 3.9% that shut one out in robots.txt.

robots.txt

A plain text file at yourdomain.com/robots.txt that tells crawlers which parts of a site they may visit. Well-behaved crawlers follow it; it is a request, not a lock.

In our audits: 81.1% of the sites we checked used it to point crawlers to their sitemap.

XML sitemap

A file listing the pages a site wants search engines to find, with the dates they last changed. It helps them find pages; it does not make Google index them.

In our audits: 83% of the sites we checked had one we could find, and 9.2% had one their robots.txt never mentioned.

noindex

A rule in a page's code or headers telling search engines to leave the page out of their results. A noindexed page can't appear in Google, or in Google's AI Overviews.

In our audits: 9.2% of the sites we checked had at least one noindexed page, sometimes on purpose, often not.

Canonical tag

A tag naming the preferred address of a page when the same content can load at several addresses, so search engines know which one to show.

In our audits: 18.5% of the sites we checked had a page with no canonical tag.

Meta description

The short summary written for a page, often shown under its link in search results. Google writes its own when the summary doesn't match what was searched.

In our audits: 44.3% of the sites we checked had a page with none.

H1 (main heading)

The main heading of a page. Google is fine with more than one; what matters is that the main heading says what the page is about.

In our audits: 39.6% of the sites we checked had a page with no H1 at all.

nosnippet and max-snippet

Rules that stop or limit how much of a page Google may quote. Google's documentation says they also limit using the page as input for AI Overviews and AI Mode.

JavaScript rendering

Building a page's content in the browser with JavaScript instead of sending it ready-made from the server. Many AI crawlers read only the HTML they are sent.

In our audits: 0.4% of the sites we checked depended on it for their main content.

Structured data (schema markup)

Code, usually JSON-LD, that states facts about a page in the shared schema.org vocabulary: who the business is, what a page offers, who wrote it.

In our audits: 82.6% of the sites we checked had some.

sameAs

A structured-data field listing a business's other official profiles, so a machine can confirm the website and those profiles describe the same business.

In our audits: 47.5% of the sites we checked had none.

llms.txt

A proposed text file listing a site's key pages for AI tools. Google says its AI features need no new machine-readable or AI text files, so we don't count a missing one as a problem.

E-E-A-T

Experience, expertise, authoritativeness and trust: the qualities Google asks its quality raters to look for in a page and in the people and business behind it.

Scaled content abuse

Google's spam policy against producing many pages mainly to rank, with little value for readers, whether people or AI tools wrote them. It is judged page by page, not by the tool used.

Entity

A business, person or place that a search engine or AI tool recognises as one specific thing, because its name and details match across its website, profiles and listings.

Brand mention

Another site naming a business, with or without a link. Search engines and AI tools can read it as evidence that the business exists and that people talk about it.

In our audits: 96.7% of the sites we checked showed no press or "as seen in" mentions on the pages we read.

WCAG

The Web Content Accessibility Guidelines, the W3C standard an accessibility audit measures a website against.

ADA (for websites)

The Americans with Disabilities Act. The US Justice Department says it covers the websites of businesses open to the public, without setting a detailed technical rule for them. General information, not legal advice.

Service-area business

A business that serves customers where they are rather than at a shop or office. On Google Business Profile it can hide its address and list the areas it serves instead.